JP2017021257A - Encoder, encoding method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an encoder, an encoding method, and a program that keep down deterioration in audio quality when an audio signal is encoded by using a masking threshold value based on a hearing characteristic.SOLUTION: An encoder converts an audio signal into a frequency spectrum for each frame and quantizes and encodes the frequency spectrum. The encoder comprises: a threshold value generation unit for generating an initial masking threshold value when the frequency spectrum is quantized; a threshold value correction unit for correcting the initial masking threshold value on the basis of the amount of bits given to the quantization of the frequency spectrum and a condition whether a missing band due to quantization is permitted or not; and a threshold value determination unit for determining which is used for quantization, a first masking threshold value corrected according to a condition that a missing band due to quantization is permitted on the basis of the degree of variation of the frequency spectrum in a plurality of consecutive frames or a second masking threshold value corrected according to a condition that a missing band due to quantization is not permitted.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an encoding device, an encoding method, and a program.

  As a coding method of an audio signal or a sound signal (hereinafter collectively referred to as “audio signal”), there is a method of reducing the amount of information using human auditory characteristics such as an Advanced Audio Cording (AAC) method. In this type of encoding method, the number of bits required for encoding without increasing perceivable noise by suppressing the quantization error that increases when an audio signal is quantized with a small number of bits to a predetermined masking threshold or less (no increase in perceivable noise) That is, the amount of information) is reduced.

  The ideal value of the masking threshold is the upper limit value of the quantization error amount that cannot be perceived by humans. Therefore, the masking threshold is calculated based on the psychoacoustic model. Hereinafter, an ideal masking threshold calculated based on the psychoacoustic model is referred to as an initial masking threshold.

  However, in coding under a low bit rate condition such as 64 kbps or less, the number of usable bits is small, and the quantization error cannot often be suppressed below the initial masking threshold. If the quantization error cannot be suppressed below the initial masking threshold, the initial masking threshold is corrected based on the bit rate condition (see, for example, Non-Patent Document 1).

  In addition, as a method of efficiently using a limited amount of bits in encoding under a low bit rate condition, priorities are assigned to each band of the frequency spectrum, and the bit amount used for quantization (encoding) is changed between the bands. A method of adjusting is known (see, for example, Patent Document 1).

  Furthermore, as a method for preventing deterioration in sound quality in encoding under a low bit rate condition, a method is known in which bit amount allocation is determined based on a scale factor representative value obtained from a plurality of adjacent scale factors, and quantization is performed. (For example, see Patent Document 2).

JP-A-6-164409 JP-A-6-318875

"3GPP TS 26.403 V9.0.0", [online], 3GPP, March 8, 2015 search, Internet <URL: http://www.arib.or.jp/IMT-2000/V900Jul11/5_Appendix/Rel9/ 26 / 26403-900.pdf>

  Masking threshold correction methods are roughly classified into a method that is performed under a condition that allows a band to be lost due to quantization, and a method that is performed under a condition that does not allow a band to be lost.

  When the masking threshold is corrected under a condition that allows the loss of bands, if the correction amount increases, a band including sound that can be perceived by humans may be lost due to quantization. If a band including a sound that can be perceived by quantization (encoding) is lost, when the encoded audio signal is reproduced (decoded), a person who listens to the reproduced sound is uncomfortable. For this reason, when the number of missing bands increases, the sound quality deteriorates. Such deterioration of sound quality due to lack of a band becomes remarkable in an audio signal in which the peak position of the frequency spectrum fluctuates significantly, for example.

  On the other hand, when the masking threshold value is corrected under a condition that does not allow the loss of bands, an upper limit value is set for the masking threshold value of each band. Therefore, when the correction amount reaches the upper limit and a band that cannot be further corrected is generated, the correction amount of that band cannot be increased (in other words, the number of allocated bits cannot be reduced), and the correction amount of other bands is increased. It becomes. Therefore, the masking threshold value for the band having a large difference between the initial masking threshold value and the upper limit value is excessively corrected, and the number of bits allocated for encoding the band is reduced. A band having a large difference between the initial masking threshold and the upper limit value is a band important for sound quality. That is, when the masking threshold is corrected under conditions that do not allow band loss, the number of bits allocated to a band important for sound quality decreases, leading to deterioration of sound quality. Such deterioration of sound quality due to excessive correction of the masking threshold becomes significant in an audio signal in which the fluctuation of the peak position of the frequency spectrum such as harpsichord is gentle.

  In one aspect, an object of the present invention is to suppress deterioration in sound quality when an audio signal is encoded using a masking threshold based on auditory characteristics.

  An encoding device according to one aspect of the present invention is an encoding device that converts an audio signal into a frequency spectrum for each frame and quantizes and encodes the frequency spectrum, and includes a threshold generation unit, a threshold correction unit, And a threshold value determination unit. The threshold generation unit generates an initial masking threshold when quantizing the frequency spectrum based on the frequency spectrum. The threshold correction unit corrects the initial masking threshold based on a bit amount given to the quantization of the frequency spectrum and a condition as to whether or not band loss due to the quantization is allowed. The threshold value determination unit quantizes either the first masking threshold value corrected under a condition that allows loss of a band due to quantization or the second masking threshold value corrected under a condition that does not allow a loss of band due to quantization. Decide whether to use it. Here, the threshold value determination unit determines which of the first masking threshold value and the second masking threshold value is used for quantization based on the degree of change of the frequency spectrum in a plurality of consecutive frames. To do.

  According to the above-described aspect, it is possible to suppress deterioration in sound quality when an audio signal is encoded using a masking threshold based on auditory characteristics.

It is a functional block diagram of the encoding apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the structural example of the correction threshold value determination part in the encoding apparatus which concerns on 1st Embodiment. It is a flowchart (the 1) which shows the encoding process in the encoding apparatus which concerns on 1st Embodiment. It is a flowchart (the 2) which shows the encoding process in the encoding apparatus which concerns on 1st Embodiment. It is a flowchart (the 3) which shows the encoding process in the encoding apparatus which concerns on 1st Embodiment. It is a graph which shows an example of the relationship between the masking threshold corrected by the 1st correction part, and a frequency spectrum. It is a graph which shows an example of the relationship between the masking threshold corrected by the 2nd correction part, and a frequency spectrum. FIG. 25 is a schematic diagram illustrating an example of a hardware configuration of a computer that operates as an encoding device. It is a functional block diagram of the encoding apparatus which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the structural example of the correction threshold value determination part in the encoding apparatus which concerns on 2nd Embodiment. It is a flowchart which shows the procedure of the correction process of the masking threshold value in the encoding process which concerns on 2nd Embodiment. It is a functional block diagram of the encoding apparatus which concerns on the 3rd Embodiment of this invention. It is a block diagram which shows the structural example of the correction threshold value determination part in the encoding apparatus which concerns on 3rd Embodiment. It is a flowchart which shows the procedure of the correction process of the masking threshold value in the encoding process which concerns on 3rd Embodiment. It is a block diagram which shows the modification of the encoding apparatus which concerns on 3rd Embodiment. It is a flowchart which shows a part of encoding process in the modification of 3rd Embodiment. It is a functional block diagram of the encoding apparatus which concerns on the 4th Embodiment of this invention. It is a block diagram which shows the structural example of the correction threshold value determination part in the encoding apparatus which concerns on 4th Embodiment. It is a flowchart which shows a part of encoding process which concerns on 4th Embodiment. It is a flowchart which shows a part of other encoding process based on 4th Embodiment. It is a functional block diagram of the encoding apparatus which concerns on the 5th Embodiment of this invention. It is a block diagram which shows the structural example of the correction threshold value determination part in the encoding apparatus which concerns on 5th Embodiment. It is a flowchart which shows a part of encoding process which concerns on 5th Embodiment. It is a flowchart which shows another part of the encoding process which concerns on 5th Embodiment.

[First Embodiment]
FIG. 1 is a functional block diagram of an encoding apparatus according to the first embodiment of the present invention.

  As shown in FIG. 1, the encoding device 1 according to the present embodiment includes a block switching unit 10, an MDCT processing unit 11, a masking threshold generation unit 12, an auditory characteristic calculation unit 13, and a masking threshold correction unit 14. A quantization unit 15, an encoding unit 16, and a multiplexing unit 17. In addition, the encoding device 1 includes a storage unit 18.

  The block switching unit 10 switches the block length when performing Modified Discrete Cosine Transform (MDCT) processing on the input signal based on the characteristics of the input signal (audio signal). For example, in AAC encoding, the block length is switched to a long block (1024 points) or a short block (128 points).

  The MDCT processing unit 11 performs MDCT processing with a window length corresponding to the long block or the short block on the input signal, and converts the input signal into a frequency spectrum. For example, in AAC coding, if the block length is a long block, MDCT processing with a window length of 2048 is performed, and if the block length is short, MDCT processing with a window length of 256 is performed.

The masking threshold generation unit 12 performs auditory psychological analysis on the input signal, and generates an optimal masking threshold (initial masking threshold) sfbThr ₀ (sfb) in the quantization of the frequency spectrum obtained from the input signal. The initial masking threshold sfbThr ₀ (sfb) is generated for each band (scale factor band sfb in the AAC scheme; hereinafter also referred to as “band sfb”). Further, the masking threshold value generator 12 determines the band sfb to be encoded based on the power value (input power) mdct_pow (sfb) and the initial masking threshold value sfbThr ₀ (sfb) in each frequency band sfb. Further, after determining the band sfb to be encoded, the masking threshold generation unit 12 determines whether quantization using the initial masking threshold is possible, in other words, whether the initial masking threshold needs to be corrected. To do. When quantization using the initial masking threshold is possible, the masking threshold generation unit 12 passes the initial masking threshold to the quantization unit 15. On the other hand, when quantization cannot be performed using the initial masking threshold, the masking threshold generation unit 12 passes the initial masking threshold sfbThr ₀ (sfb) to the auditory characteristic calculation unit 13 and the first correction unit 141 of the masking threshold correction unit 14. At the same time, it is stored in the storage unit 18.

  The auditory characteristic calculator 13 calculates the auditory characteristic necessary for correcting the masking threshold. The auditory characteristic calculation unit 13 of the present embodiment calculates a signal mask ratio (SMR) of each band sfb as the auditory characteristic. The auditory characteristic calculation unit 13 passes the calculated signal-to-mask ratio smr (sfb) to the first correction unit 141 of the masking threshold correction unit 14 and stores it in the storage unit 18.

  The masking threshold correction unit 14 corrects the initial masking threshold based on the auditory characteristics (signal to mask ratio smr (sfb)) and the bit rate condition. The masking threshold value correction unit 14 of the present embodiment includes a first correction unit 141, a second correction unit 142, and a correction threshold value determination unit 143.

The first correction unit 141 corrects the initial masking threshold under a condition that allows loss of a band due to quantization. The second correction unit 142 corrects the initial masking threshold under a condition that does not allow band loss due to quantization. The correction threshold value determination unit 143 quantizes either the masking threshold value sfbThr ₁ (sfb) corrected by the first correction unit 141 or the masking threshold value sfbThr ₂ (sfb) corrected by the second correction unit 142. Decide whether to use it. Hereinafter, the masking threshold sfbThr ₁ (sfb) corrected by the first correction unit 141 is also referred to as a first correction threshold, and the masking threshold sfbThr ₂ (sfb) corrected by the second correction 142 is the second correction threshold. Also called.

In the masking threshold correction unit 14, first, the first correction unit 141 corrects the initial masking threshold sfbThr ₀ (sfb) to the first correction threshold sfbThr ₁ (sfb). Then, the correction threshold value determination unit 143 determines whether or not to adopt the _first correction threshold value sfbThr ₁ (sfb). As will be described later, the correction threshold value determination unit 143 determines whether or not to adopt the first correction threshold value based on the missing SFB variation rate for the first correction threshold value. When it is determined that the first correction threshold value is adopted, the correction threshold value determination unit 143 determines the first correction threshold value as a masking threshold value used for quantization, and passes the first correction threshold value to the quantization unit 15. On the other hand, when it is determined that the first correction threshold value is not adopted, the correction threshold value determination unit 143 determines that the second correction threshold value is used for quantization, and causes the second correction unit 142 to correct the initial masking threshold value. The second correction unit 142 corrects the initial masking threshold value to the second correction threshold value sfbThr ₂ (sfb), and then passes the _second correction threshold value sfbThr ₂ (sfb) to the quantization unit 15.

The quantization unit 15 uses one of the initial masking threshold value sfbThr ₀ (sfb), the first correction threshold value sfbThr ₁ (sfb), and the second correction threshold value sfbThr ₂ (sfb) to be encoded in the frequency spectrum. Quantize the band.

  The encoding unit 16 encodes a value obtained by quantizing the frequency spectrum. In the case of encoding in the AAC scheme, the encoding unit 16 performs Huffman encoding on the value obtained by quantization.

  The multiplexing unit 17 multiplexes the encoded audio signal to generate an encoded stream.

The storage unit 18 stores information used for correction of the masking threshold, such as the initial masking threshold sfbThr ₀ (sfb) and the signal-to-mask ratio smr (sfb), in a state where it can be read and rewritten.

  FIG. 2 is a block diagram illustrating a configuration example of the correction threshold value determination unit in the encoding device according to the first embodiment.

  As illustrated in FIG. 2, the correction threshold value determination unit 143 includes a missing SFB fluctuation rate calculation unit 143a, a missing SFB information holding unit 143b, and a determination unit 143c.

The missing SFB fluctuation rate calculation unit 143a calculates the fluctuation rate Vsfb of the band sfb missing due to quantization based on the first correction threshold value sfbThr ₁ (sfb) and the frequency spectrum power value mdct_pow (sfb). Hereinafter, the fluctuation rate Vsfb calculated by the missing SFB fluctuation rate calculation unit 143a is referred to as a missing SFB fluctuation rate.

The missing SFB fluctuation rate Vsfb in the present embodiment represents a ratio of a band that is continuously missing over several frames in a band that is missing in a current encoding target frame (T frame). The missing band is a band in which the relationship between the first correction threshold value sfbThr ₁ (sfb) and the power value mdct_pow (sfb) of the frequency spectrum is sfbThr ₁ (sfb)> mdct_pow (sfb). Further, in the present embodiment, regarding a band that is missing in a T frame, if the same band or a band before and after the same band is missing in a frame temporally prior to the T frame, it is determined that the band is missing continuously.

  The missing SFB information holding unit 143b holds missing SFB information for several frames used for calculating the missing SFB fluctuation rate Vsfb. The missing SFB information is a set of missing bands in each frame. In this embodiment, as shown in FIG. 2, the missing SFB information of the T-1 frame (one frame before the T frame) and the missing SFB information of the T-2 frame (two frames before the T frame) are retained. .

The determination unit 143c determines whether or not to adopt the _first correction threshold value sfbThr ₁ (sfb) as a masking threshold value for quantization based on the calculated missing SFB fluctuation rate Vsfb. Determining unit 143c, when missing SFB variation rate Vsfb the threshold TH ₁ below a predetermined determines to adopt the first correction threshold masking threshold when quantizing. That is, it is determined that the first correction threshold is used for quantization. In this case, the determination unit 143 passes the first correction threshold value to the quantization unit 15.

On the other hand, when the missing SFB fluctuation rate Vsfb is smaller than the threshold value, the determination unit 143 determines not to employ the first correction threshold value, that is, to use the second correction threshold value sfbThr ₂ (sfb) for quantization. In this case, the determination unit 143c causes the second correction unit 142 to correct the initial masking threshold.

As described above, in the encoding device 1 according to the present embodiment, when the first correction threshold value sfbThr ₁ (sfb) corrected under the condition that allows band loss is satisfied, the first correction is performed. The frequency spectrum is quantized using a threshold value. On the other hand, when the first correction threshold value does not satisfy the predetermined adoption condition, the frequency spectrum is quantized using the _second correction threshold value sfbThr ₂ (sfb) corrected under a condition that does not allow band loss.

Hereinafter, the content of the encoding process in the encoding device 1 according to the present embodiment will be described.
FIG. 3A is a flowchart (part 1) illustrating an encoding process in the encoding device according to the first embodiment. FIG. 3B is a flowchart (part 2) illustrating the encoding process in the encoding device according to the first embodiment. FIG. 3C is a flowchart (part 3) illustrating the encoding process in the encoding device according to the first embodiment.

  The encoding apparatus 1 according to the present embodiment performs encoding processing as shown in FIGS. 3A to 3C on each encoding unit data such as a frame in an input signal (audio signal).

  As shown in FIG. 3A, the encoding device 1 first converts an input signal for one frame into a frequency spectrum, and calculates a power value mdct_pow (sfb) of each band sfb (step S10). The block switching unit 10 and the MDCT processing unit 11 perform the process in step S10.

  The block switching unit 10 selects and switches whether the block length of the MDCT processing is a long block or a short block. The block length is selected based on a known selection method, for example, an input signal power fluctuation ratio and a predicted gain fluctuation ratio.

  The MDCT processing unit 11 performs an MDCT process with a window length corresponding to the block length selected by the block switching unit 10 and converts the input signal into a frequency spectrum. Thereafter, the MDCT processing unit 11 calculates the power value mdct_pow (sfb) of each band sfb based on the obtained frequency spectrum. The MDCT process and the calculation of the power value are performed using either a known conversion method or calculation method, respectively.

Next, the encoding device 1 generates an initial masking threshold sbfThr ₀ (sfb) for quantizing the frequency spectrum (step S12). The process of step S12 is performed by the masking threshold value generator 12.

The masking threshold value generation unit 12 performs auditory psychological analysis on the input signal and obtains an initial masking threshold value sfbThr ₀ (sfb) for each band sfb. The initial masking threshold value sfbThr ₀ (sfb) is calculated using any known calculation method based on the minimum audible level, the masking effect, etc. in each band sfb.

In addition, when the masking threshold generation unit 12 generates the initial masking threshold sfbThr ₀ (sfb), the masking threshold generation unit 12 then generates an encoding target based on the initial masking threshold sfbThr ₀ (sfb) and the frequency spectrum power value mdct_pow (sfb). A band is determined (step S14). In the process of step S14, the masking threshold value generation unit 12 sets only the band of sfbThr ₀ (sfb) <mdct_pow (sfb) among all the bands of the frequency spectrum as the encoding target.

  After determining the encoding target band, the masking threshold value generator 12 calculates an initial PE value and a target PE value in order to determine whether or not to correct the initial masking threshold value (step S16). In the present embodiment, it is determined whether or not to correct the masking threshold based on whether or not the initial PE value is larger than the target PE value (step S18).

Here, the PE value is a value of perceptual entropy which is one of acoustic parameters, and represents the number of bits necessary for quantization so as not to perceive noise. The initial PE value is a PE value calculated based on the power value mdct_pow (sfb) and the initial masking threshold sfbThr ₀ (sfb) in the band to be encoded. The target PE value is a PE value calculated based on the number of bits that can be used for encoding. The initial PE value and the target PE value are calculated using any known calculation method (for example, the calculation method described in Non-Patent Document 1).

  The perceptual entropy value is related to the number of bits required for quantization as described above. If the initial PE value is larger than the target PE value, the amount of bits used for quantization using the initial masking threshold can be used. It can be determined that the number of bits exceeds the maximum. On the other hand, when the initial PE value is less than or equal to the target PE value, it can be determined that the amount of bits used in quantization using the initial masking threshold falls within the number of usable bits. Therefore, it can be determined whether quantization using the initial masking threshold is possible based on the magnitude relationship between the initial PE value and the target PE value, that is, whether the initial masking threshold needs to be corrected.

When the initial PE value is less than or equal to the target PE value (step S18; No), the masking threshold value generation unit 12 determines that the masking threshold value is not corrected, and passes the initial masking threshold value sfbThr ₀ (sfb) to the quantization unit 15. In this case, as illustrated in FIG. 3C, the encoding apparatus 1 quantizes the frequency spectrum using the initial masking threshold sfbThr ₀ (sfb) (step S30). The quantization unit 15 performs the quantization in step S30. The quantization unit 15 quantizes the frequency spectrum using any of the known quantization methods.

On the other hand, when the initial PE value is larger than the target PE value (step S18; Yes), the masking threshold value generator 12 determines to correct the masking threshold value. In this case, the masking threshold generation unit 12 passes the initial masking threshold sfbThr ₀ (sfb) to the auditory characteristic calculation unit 13. Further, the masking threshold value generation unit 12 passes the initial masking threshold value sfbThr ₀ (sfb) and the target PE value to the first correction unit 141 of the masking threshold value correction unit 14 and causes the storage unit 18 to store them. Thereafter, the encoding apparatus 1 performs a masking threshold value correction process as in steps S20 to S28 illustrated in FIG. 3B.

  When correcting the masking threshold, the encoding apparatus 1 next calculates an auditory characteristic based on the frequency spectrum or the like (step S20). The process of step S20 is performed by the auditory characteristic calculation unit 16.

The auditory characteristic calculator 16 calculates a signal-to-mask ratio smr (sfb) in each band sfb, that is, a difference value between the power value mdct_pow (sfb) and the initial masking threshold value sfbThr ₀ (sfb) in each band. When the signal-to-mask ratio smr (sfb) is calculated, the auditory characteristic calculation unit 16 passes the calculated signal-to-mask ratio smr (sfb) to the first correction unit 141 of the masking threshold correction unit 14 and also stores the storage unit 18. Remember me.

  When the calculation of the auditory characteristic (signal to mask ratio) is completed, the encoding apparatus 1 next corrects the initial masking threshold based on the auditory characteristic and the like under a condition that allows band loss due to quantization (step S22). ). Step S22 is performed by the first correction unit 141 of the masking threshold correction unit 14.

For example, the first correction unit 141 uses the following formula (1), and the PE value calculated based on the power value mdct_pow (sfb) and the corrected masking threshold value sfbThr ₁ (sfb) becomes the target PE value. The masking threshold is corrected as follows.

  In Equation (1), r is a correction parameter (see Non-Patent Document 1).

Further, the first correction unit 141 performs the correction process in step S22 without setting an upper limit value for the corrected masking threshold value sfbThr ₁ (sfb) in order to allow a loss of the band. When the first correction unit 141 completes the correction of the masking threshold value, the corrected masking threshold value (first correction threshold value sfbThr ₁ (sfb)) is used as the missing SFB fluctuation rate calculation unit 143a of the correction threshold value determination unit 143. And to the determination unit 143c.

  When the process of correcting the initial masking threshold to the first correction threshold is completed, the encoding apparatus 1 next calculates the missing SFB variation rate Vsfb of the corrected masking threshold (first correction threshold) (step S24). ). Step S24 is performed by the missing SFB fluctuation rate calculation unit 143a.

The missing SFB fluctuation rate calculation unit 143a first uses the _first correction threshold sfbThr ₁ (sfb) and the power value mdct_pow (sfb) for the current encoding target frame (T frame) as the process of step S24. Band to be extracted. Next, the missing SFB fluctuation rate calculation unit 143a reads the missing SFB information of the T-1 frame and the missing SFB information of the T-2 frame from the missing SFB information holding unit 143b, and performs each missing for each band in which the T frame is missing. It collates with the missing band included in the SFB information. If each missing SFB information includes the same band as the one missing band in the T frame or the bands before and after that, it is determined that the one missing band is a continuously missing band. Thereafter, the missing SFB fluctuation rate calculation unit 143a passes the value obtained by dividing the number of continuously missing bands by the total number of missing bands in the T frame to the determination unit 143c as the missing SFB fluctuation rate Vsfb. Further, when the calculation of the missing SFB fluctuation rate Vsfb is completed, the missing SFB fluctuation rate calculation unit 143a updates the information held in the missing SFB information holding unit 143b. That is, the missing SFB information of the T-1 frame is held as the missing SFB information of the T-2 frame, and the missing SFB information of the T frame is held as the missing SFB information of the T-1 frame.

Note that the missing SFB fluctuation rate Vsfb is not limited to the above method, and may be calculated by another method. Further, when the missing SFB information of the previous frame and the previous frame is not held as in encoding the first few frames in the audio signal, the missing SFB fluctuation rate calculation unit 143a sets a predetermined value. The missing SFB fluctuation rate Vsfb. The predetermined value is, for example, to a value smaller than the threshold value TH ₁ to be compared with the missing SFB variation rate Vsfb in the next step.

When the calculation of the missing SFB fluctuation rate Vsfb is finished, the encoding apparatus 1 next determines the missing SFB fluctuation rate Vsfb in order to determine which of the first correction threshold value and the second correction threshold value is used for quantization. comparing the threshold value TH ₁ determined in advance (step S26). Step S26 is performed by the determination unit 143c. When Vsfb> TH ₁ (step S26; Yes), the determination unit 143c determines that the first correction threshold is adopted as the masking threshold for quantization, and passes the first correction threshold to the quantization unit 15. Thereby, the masking threshold used for quantization is determined to be the first correction threshold sfbThr ₁ (sfb). In this case, as illustrated in FIG. 3C, the quantization unit 15 quantizes the frequency spectrum using the first correction threshold, that is, the masking threshold corrected under the condition that allows the loss of the band (step S32). On the other hand, when Vsfb ≦ TH ₁ (step S26; No), the determination unit 143c determines that the first correction threshold is not adopted as the masking threshold for quantization, and sets the masking threshold to the second correction unit 142. Let me correct it. Thereby, the masking threshold value used for quantization is determined to be the second correction threshold value sfbThr ₂ (sfb).

When the second correction unit 142 receives a signal instructing correction of the masking threshold value from the determination unit 143c, the second correction unit 142 corrects the masking threshold value under a condition that does not allow band loss due to quantization based on auditory characteristics or the like (step S102). S28). In step S28, similarly to step S22, the masking threshold value is corrected using Equation (1) so that the PE value calculated from the corrected masking threshold value and the power value becomes the target PE value. However, in step S28, the initial masking threshold value is corrected under a condition that does not allow band loss. Therefore, the second correction unit 142 sets the upper limit value at which no band is lost to the corrected masking threshold value sfbThr ₂ (sfb), and performs the correction process in step S28. The band lost due to quantization is a band whose masking threshold is larger than the power value mdct_pow (sfb) of the frequency spectrum. Therefore, the upper limit value of the masking threshold value sfbThr ₂ (sfb) after correction is set to a value equal to or lower than the power value mdct_pow (sfb) of the frequency spectrum.

  After completing the correction of the masking threshold, the second correction unit 142 passes the corrected masking threshold (second correction threshold) to the quantization unit 15. In this case, as shown in FIG. 3C, the quantizing unit 15 quantizes the frequency spectrum using the masking threshold corrected under the condition that the loss of the band is not allowed (step S34).

When the frequency spectrum is quantized using one of the initial masking threshold value sfbThr ₀ (sfb), the first correction threshold value sfbThr ₁ (sfb), and the second correction threshold value sfbThr ₂ (sfb), the quantization unit 15 Then, the quantized value is passed to the encoding unit 16. Then, the encoding unit 16 encodes the quantized value using a known encoding method such as fixed Huffman encoding (step S36). When the encoding is completed, the encoding unit 16 passes the encoded data to the multiplexing unit 17. Thereby, the encoding process for one frame of the input signal (audio signal) is completed.

  When the encoding process is completed, the encoding device 1 (multiplexer 17) generates and outputs an encoded stream in which header information or the like is added to the encoded audio signal (audio data).

  The missing SFB fluctuation rate Vsfb calculated in the present embodiment is a ratio of the number of bands sfb missing in three consecutive frames in the missing bands in the current encoding target frame. Therefore, the missing SFB fluctuation rate Vsfb increases as the proportion of the band missing continuously in a plurality of frames increases. That is, the smaller the proportion of bands that are continuously missing in a plurality of frames, the smaller the missing SFB fluctuation rate Vsfb.

  In an audio signal in which the peak position (power value pattern) of the frequency spectrum changes drastically, the missing band also changes drastically, and therefore, the missing band in several consecutive frames varies from frame to frame. Therefore, quantizing an audio signal whose frequency spectrum peak position changes drastically using a masking threshold value corrected under conditions that allow omissions, for example, leads to deterioration in sound quality such that the original sound that is continuously heard is interrupted. . Therefore, when this type of audio signal is quantized, it is preferable to quantize using a masking threshold corrected under conditions that do not allow band loss.

An audio signal in which the peak position of the frequency spectrum changes drastically and the bands that are lost in several consecutive frames are separated has a small ratio of the bands that are continuously lost in a plurality of frames. Therefore, the missing SFB fluctuation rate Vsfb calculated by the above method becomes small. Thus, in this embodiment, if the missing SFB variation rate Vsfb audio signal is smaller than a predetermined threshold value TH _1, using the masking threshold corrected under a condition that does not allow missing quantization. As a result, an audio signal whose peak position of the frequency spectrum changes drastically can be quantized using a masking threshold corrected under a condition that does not allow omission. Therefore, it is possible to suppress deterioration in sound quality such that the original sound that can be heard continuously is interrupted after encoding.

  On the other hand, in the masking threshold value corrected under the condition that the omission is not allowed, a band important in terms of sound quality is excessively corrected. If quantization is performed using such an overcorrected masking threshold, a perceptible quantization error is likely to occur, leading to deterioration of sound quality. Such deterioration of sound quality due to excessive correction becomes prominent when quantization is performed using an audio signal in which the peak position of a frequency spectrum such as a harpsichord is moderately changed, or a masking threshold value corrected under conditions that do not allow omission.

An audio signal with a gradual change in the peak position of the frequency spectrum has a gradual fluctuation in the missing band, and there are many missing bands over several consecutive frames. For this reason, an audio signal with a gradual change in the peak position of the frequency spectrum has a large missing SFB fluctuation rate Vsfb calculated by the above method. In this embodiment, if the missing SFB variation rate Vsfb of the audio signal of a predetermined threshold value TH ₁ or more, using the masking threshold corrected in conditions permitting missing quantization. Therefore, an audio signal with a gradual change in the peak position of the frequency spectrum is quantized using a masking threshold value corrected under conditions that allow omission. Therefore, it is possible to suppress deterioration in sound quality due to excessive correction of the masking threshold of a band important for sound quality.

  The masking threshold value correction process will be specifically described with reference to FIGS. 4A and 4B.

  FIG. 4A is a graph showing an example of the relationship between the masking threshold corrected by the first correction unit and the frequency spectrum.

When the processing of steps S10 to S14 is performed on the input signal for one frame, for example, the power value mdct_pow (sfb) and the initial masking threshold value sfbThr ₀ (sfb) of each band sfb of the frequency spectrum as shown in FIG. 4A are obtained. can get. In FIG. 4A, the power value mdct_pow (sfb) and the initial masking threshold value sfbThr ₀ (sfb) of each band sfb of the frequency spectrum for three consecutive frames of T-2 frame, T-1 frame, and T frame are shown. Show. Of these three frames, the lower T frame is the current frame to be encoded. The middle T-1 frame is a frame immediately before the T frame, and the upper T-2 frame is a frame two frames before the T frame. In the graph shown in FIG. 4A, the lower T frame and the upper T-2 frame have approximately the same shape of the power value mdct_pow (sfb) and the initial masking threshold sfbThr ₀ (sfb). The outline of one frame is completely different.

As described above, the initial masking threshold value sfbThr ₀ (sfb) is an optimal masking threshold value for quantization of the corresponding frequency spectrum. Therefore, when it is determined that the quantization using the initial masking threshold sfbThr ₀ (sfb) is possible by the processing in steps S16 and S18, the encoding device 1 uses the initial masking threshold sfbThr ₀ (sfb) to perform frequency spectrum analysis. Is quantized (step S30).

  However, when encoding is performed under a low bit rate condition, that is, when the number of bits that can be used for encoding a frequency spectrum is small, the quantization error often cannot be made lower than the initial masking threshold. When the quantization error cannot be made equal to or less than the initial masking threshold, the encoding apparatus 1 performs correction for increasing (relaxing) the masking threshold within a range where the sound quality is not deteriorated as much as possible based on the bit rate condition, auditory characteristics, and the like.

In the encoding apparatus 1 according to the present embodiment, as described above, first, the masking threshold is corrected under a condition that allows loss of a band due to quantization (step S22). At this time, the corrected masking threshold values ( _first correction threshold values) sfbThr ₁ (sfb) in the three frames shown in FIG. 4A have the outlines indicated by the solid broken lines. Whether to use the _first correction threshold value sfbThr ₁ (sfb) for the quantization of the T frame is determined based on the missing SFB fluctuation rate Vsfb.

The band that is lost when the first correction threshold sfbThr ₁ (sfb) is used for the quantization of the T frame is a band that satisfies sfbThr ₁ (sfb)> mdct_pow (sfb). In the example shown in the lower part of FIG. 4A, there are three bands sfb5, sfb9, and sfb12 that are lost due to quantization. On the other hand, the bands sfb3 and sfb8 are missing in the T-1 frame, and the bands sfb5, sfb9, and sfb12 are missing in the T-2 frame.

In the T-1 frame, three bands of the band sfb5 and the bands sfb4 and sfb6 before and after the band sfb5 are not missing. In the T-1 frame, the band sfb12 and the previous band sfb11 are not lost. On the other hand, the band sfb8 before the band sfb9 is missing in the T-1 frame, and the band sfb9 is missing in the T-2 frame. That is, in the example shown in FIG. 4A, it is only the band sfb9 that is determined to be continuously missing over three frames among the three bands sfb5, sfb9, and sfb12 in which the T frame is missing. Therefore, the missing SFB fluctuation rate Vsfb for the T frame in the example shown in FIG. 4A is 1/3. Therefore, if a value smaller than the threshold value TH ₁ is 1/3 compared to the missing SFB variation rate Vsfb in step S26, the determination in step S26 is "Yes", the frequency spectrum of the T frame first correction threshold sfbThr ₁ (sfb) is used for quantization. On the other hand, if the value of the threshold value TH ₁ is 1/3 or more, the determination is "No" in step S26, the frequency spectrum of the T frame to be quantized using the second correction threshold sfbThr ₂ (sfb) Become.

  FIG. 4B is a graph showing an example of the relationship between the masking threshold corrected by the second correction unit and the frequency spectrum.

The corrected masking threshold value sfbThr ₂ (sfb) when the masking threshold value is corrected by the second correction unit 142 has, for example, an outline as shown by a solid line in FIG. 4B. FIG. 4B shows an example in which the second masking unit 142 corrects the initial masking threshold for the T frame shown in the lower part of FIG. 4A. In FIG. 4B, the first correction threshold value sfbThr ₁ (sfb) is indicated by a two-dot chain line for comparison.

The second correction threshold value sfbThr ₂ (sfb) is a masking threshold value corrected under a condition that does not allow band loss. Therefore, as shown in FIG. 4B, sfbThr ₂ (sfb) <mdct_pow (sfb) is satisfied in all the bands including the bands sfb5, sfb9, and sfb12 that are missing in the first correction threshold. However, in order to prevent the loss of the bands sfb5, sfb9, and sfb12, the number of bits in these bands cannot be reduced, and the number of bits is reduced by increasing the masking threshold correction amount in other bands. Therefore, the masking threshold for the bands sfb1 and sfb2 is excessively corrected.

In encoding an audio signal, it is ideal to quantize using an initial masking threshold. Therefore, when correcting the masking threshold, it is preferable that the similarity between the outline of the corrected masking threshold and the outline of the initial masking threshold is high. However, when the masking threshold is corrected under a condition that does not allow the loss of the band, as shown in FIG. 4B, the deviation of the approximate similarity with the initial masking threshold increases. Therefore, in the case of an audio signal in which the fluctuation of the peak position of the frequency spectrum is gentle, for example, in the case of an audio signal in which a spectrum pattern similar to the T frame in FIG. 4B continues, a quantization error in an excessively corrected band is perceived. It leads to deterioration of sound quality. Therefore, fluctuation gentle missing SFB volatility Vsfb the peak position of the frequency spectrum is greater than the threshold value TH _1, is quantized using the masking threshold corrected in conditions permitting missing. This suppresses deterioration in sound quality that may occur when quantization is performed using a masking threshold corrected under conditions that do not allow band loss.

  The encoding apparatus 1 of the present embodiment that performs the encoding process as described above can be realized by, for example, a computer and a program that causes the computer to execute the encoding process. Hereinafter, the encoding apparatus 1 realized by a computer and a program will be described with reference to FIG.

  FIG. 5 is a schematic diagram illustrating an example of a hardware configuration of a computer that operates as an encoding apparatus.

  As shown in FIG. 5, the computer 5 that operates as an encoding device includes a Central Processing Unit (CPU) 50, a main storage device 51, an auxiliary storage device 52, an input device 53, and an output device 54. . The computer 5 further includes a digital signal processor (DSP) 55, a storage medium driving device 56, and an interface device 57. These elements 50 to 57 in the computer 5 are connected to each other by a bus 59 so that data can be exchanged between the elements.

  The CPU 50 is an arithmetic processing unit that controls the overall operation of the computer 5 by executing various programs including an operating system.

  The main storage device 51 includes a read only memory (ROM) 51a and a random access memory (RAM) 51b. In the ROM 51a, for example, a predetermined basic control program read by the CPU 50 when the computer 5 is started is recorded in advance. The RAM 51b is used as a working storage area as necessary when the CPU 50 executes various programs. In the present embodiment, the RAM 51b is used for temporary storage of, for example, an audio signal to be encoded and a masking threshold.

  The auxiliary storage device 52 is a storage device with a larger capacity than the main storage device 51 such as a hard disk drive (HDD) or a solid state drive (SSD). The auxiliary storage device 52 stores various programs executed by the CPU 50, various data, and the like. Examples of the program stored in the auxiliary storage device 52 include an audio player program that encodes and reproduces an audio signal. The data stored in the auxiliary storage device 52 includes, for example, audio signal data encoded by the player.

  The input device 53 is, for example, a keyboard device or a mouse device. When the input device 53 is operated by an operator of the computer 5, the input device 53 transmits input information associated with the operation content to the CPU 50.

  The output device 54 is, for example, a liquid crystal display or a speaker. The liquid crystal display displays various texts, images, and the like according to display data transmitted from the CPU 50 or the like. The speaker outputs audio data and audio data transmitted from the CPU 50, DSP 55, and the like.

  The DSP 55 is an arithmetic processing unit that performs audio signal encoding processing, decoding (reproduction) processing, and the like in accordance with control signals from the CPU 50.

  The storage medium driving device 56 reads programs and data recorded in a portable storage medium (not shown) and writes data stored in the auxiliary storage device 52 to the portable storage medium. As the portable storage medium, for example, a flash memory equipped with a USB standard connector can be used. Further, as a portable storage medium, an optical disc such as a Compact Disk (CD), a Digital Versatile Disc (DVD), and a Blu-ray Disc (Blu-ray is a registered trademark) can be used.

  The interface device 57 is, for example, an audio input / output device or a communication control device. The audio input / output device inputs and outputs an audio signal by connecting the computer 5 to a microphone or an audio device, for example. The communication control device communicatively connects the computer 5 and a communication network such as the Internet, and transmits and receives audio data and the like by communication with an external communication device or the like via the communication network.

  In this computer 5, the CPU 50 reads out the program including the above-described encoding process from the auxiliary storage device 52, and executes the encoding process of the audio signal in cooperation with the DSP 55, the main storage device 51, the auxiliary storage device 52, and the like. To do. At this time, the CPU 50 causes the DSP 55 to execute arithmetic processing in the encoding process. The DSP 55 converts the audio signal into a frequency spectrum and generates an initial masking threshold. The audio signal may be read from a portable storage medium such as a music CD, for example, or may be input to the computer 5 by communication via the interface device 57. Further, the DSP 55 calculates an initial PE value and a target PE value, and determines whether or not the audio signal can be quantized using the initial masking threshold from the magnitude relationship between them. If the quantization cannot be performed using the initial masking threshold, the DSP 55 calculates the auditory characteristic and corrects the masking threshold under a condition that allows the lack of a band due to the quantization. Further, the DSP 55 calculates the missing SFB fluctuation rate Vsfb based on the corrected masking threshold (first correction threshold), and determines whether or not the first correction threshold is used for quantization. When it is determined that the first correction threshold is used for quantization, the frequency spectrum is quantized using the first correction threshold. On the other hand, when it is determined that the first correction threshold value is not used for quantization, the initial masking threshold value is corrected under a condition that does not allow band loss, and the frequency spectrum is quantized using the corrected masking threshold value. Further, after quantization, the DSP 55 performs encoding and multiplexing. In addition, the DSP 55 stores the initial masking threshold, the target PE value, the auditory characteristic (signal-to-mask ratio), and the like in the RAM 51b and the auxiliary storage device 52, and the RAM 51b and auxiliary during the above processing. Processing to read from the storage device 52 is performed.

  Audio signal data (audio data) encoded by the computer 5 is stored in, for example, the auxiliary storage device 52, and is decoded (reproduced) by the computer 5 as necessary. Further, if the computer 5 includes a communication control device as the interface device 57, for example, audio data can be provided (distributed) to another computer or the like via a communication network.

  Note that the computer 5 used as the encoding device 1 is not limited to the configuration shown in FIG. 5, and may be configured such that the CPU 50 performs audio signal encoding processing. In addition, the computer 5 used as the encoding device 1 is not limited to a general-purpose computer that realizes a plurality of functions by executing various programs, but may be an audio device specialized for encoding and decoding audio signals. Good.

  As described above, according to the first embodiment, when the masking threshold used for quantization is corrected, first, the masking threshold is set under the condition that the first correction unit 141 allows the loss of the band due to quantization. to correct. Then, based on the fluctuation rate (time continuity) of the band sfb that is lost when the masking threshold corrected by the first correction unit 141 is used for quantization, either of the conditions that allow the band to be lost or the conditions that do not allow it It is determined whether the masking threshold corrected under the above conditions is used for quantization. Therefore, it is possible to perform quantization using a masking threshold value that is less susceptible to deterioration in sound quality, between masking corrected by allowing band loss and corrected masking threshold not allowed depending on the characteristics of the audio signal to be encoded. it can. For example, when encoding an audio signal with significant fluctuations in the peak position of the frequency spectrum, it is quantized using a masking threshold corrected under conditions that do not allow omissions, thereby suppressing deterioration in sound quality due to continuous sound interruptions, etc. it can. On the other hand, when encoding an audio signal with a gradual fluctuation in the peak position of the frequency spectrum, quantization is performed using the masking threshold corrected under the conditions that allow omission, so that the masking threshold is excessively corrected, resulting in quantization error. It is possible to suppress deterioration of sound quality due to an increase in.

  Further, it is determined based on the missing SFB fluctuation rate Vsfb which one of the masking threshold corrected under the condition allowing the loss of the band and the masking threshold corrected under the condition not allowing the missing is used for the quantization. As described above, the missing SFB fluctuation rate Vsfb is a ratio of a band that is continuously lost for several frames out of the band sfb that is missing in the current encoding target frame. For this reason, which of the masking threshold value corrected under conditions that allow band loss and the masking threshold value corrected under conditions that do not allow loss are used for quantization without analyzing the characteristics of the input signal (audio signal) or the like? Can be easily determined.

  Further, in the present embodiment, during the encoding process, the encoding apparatus 1 automatically determines in real time and automatically whether to use a masking threshold that allows missing or a masking threshold that does not allow for quantization. Also, since the masking threshold is determined based on the missing SFB fluctuation rate calculated from the number of missing bands and the continuity of the missing, it is possible to determine and quantize a masking threshold that can further suppress deterioration in sound quality with a small amount of calculation. it can. Therefore, the encoding apparatus and the encoding method of the present embodiment can be easily applied to a purpose of outputting (distributing) an input audio signal such as a live broadcast while sequentially encoding it.

  In the present embodiment, the missing band change rate is determined using the missing SFB information for the previous frame and the previous two frames. However, the present invention is not limited to this, and missing for more frames. The fluctuation rate of the missing band may be determined using the SFB information. In addition, the fluctuation rate of the missing band is not limited to the ratio of the missing band continuously in a plurality of frames in the missing band, and may be calculated by other calculation methods.

  In the present embodiment, as illustrated in FIG. 1, two correction units including a first correction unit 141 and a second correction unit 142 are provided as correction units that correct the masking threshold. However, the correction of the masking threshold is not limited to this, and it is needless to say that the correction may be performed by one correction unit. In that case, the correction of the masking threshold is performed under the condition that the first correction for one initial masking threshold is allowed to lack a band. On the other hand, when the second correction instruction is received from the correction threshold value determination unit 143, it is performed under a condition that does not allow the loss of the band.

  Further, the auditory characteristic calculated in step S20 is not limited to the signal-to-mask ratio, and may be another characteristic.

  Furthermore, the encoding apparatus 1 is not limited to an apparatus that only encodes an audio signal as shown in FIG. 1, and may be an apparatus that encodes a video signal. An apparatus for encoding a video signal has a configuration for encoding a moving image in addition to the configuration shown in FIG. In such an apparatus, encoding of an input video signal is divided into moving image encoding and audio encoding, and then the encoded moving image and audio are multiplexed.

[Second Embodiment]
FIG. 6 is a functional block diagram of an encoding apparatus according to the second embodiment of the present invention. FIG. 7 is a block diagram illustrating a configuration example of the correction threshold value determination unit in the encoding device according to the second embodiment.

  As illustrated in FIG. 6, the encoding device 1 according to the present embodiment includes a block switching unit 10, an MDCT processing unit 11, a masking threshold generation unit 12, an auditory characteristic calculation unit 13, and a masking threshold correction unit 14. A quantization unit 15, an encoding unit 16, and a multiplexing unit 17. Further, the masking threshold correction unit 14 in the encoding device 1 according to the present embodiment includes a first correction unit 141, a second correction unit 142, and a correction threshold determination unit 143.

The functions of these components in the present embodiment are as described in the first embodiment. However, unlike the first embodiment, the correction threshold value determination unit 143 includes a masking threshold value (first correction threshold value) sfbThr ₁ (sfb) corrected by the first correction unit 141, and a second correction unit 142. And the masking threshold value (second correction threshold value) sfbThr ₂ (sfb) corrected in step. Then, the correction threshold value determination unit 143 determines which one of the first correction threshold value sfbThr ₁ (sfb) and the second correction threshold value sfbThr ₂ (sfb) is used for quantization, and then uses it for quantization. Only the correction threshold value is passed to the quantization unit 15.

As illustrated in FIG. 7, the correction threshold value determination unit 143 according to the present embodiment includes a missing SFB fluctuation rate calculation unit 143a, a missing SFB information holding unit 143b, and a determination unit 143c. Among these, the missing SFB fluctuation rate calculation unit 143a and the missing SFB information holding unit 143b are as described in the first embodiment. On the other hand, in addition to the first correction threshold sfbThr ₁ (sfb) corrected by the first correction unit 141 and the missing SFB fluctuation rate Vsfb calculated by the missing SFB fluctuation rate calculation unit 143a, the determination unit 143c performs the second correction. The second correction threshold value sfbThr ₂ (sfb) corrected by the unit 142 is input. When determining that the first correction threshold value sfbThr ₁ (sfb) is to be adopted based on the missing SFB fluctuation rate Vsfb and the threshold value TH ₁ , the determination unit 143 c passes the first correction threshold value to the quantization unit 15. On the other hand, when it is determined that the first correction threshold value sfbThr ₁ (sfb) is not adopted, the second correction threshold value sfbThr ₂ (sfb) is passed to the quantization unit 15.

In the encoding process using the encoding apparatus 1 of the present embodiment, first, the processes in steps S10 to S16 and the determination in step S18 shown in FIG. 3A are performed. When the initial PE value is equal to or less than the target PE value (step S18; No), as shown in FIG. 3C, the frequency spectrum is quantized using the initial masking threshold sfbThr ₀ (sfb) (step S30).

On the other hand, when the initial PE value is larger than the target PE value (step S18; Yes), the encoding apparatus 1 according to the present embodiment subsequently performs the masking threshold value correction process shown in FIG. When the masking threshold value correction process is performed, the masking threshold value generation unit 12 passes the initial masking threshold value sfbThr ₀ (sfb) to the auditory characteristic calculation unit 13. In addition, the initial masking threshold value generation unit 12 passes the initial masking threshold value and the target PE value to the first correction unit 141 and the second correction unit 142 of the masking threshold value correction unit 14, respectively.

  FIG. 8 is a flowchart illustrating a procedure of masking threshold correction processing in the encoding processing according to the second embodiment.

  In the masking threshold correction process, as shown in FIG. 8, first, auditory characteristics are calculated based on a frequency spectrum or the like (step S20). Step S20 is performed by the auditory characteristic calculation unit 13. The auditory characteristic calculation unit 13 calculates a signal-to-mask ratio as the auditory characteristic and passes it to the first correction unit 141 and the second correction unit 142.

Next, the masking threshold correction unit 14 of the encoding device 1 performs the masking threshold correction by the first correction unit 141 and the masking threshold correction by the second correction unit 142 in parallel. When the first correction unit 141 receives the auditory characteristic (signal to mask ratio), the first correction unit 141 corrects the initial masking threshold under a condition that allows the loss of the band (step S22). In addition, when the correction of the masking threshold is completed, the first correction unit 141 passes the corrected masking threshold to the missing SFB fluctuation rate calculation unit 143a and the determination unit 143c of the correction threshold determination unit 143. Then, the missing SFB fluctuation rate calculation unit 143a calculates the missing SFB fluctuation rate Vsfb using the corrected masking threshold sfbThr ₁ (sfb) (step S24). The missing SFB fluctuation rate calculation unit 143a calculates the missing SFB fluctuation rate Vsfb by the method described in the first embodiment. Then, the missing SFB fluctuation rate calculation unit 143a passes the calculated missing SFB fluctuation rate Vsfb to the determination unit 143c.

  On the other hand, when the second correction unit 142 receives the auditory characteristic (signal-to-mask ratio), the second correction unit 142 corrects the initial masking threshold under a condition that allows loss of the band (step S28). In addition, when the correction of the masking threshold is completed, the second correction unit 142 passes the corrected masking threshold to the determination unit 143c of the correction threshold determination unit 143.

Determining unit 143c, the first correction threshold sfbThr ₁ (sfb), a second correction threshold sfbThr ₂ (sfb), and receives the missing SFB volatility Vsfb, comparing the missing SFB volatility Vsfb with the threshold TH ₁ (Step S26). If Vsfb> TH ₁ (step S26; Yes), the determination unit 143c determines that the first correction threshold value sfbThr ₁ (sfb) is adopted, and passes the first correction threshold value to the quantization unit 15. Thereby, the masking threshold value used for quantization is determined to be the first correction threshold value sfbThr ₁ (sfb). Therefore, as illustrated in FIG. 3C, the quantization unit 15 quantizes the frequency spectrum using the first correction threshold, that is, the masking threshold corrected under the condition that allows the loss of the band (step S <b> 32).

On the other hand, when Vsfb ≦ TH ₁ (step S26; No), the determination unit 143c determines that the second correction threshold value sfbThr ₂ (sfb) is used for quantization, and the second correction threshold value is input to the quantization unit 15. hand over. Thereby, the masking threshold value used for quantization is determined to be the second correction threshold value sfbThr ₂ (sfb). Therefore, as illustrated in FIG. 3C, the quantization unit 15 quantizes the frequency spectrum using the second correction threshold, that is, the masking threshold corrected under a condition that does not allow band loss (step S <b> 34).

When the frequency spectrum using any one of the initial masking threshold value sfbThr ₀ (sfb), the first correction threshold value sfbThr ₁ (sfb), and the second correction threshold value sfbThr ₂ (sfb) is thus quantized, the quantization unit 15 Then, the quantized value is passed to the encoding unit 16. Then, the encoding unit 16 encodes the quantized value using a known encoding method such as fixed Huffman encoding as shown in FIG. 3C (step S36). When the encoding is completed, the encoding unit 16 passes the encoded data to the multiplexing unit 17. Thereby, the encoding process for one frame of the input signal (audio signal) is completed.

  When the encoding process is completed, the encoding device 1 (multiplexer 17) generates and outputs an encoded stream in which header information or the like is added to the encoded audio signal (audio data).

  As described above, also in the encoding process in the encoding apparatus 1 of the present embodiment, when correcting the masking threshold used for quantization, masking corrected under any of the conditions allowing or not allowing band loss is performed. Decide whether to use the threshold for quantization. At this time, as in the first embodiment, the encoding device 1 uses the fluctuation rate (time continuity) of the band sfb that is lost when the masking threshold corrected under the condition that allows band loss is used for quantization. Based on this, it is determined whether or not to allow a loss of bandwidth. Therefore, it is possible to perform quantization using a masking threshold value that is less susceptible to deterioration in sound quality, between masking corrected by allowing band loss and corrected masking threshold not allowed depending on the characteristics of the audio signal to be encoded. it can.

In this embodiment, when the initial masking threshold value is corrected, the masking threshold value is corrected under the condition that the first correction unit 141 allows the loss and the second correction unit 142 does not allow the deletion. The correction of the threshold value is performed in parallel. Therefore, when it is determined that the masking threshold corrected by the second correction unit 142 is used for quantization as a result of comparison between the missing SFB variation rate Vsfb and the threshold TH ₁ (step S26), the quantization is performed quickly. Can do. That is, according to the present embodiment, it is possible to suppress a delay in the encoding process when the masking threshold corrected by the first correction unit 141 does not satisfy the adoption condition. Therefore, it can be said that the encoding apparatus and the encoding method according to the present embodiment are suitable for suppressing deterioration in sound quality in a low delay encoding method such as AAC-Enhanced Low Delay (AAC-ELD).

  Note that the encoding apparatus 1 of the present embodiment can be realized by a computer and a program executed by the computer, as in the first embodiment. At this time, the computer that operates as the encoding apparatus 1 may have a hardware configuration as shown in FIG. Moreover, the program should just be the content which makes a computer perform the encoding process shown to FIG. 3A, FIG. 8, and FIG. 3C.

[Third Embodiment]
FIG. 9 is a functional block diagram of an encoding apparatus according to the third embodiment of the present invention. FIG. 10 is a block diagram illustrating a configuration example of the correction threshold value determination unit in the encoding device according to the third embodiment.

  As illustrated in FIG. 9, the encoding device 1 according to the present embodiment includes a block switching unit 10, an MDCT processing unit 11, a masking threshold generation unit 12, an auditory characteristic calculation unit 13, and a masking threshold correction unit 14. A quantization unit 15, an encoding unit 16, and a multiplexing unit 17. In addition, the encoding device 1 includes a storage unit 18. Furthermore, the masking threshold value correction unit 14 in the encoding device 1 according to the present embodiment includes a first correction unit 141, a second correction unit 142, and a correction threshold value determination unit 144.

  In the encoding device 1 of the present embodiment, the functions of the respective units other than the correction threshold value determination unit 144 are as described in the first embodiment.

  As illustrated in FIG. 10, the correction threshold value determination unit 144 according to the present embodiment includes a quantization unit 144a, a missing MDCT fluctuation rate calculation unit 144b, a missing MDCT information holding unit 144c, and a determination unit 144d.

The quantization unit 144a quantizes the frequency spectrum using the bit rate condition, the signal-to-mask ratio, and the masking threshold value sfbThr ₁ (sfb) corrected by the first correction unit 141.

  The missing MDCT fluctuation rate calculation unit 144b calculates the missing MDCT fluctuation rate Vmdct based on the quantization result in the quantization unit 144a and the information held by the missing MDCT information holding unit 144c. The missing MDCT fluctuation rate Vmdct is a value obtained by calculating the fluctuation rate of the band missing due to quantization in units of MDCT coefficients. This missing MDCT fluctuation rate Vmdct represents a ratio of a band that is continuously missing over several frames in a band that is missing due to quantization with respect to a current encoding target frame (T frame). That is, the missing MDCT fluctuation rate Vmdct is calculated by the same calculation method as the missing SFB fluctuation rate Vsfb in the first embodiment.

  The missing MDCT information holding unit 144c holds missing MDCT information for several frames used for calculating the missing MDCT fluctuation rate Vmdct. Missing MDCT information is a set of bands missing due to quantization in each frame. Note that the missing band in the missing MDCT information is the band of the MDCT coefficient unit after quantization. In the present embodiment, the missing MDCT information holding unit 144c, as shown in FIG. 10, includes missing MDCT information of T-1 frame (one frame before T frame) and T-2 frame (two frames before T frame). Of missing MDCT information.

The determination unit 144d determines whether or not to adopt the _first correction threshold value sfbThr ₁ (sfb) as a masking threshold value for quantization based on the calculated missing MDCT fluctuation rate Vmdct. Determining unit 144d, when the missing MDCT variation rate Vmdct is larger than the threshold TH ₂ a predetermined, to adopt a first correction threshold sfbThr masking threshold when quantizing ₁ (sfb) determines. That is, in this case, the determination unit 144d determines to use the first correction threshold value for quantization. In this case, the determination unit 144d passes the first correction threshold value to the quantization unit 15.

On the other hand, if the missing MDCT variation rate Vmdct the threshold TH ₂ or less, the determination unit 144d does not adopt the first correction threshold sfbThr ₁ (sfb), used in quantization i.e. second correction threshold sfbThr ₂ (sfb) And decide. In this case, the determination unit 144d causes the second correction unit 142 to correct the initial masking threshold value sfbThr ₀ (sfb).

  As described above, in the encoding device 1 according to the present embodiment, the frequency spectrum is quantized using the first correction threshold value corrected under a condition that allows band loss, and the result satisfies a predetermined condition. , The first correction threshold is determined to be used for quantization. On the other hand, if the result of quantization using the first correction threshold does not satisfy the predetermined condition, it is determined that the second correction threshold corrected under the condition that does not allow band loss is used for quantization. That is, in the encoding device 1 according to the present embodiment, either the first correction threshold or the second correction threshold is used for quantization based on the value of the frequency spectrum quantized using the first correction threshold. Decide whether or not.

In the encoding process using the encoding apparatus 1 of the present embodiment, first, the processes in steps S10 to S16 and the determination in step S18 shown in FIG. 3A are performed. When the initial PE value is equal to or less than the target PE value (step S18; No), as shown in FIG. 3C, the frequency spectrum is quantized using the initial masking threshold sfbThr ₀ (sfb) (step S30).

  On the other hand, when the initial PE value is larger than the target PE value (step S18; Yes), the encoding apparatus 1 of the present embodiment continues to perform the masking threshold value correction process shown in FIG. When performing the masking threshold correction process, the initial masking threshold generation unit 12 passes the initial masking threshold to the auditory characteristic calculation unit 13. Further, the initial masking threshold value generation unit 12 passes the initial masking threshold value and the target PE value to the first correction unit 141 of the masking threshold value correction unit 14 and causes the storage unit 18 to store the initial masking threshold value and the target PE value.

  FIG. 11 is a flowchart illustrating a procedure of masking threshold correction processing in the encoding processing according to the third embodiment.

  In the masking threshold correction process, as shown in FIG. 11, auditory characteristics are first calculated based on a frequency spectrum or the like (step S20). Step S20 is performed by the auditory characteristic calculation unit 13. The auditory characteristic calculation unit 13 calculates a signal-to-mask ratio as the auditory characteristic, and passes the calculated signal-to-mask ratio to the first correction unit 141 and stores it in the storage unit 18.

When the first correction unit 141 receives the auditory characteristic (signal to mask ratio), the first correction unit 141 corrects the initial masking threshold under a condition that allows the loss of the band (step S22). When completing the correction of the masking threshold, the first correction unit 141 passes the corrected masking threshold (first correction threshold) to the quantization unit 144a and the determination unit 144d of the correction threshold determination unit 144. Then, the quantization unit 144a of the correction threshold value determination unit 144 quantizes the frequency spectrum using the _first correction threshold value sfbThr ₁ (sfb) (step S23). When the quantization unit 144a finishes the quantization of the frequency spectrum, the quantization unit 144a passes the result to the missing MDCT fluctuation rate calculation unit 144b and the determination unit 144d.

When the missing MDCT fluctuation rate calculation unit 144b receives the quantization result, the missing MDCT fluctuation rate calculation unit 144b calculates the missing MDCT fluctuation rate Vmdct based on the result (step S25). The missing MDCT fluctuation rate calculation unit 144a is a frequency spectrum quantized using the _first correction threshold sfbThr ₁ (sfb) in the same manner as the method for calculating the missing SFB fluctuation rate Vsfb described in the first embodiment. The missing MDCT fluctuation rate Vmdct at is calculated. However, in step S25, the fluctuation rate is calculated by extracting a missing band not in SFB units but in quantized MDCT coefficient units. Then, the missing MDCT fluctuation rate calculation unit 144b passes the calculated missing MDCT fluctuation rate Vmdct to the determination unit 144d. In addition, when the missing MDCT fluctuation rate calculation unit 144b finishes calculating the missing MDCT fluctuation rate Vmdct, the missing MDCT fluctuation rate calculation unit 144b updates the information held by the missing MDCT information holding unit 144c. That is, the missing MDCT information of the T-1 frame is held as missing MDCT information of the T-2 frame, and the missing MDCT information of T frame is held as the missing MDCT information of the T-1 frame.

Note that, when the missing MDCT information of the previous frame and the previous frame is not held as in encoding the first few frames in the audio signal, the missing MDCT fluctuation rate calculation unit 144b sets a predetermined value. The missing MDCT fluctuation rate Vmdct. The predetermined value is, for example, to a value smaller than the threshold value TH ₂ to be compared with the missing MDCT variation rate Vmdct in the next step.

Missing MDCT variation rate determining section 144d for receiving the Vmdct compares the threshold value TH ₂ determined in advance and the missing MDCT volatility Vmdct (step S27). If Vmdct> TH ₂ (step S27; Yes), the determination unit 144d determines that the first correction threshold value sfbThr ₁ (sfb) is adopted as the masking threshold value for quantization, and sets the first correction threshold value. The data is passed to the quantization unit 15. Thereby, the masking threshold used for quantization is determined to be the first correction threshold sfbThr ₁ (sfb). In this case, as illustrated in FIG. 3C, the quantization unit 15 quantizes the frequency spectrum using the first correction threshold, that is, the masking threshold corrected under a condition that allows the loss of the band (step S32). On the other hand, if Vmdct ≦ TH ₂ (step S27; No), the determination unit 144d determines that the first correction threshold value sfbThr ₁ (sfb) is not adopted as the masking threshold value for quantization, and the second correction unit 142 causes the masking threshold to be corrected. Thereby, the masking threshold value used for quantization is determined to be the second correction threshold value sfbThr ₂ (sfb). The second correction unit 142 corrects the masking threshold under conditions that do not allow the loss of bands due to quantization based on auditory characteristics or the like (step S28), and the corrected masking threshold (second correction threshold) is quantized. To the conversion unit 15. In this case, as illustrated in FIG. 3C, the quantization unit 15 quantizes the frequency spectrum using the _second correction threshold value sfbThr ₂ (sfb) corrected under a condition that does not allow the loss of the band (step S34). .

When the frequency spectrum is quantized using one of the initial masking threshold value sfbThr ₀ (sfb), the first correction threshold value sfbThr ₁ (sfb), and the second correction threshold value sfbThr ₂ (sfb), the quantization unit 15 Then, the quantized value is passed to the encoding unit 16. Then, the encoding unit 16 encodes the quantized value using a known encoding method such as fixed Huffman encoding (step S36). When the encoding is completed, the encoding unit 16 passes the encoded data to the multiplexing unit 17. Thereby, the encoding process for one frame of the input signal (audio signal) is completed.

  When the encoding process is completed, the encoding device 1 (multiplexer 17) generates and outputs an encoded stream in which header information or the like is added to the encoded audio signal (audio data).

The missing MDCT variation rate Vmdct calculated in the present embodiment and the missing SFB variation rate Vsfb in the first embodiment differ only in whether the band is in the MDCT coefficient unit after quantization or in the SFB unit. It is the same. Therefore, also in the present embodiment, the missing MDCT variation rate Vmdct increases as the proportion of the band continuously missing in a plurality of frames increases. In other words, the smaller the proportion of bands that are continuously missing in a plurality of frames, the smaller the missing MDCT variation rate Vmdct. Therefore, if the missing MDCT variation rate Vmdct the threshold TH ₂ below, the audio signal to be encoded can be determined the peak position of the frequency spectrum is changing drastically. Therefore, in the encoding process of the present embodiment, if the missing MDCT variation rate Vmdct the threshold TH ₂ hereinafter be quantized using a second correction threshold sfbThr corrected under a condition that does not allow missing ₂ (sfb). As a result, it is possible to suppress deterioration in sound quality such that an original sound that can be heard continuously in an audio signal whose frequency spectrum peak position changes drastically is interrupted after encoding.

Also, the encoding process of the present embodiment, if the missing MDCT variation rate Vmdct is greater than the threshold value TH _2, quantized using a first correction threshold sfbThr corrected in conditions permitting missing ₁ (sfb). Therefore, it is possible to suppress deterioration in sound quality due to excessive correction of the masking threshold of a band important for sound quality when an audio signal such as a harpsichord in which the peak position of the frequency spectrum is gradually changed is encoded.

Further, in the present embodiment, based on the missing MDCT fluctuation rate Vmdct, it is determined which of the masking threshold corrected by the first correction unit 141 and the masking threshold corrected by the second correction unit 142 is used for quantization. . The missing SFB fluctuation rate Vsfb used in the first embodiment is a value calculated by dividing the MDCT coefficient in the frequency spectrum into a plurality of scale factor bands sfb and in units of scale factor bands sfb. On the other hand, as described above, the missing MDCT variation rate Vmdct of the present embodiment represents the variation rate of the missing band in the frequency spectrum quantized using the masking threshold corrected by the first correction unit 141 as described above. It is a value calculated in units. Therefore, according to the present embodiment, the fluctuation rate of the missing band can be calculated with a finer granularity than the missing SFB fluctuation rate Vsfb, and whether or not the first correction threshold value sfbThr ₁ (sfb) is adopted. The determination accuracy can be increased. Therefore, it is possible to further suppress deterioration in sound quality due to lack of bands.

  Note that the encoding apparatus 1 of the present embodiment can be realized by a computer and a program executed by the computer, as in the first embodiment. At this time, the computer that operates as the encoding apparatus 1 may have a hardware configuration as shown in FIG. Moreover, the program should just be the content which makes a computer perform the encoding process shown to FIG. 3A, FIG. 11, and FIG. 3C.

[Modification of Third Embodiment]
The correction threshold value determination unit 144 of the encoding device 1 according to the present embodiment is not limited to the configuration illustrated in FIG. 10 and may have a configuration illustrated in FIG.

  FIG. 12 is a block diagram illustrating a modification of the encoding device according to the third embodiment. FIG. 13 is a flowchart showing a part of the encoding process in the modified example of the third embodiment.

In the correction threshold value determination unit 144 illustrated in FIG. 12, the value of the frequency spectrum quantized by the quantization unit 144a is not the masking threshold value (first correction threshold value) corrected by the first correction unit 141, but the determination unit 144d. Is input. When the determination unit 144d determines that the first correction threshold sfbThr ₁ (sfb) is used for quantization based on the missing MDCT variation rate Vmdct by the determination in step S26 illustrated in FIG. 11, the quantization unit 144a The quantized frequency spectrum value is passed to the encoding unit 16. Therefore, when it is determined that the first correction threshold sfbThr ₁ (sfb) is used for quantization, the encoding apparatus 1 performs encoding using the quantized value (step S36), as shown in FIG. )I do.

On the other hand, when the determination in step S26 shown in FIG. 11 determines that the second correction threshold sfbThr ₂ (sfb) is used for quantization, the determination unit 144d causes the second correction unit 142 to correct the masking threshold ( Step S28).

As described above, the quantization unit 144a of the correction threshold value determination unit 144 performs the same processing as the quantization unit 15 provided outside, and quantizes the frequency spectrum. Therefore, if the bit rate condition, the initial masking threshold, and the frequency spectrum are the same, the result of quantization by the quantization unit 144a is the result of quantization by the quantization unit 15 provided outside the masking threshold correction unit 14. Be the same. Therefore, when the first correction threshold value sfbThr ₁ (sfb) is used for quantization, the value quantized by the quantization unit 144a of the correction threshold value determination unit 144 is encoded before the determination by the determination unit 144d. The process of step S32 shown in 3C can be omitted. Therefore, according to the modified example illustrated here, when the first correction threshold SfbThr ₁ a (sfb) was determined using the quantization, the first correction threshold sfbThr ₁ (sfb) Quantization Using after decision It is possible to suppress a decrease in processing efficiency due to performing again.

[Fourth Embodiment]
FIG. 14 is a functional block diagram of an encoding apparatus according to the fourth embodiment of the present invention. FIG. 15 is a block diagram illustrating a configuration example of a correction threshold value determination unit in the encoding device according to the fourth embodiment.

  As illustrated in FIG. 14, the encoding device 1 according to the present embodiment includes a block switching unit 10, an MDCT processing unit 11, a masking threshold generation unit 12, an auditory characteristic calculation unit 13, and a masking threshold correction unit 14. A quantization unit 15, an encoding unit 16, and a multiplexing unit 17.

In the encoding apparatus 1 according to the present embodiment, the functions of the respective units other than the masking threshold correction unit 14 are as described in the first embodiment. However, the quantization unit 15 is used only when the frequency spectrum is quantized using the initial masking threshold sfbThr ₀ (sfb).

  The masking threshold value correction unit 14 in the present embodiment includes a first correction unit 141, a second correction unit 142, a correction threshold value determination unit 144, a first quantization unit 145, and a second quantization unit 146. And comprising.

  The first correction unit 141 corrects the masking threshold under a condition that allows loss of a band due to quantization. In addition, the second correction unit 142 corrects the masking threshold under a condition that does not allow band loss due to quantization.

  The first quantization unit 145 quantizes the frequency spectrum by using the masking threshold (first correction threshold) corrected by the first correction unit 141. Further, the second quantization unit 146 quantizes the frequency spectrum using the masking threshold (second correction threshold) corrected by the second correction unit 142.

  Which one of the frequency spectrum value quantized by the first quantizing unit 145 and the frequency spectrum value quantized by the second quantizing unit 146 is used for encoding by the correction threshold value determining unit 144? To decide. As illustrated in FIG. 15, the correction threshold value determination unit 144 includes a missing MDCT fluctuation rate calculation unit 144b, a missing MDCT information holding unit 144c, and a determination unit 144d.

  The missing MDCT fluctuation rate calculation unit 144b and the missing MDCT information holding unit 144c are as described in the third embodiment.

  Based on the missing MDCT fluctuation rate Vmdct calculated by the missing MDCT fluctuation rate calculation unit 144b, the determination unit 144d uses the frequency spectrum value quantized by the first quantization unit 145 or the second quantization unit 146. The quantized frequency spectrum value is passed to the encoding unit 16.

  As described above, the encoding apparatus 1 according to the present embodiment performs quantization using the masking threshold corrected by the first correction unit 141 and quantization using the masking threshold corrected by the second correction unit 142. After performing the above, it is determined which quantization result is to be adopted. That is, the encoding apparatus 1 according to the present embodiment determines which quantization result to use, thereby determining the masking threshold corrected by the first correction unit 141 and the masking corrected by the second correction unit 142. It is indirectly determined which of the threshold value is used for quantization.

In the encoding process using the encoding apparatus 1 of the present embodiment, first, the processes in steps S10 to S16 and the determination in step S18 shown in FIG. 3A are performed. When the initial PE value is equal to or less than the target PE value (step S18; No), the frequency spectrum is quantized using the initial masking threshold sfbThr ₀ (sfb).

On the other hand, when the initial PE value is larger than the target PE value (step S18; Yes), the encoding apparatus 1 according to the present embodiment continues to perform various processes including the masking threshold correction processing illustrated in FIGS. 16A and 16B. Process. When the masking threshold value correction process is performed, the masking threshold value generation unit 12 passes the initial masking threshold value sfbThr ₀ (sfb) to the auditory characteristic calculation unit 13. In addition, the masking threshold value generation unit 12 passes the initial masking threshold value and the target PE value to each of the first correction unit 141 and the second correction unit 142 of the masking threshold value correction unit 14.

  FIG. 16A is a flowchart showing a part of the encoding process according to the fourth embodiment. FIG. 16B is a flowchart illustrating another part of the encoding process according to the fourth embodiment.

  When correcting the masking threshold in the encoding apparatus 1 of the present embodiment, as shown in FIG. 16A, first, auditory characteristics are calculated based on a frequency spectrum or the like (step S20). Step S20 is performed by the auditory characteristic calculation unit 13. The auditory characteristic calculation unit 13 calculates a signal-to-mask ratio smr (sfb) as the auditory characteristic and passes it to the first correction unit 141 and the second correction unit 142.

  Next, the masking threshold value correction unit 14 of the encoding device 1 corrects the masking threshold value by the first correction unit 141, the quantization by the first quantization unit 145, and the masking threshold value correction by the second correction unit 142. And the quantization by the 2nd quantization part 146 is performed in parallel.

When the first correction unit 141 receives the auditory characteristic (signal to mask ratio), the first correction unit 141 corrects the initial masking threshold under a condition that allows the loss of the band (step S22). After completing the correction of the masking threshold value, the first correction unit 141 passes the corrected first correction threshold value sfbThr ₁ (sfb) to the first quantization unit 145.

  The first quantizing unit 145 quantizes the frequency spectrum using the received masking threshold (first correction threshold) (step S23). After completing the quantization, the first quantizing unit 145 passes the quantized frequency spectrum value to the missing MDCT fluctuation rate calculating unit 144b and the determining unit 144d of the correction threshold value determining unit 144.

The missing MDCT fluctuation rate calculation unit 144b calculates a missing MDCT fluctuation rate Vmdct based on the result of quantization using the _first correction threshold sfbThr ₁ (sfb) (step S25). In step S25, as described above, the missing band in the quantization result at the present time (T frame) and the missing T-1 frame and T-2 frame held in the missing MDCT information holding unit 144b are lost. It is calculated by comparing with the band information. The missing MDCT fluctuation rate calculation unit 144b passes the calculated missing MDCT fluctuation rate Vmdct to the determination unit 144d.

  On the other hand, when the second correction unit 142 receives the auditory characteristic (signal-to-mask ratio), the second correction unit 142 corrects the initial masking threshold under a condition that does not allow the loss of the band (step S28). After completing the correction of the masking threshold value, the second correction unit 142 passes the corrected masking threshold value to the second quantization unit 146.

  The second quantizing unit 146 quantizes the frequency spectrum using the received masking threshold (step S34). When the second quantization unit 146 finishes the quantization, the second quantization unit 146 passes the quantized frequency spectrum value to the determination unit 144d of the correction threshold value determination unit 144.

Determining unit 144d, the first correction threshold quantization using the result, the result of the quantization using the second correction threshold, and receives the missing MDCT volatility Vmdct, missing MDCT volatility Vmdct and the threshold TH ₂ Are compared (step S27). When Vmdct> TH ₂ (step S27; Yes), the determination unit 144d determines that the first correction threshold value sfbThr ₁ (sfb) is adopted, and encodes the result quantized using the first correction threshold value. To the conversion unit 16. Thereby, the masking threshold value used for quantization is determined as the first correction threshold value sfbThr ₁ (sfb) afterwards. Further, the determination unit 144d passes the result obtained by quantizing the frequency spectrum using the first correction threshold value to the encoding unit 16. Therefore, as shown in FIG. 16B, after the determination in step S27, the quantized value is subsequently encoded (step S36).

On the other hand, when Vmdct ≦ TH ₂ (step S27; No), the determination unit 144d determines that the second correction threshold value sfbThr ₂ (sfb) is adopted for quantization, and performs quantization using the second correction threshold value. The result is passed to the encoding unit 16. As a result, the masking threshold value used for quantization is subsequently determined as the _second correction threshold value sfbThr ₂ (sfb). Further, the determination unit 144d passes the result of quantizing the frequency spectrum using the second correction threshold value to the encoding unit 16. Therefore, as illustrated in FIG. 16B, the encoding device 1 continues to encode the value quantized by the encoding unit 16 after the determination in step S26 (step S36).

  The encoding unit 16 encodes the quantized value using a known encoding method such as fixed Huffman encoding. When the encoding is completed, the encoding unit 16 passes the encoded data to the multiplexing unit 17. Thereby, the encoding process for one frame of the input signal (audio signal) is completed.

  When the encoding process is completed, the encoding device 1 (multiplexer 17) generates and outputs an encoded stream in which header information or the like is added to the encoded audio signal (audio data).

  Thus, in this embodiment, the correction of the masking threshold by the first correction unit 141 and the quantization by the first quantization unit 145, and the correction of the masking threshold by the second correction unit 142 and the second quantization are performed. The quantization by the unit 146 is performed in parallel. Therefore, after determining that the masking threshold corrected by the first correction unit 141 is not adopted, the encoding process using the frequency spectrum value quantized using the masking threshold corrected by the second correction unit 142 is performed. Can be started immediately. Therefore, according to the encoding device 1 and the encoding method of the present embodiment, it is possible to suppress a decrease in processing efficiency when quantization is performed using the masking threshold corrected by the second correction unit 142.

  In the present embodiment, the fluctuation rate Vmdct of the band that is missing in the band of the MDCT coefficient unit after quantization is calculated. Therefore, the fluctuation rate of the missing band can be calculated with a finer granularity than the case of calculating the fluctuation rate Vsfb of the missing band in SFB units, and the determination accuracy of whether or not to adopt the first correction threshold value Can be high. Therefore, it is possible to further suppress deterioration in sound quality due to lack of bands.

  Note that the encoding apparatus 1 of the present embodiment can be realized by a computer and a program executed by the computer, as in the first embodiment. At this time, the computer that operates as the encoding apparatus 1 may have a hardware configuration as shown in FIG. Moreover, the program should just be the content which makes a computer perform the encoding process shown to FIG. 3A, FIG. 16A, and FIG. 16B.

[Fifth Embodiment]
FIG. 17 is a functional block diagram of an encoding apparatus according to the fifth embodiment of the present invention. FIG. 18 is a block diagram illustrating a configuration example of a correction threshold value determination unit in the encoding device according to the fifth embodiment.

  As illustrated in FIG. 17, the encoding device 1 according to the present embodiment includes a block switching unit 10, an MDCT processing unit 11, a masking threshold generation unit 12, an auditory characteristic calculation unit 13, and a masking threshold correction unit 14. A quantization unit 15, an encoding unit 16, and a multiplexing unit 17. In addition, the encoding device 1 includes a correction threshold value determination unit 19.

  In the encoding device 1 of the present embodiment, the functions of the respective units other than the correction threshold value determination unit 19 are as described in the first embodiment. In the encoding device 1 shown in FIG. 17, the correction threshold value determination unit 19 is provided outside the masking threshold value correction unit 14. Therefore, the masking threshold correction unit 14 includes only the first correction unit 141 and the second correction unit 142 that correct the masking threshold.

  The correction threshold value determination unit 19 in the encoding device 1 of the present embodiment is different from the correction threshold value determination units 143 and 144 of the first to fourth embodiments, and includes a first correction unit 141 and a second correction unit 142. To determine which of the masking threshold values is to be corrected. Further, the correction threshold value determination unit 19 of the present embodiment determines a correction unit that corrects the masking threshold value based on the autocorrelation value of the power value mdct_pow (sfb) of the frequency spectrum.

  As shown in FIG. 18, the correction threshold value determination unit 19 of the present embodiment includes an autocorrelation value calculation unit 19a, a spectrum power value holding unit 19b, and a determination unit 19c.

  The autocorrelation value calculation unit 19a calculates the autocorrelation value corre of the power value (spectrum power value) mdct_pow (sfb) of the frequency spectrum. In the present embodiment, the autocorrelation value corre is calculated by the following equation (2).

In equation (2), corre (sfb) is an autocorrelation value of the spectrum power value mdct_pow (sfb) of the band sfb. Further, mdct_pow _T (sfb) is a spectrum power value of a current encoding target frame (T frame). Also, mdct_pow _T-1 (sfb) is the spectrum power value of the frame immediately before the current encoding target frame (T-1 frame). Further, mdct_pow _T-2 (sfb) is a spectrum power value of a frame (T-2 frame) two frames before the current encoding target frame.

  The spectrum power value holding unit 19b holds the spectrum power value of the T-1 frame and the spectrum power value of the T-2 frame used for calculating the autocorrelation value corre.

  The determination unit 19c determines which of the first correction unit 141 and the second correction unit 142 should correct the masking threshold based on the calculated autocorrelation value corre.

In the encoding process using the encoding apparatus 1 of the present embodiment, first, the processes in steps S10 to S16 and the determination in step S18 shown in FIG. 3A are performed. If the initial PE value is equal to or less than the target PE value (step S18; No), the frequency spectrum is quantized using the initial masking threshold sfbThr ₀ (sfb) (step S30).

  On the other hand, when the initial PE value is larger than the target PE value (step S18; Yes), the encoding device 1 of the present embodiment continues to perform various processes including the masking threshold correction processing shown in FIGS. 19A and 19B. Process. When the masking threshold value correction process is performed, the masking threshold value generation unit 12 passes the initial masking threshold value to the auditory characteristic calculation unit 13. In addition, the masking threshold value generation unit 12 passes the initial masking threshold value and the target PE value to each of the first correction unit 141 and the second correction unit 142 of the masking threshold value correction unit 14.

  FIG. 19A is a flowchart showing a part of the encoding process according to the fifth embodiment. FIG. 19B is a flowchart showing another part of the encoding process according to the fifth embodiment.

  When correcting the masking threshold in the encoding device 1 of the present embodiment, as shown in FIG. 19A, first, auditory characteristics are calculated based on a frequency spectrum or the like (step S20). Step S20 is performed by the auditory characteristic calculation unit 13. The auditory characteristic calculation unit 13 calculates a signal-to-mask ratio as the auditory characteristic and passes it to the first correction unit 141 and the second correction unit 142.

  When correcting the masking threshold, the encoding apparatus 1 determines which of the first correction unit 141 and the second correction unit 142 corrects the masking threshold in parallel with the calculation of the auditory characteristic. Processing is performed (steps S40 to S46).

  In the process of determining a correction unit that corrects the masking threshold, first, the autocorrelation value corre of the spectral power value mdct_pow (sfb) is calculated (step S40). Step S40 is performed by the autocorrelation value calculation unit 19a of the correction threshold value determination unit 19.

  The autocorrelation value calculation unit 19a reads the spectrum power value of the T-1 frame and the spectrum power value of the T-2 frame from the spectrum power value holding unit 19b, and calculates the autocorrelation value corre using Expression (2). The autocorrelation value calculation unit 19a passes the calculated autocorrelation value corre to the determination unit 19c. In addition, when the autocorrelation value calculation unit 19a finishes calculating the autocorrelation value corre, the autocorrelation value calculation unit 19a updates the power value held by the spectrum power value holding unit 19b. That is, the power value of the T-1 frame is held as the power value of the T-2 frame, and the power value of the T frame is held as the power value of the T-1 frame.

The determination unit 19c that has received the autocorrelation value corre determines the magnitude relationship between the autocorrelation value corre and a predetermined threshold value TH3 (step S42). If corre> TH ₃ (step S42; Yes), the determination unit 19c selects masking threshold correction by the first correction unit 141 (step S44). In this case, the determination unit 19c sends a signal instructing only the first correction unit 141 to correct the masking threshold.

On the other hand, when corre ≦ TH ₃ (step S42; No), the determination unit 19c selects the correction of the masking threshold by the second correction unit 142 (step S46). In this case, the determination unit 19c sends a signal instructing only the second correction unit 142 to correct the masking threshold.

  Thus, when a signal instructing correction of the masking threshold value is sent to the first correction unit 141 or the second correction unit 142, only the correction unit that receives the signal corrects the masking threshold value (step S48). . When the masking threshold is corrected by the first correction unit 141, step S48 is performed under a condition that allows band loss due to quantization. On the other hand, when the masking threshold value is corrected by the second correction unit 142, step S48 is performed under a condition that does not allow band loss due to quantization. Further, when the masking threshold value is corrected by either the first correction unit 141 or the second correction unit 142, for example, the correction value is calculated from the corrected masking threshold value and the power value using the above equation (1). The initial masking threshold is corrected so that the PE value becomes the target PE value.

  When the correction of the masking threshold is completed, the first correction unit 141 passes the corrected masking threshold (first correction threshold) to the quantization unit 15. Similarly, when the correction of the masking threshold is completed, the second correction unit 142 passes the corrected masking threshold (second correction threshold) to the quantization unit 15.

  In the encoding apparatus 1 of the present embodiment, when correcting the initial masking threshold for the current encoding target frame (frequency spectrum), only one of the first correction unit 141 and the second correction unit 142 is used. to correct. Therefore, when the masking threshold is corrected, the quantization unit 15 quantizes the frequency spectrum using the first or second correction threshold as shown in FIG. 19B (step S50). If the masking threshold is not corrected, the quantization unit 15 quantizes the frequency spectrum using the initial masking threshold (step S30).

When the frequency spectrum is quantized using one of the initial masking threshold value sfbThr ₀ (sfb), the first correction threshold value sfbThr ₁ (sfb), and the second correction threshold value sfbThr ₂ (sfb), the quantization unit 15 Then, the quantized value is passed to the encoding unit 16. Then, the encoding unit 16 encodes the quantized value using a known encoding method such as fixed Huffman encoding (step S36). When the encoding is completed, the encoding unit 16 passes the encoded data to the multiplexing unit 17. Thereby, the encoding process for one frame of the input signal (audio signal) is completed.

  When the encoding process is completed, the encoding device 1 (multiplexer 17) generates and outputs an encoded stream in which header information or the like is added to the encoded audio signal (audio data).

The autocorrelation value corre calculated from the equation (2) becomes larger as the correlation between the power values between frames is higher, in other words, as the difference between the power values between frames is smaller. Therefore, when the change in the peak position of the frequency spectrum in a plurality of consecutive frames is severe, the autocorrelation value corre becomes small. In the present embodiment, as in each of the embodiments described above, when the change in the peak position of the frequency spectrum is severe, the frequency spectrum is quantized with a masking threshold corrected under a condition that does not allow the loss of the band. Therefore, as described above, when the autocorrelation value corre is threshold TH ₃ or less (step S42; No), the determination unit 19c selects a correction of the masking threshold by the second correction unit 142 (step S46). As a result, when an audio signal in which a change in the peak position of the frequency spectrum is severe is encoded, it is possible to suppress deterioration in sound quality such that the sound that can be heard continuously is interrupted in the original sound.

  On the other hand, when the change in the peak position of the frequency spectrum in successive frames such as harpsichords is gentle, the autocorrelation value corre increases. Therefore, when the autocorrelation value corre is larger than the threshold (step S42; Yes), the determination unit 19c selects the correction of the masking threshold by the first correction unit 141 (step S44). As a result, when an audio signal with a gradual change in the peak position of the frequency spectrum is encoded, the masking threshold of a sound quality important band is excessively corrected and quantization error (quantization noise) can be perceived. It is possible to suppress the deterioration of the sound quality due to.

  In the encoding process of this embodiment, the masking threshold correction process may be performed only once by either the first correction unit 141 or the second correction unit 142. Moreover, which of the first correction unit 141 and the second correction unit 142 performs correction is selected based on the autocorrelation value corre of the spectral power value mdct_pow (sfb). Therefore, the amount of processing is reduced compared to the above-described embodiments in which the fluctuation rate of the band that is lost due to quantization using the first correction threshold is calculated, and the masking threshold used for quantization is determined based on the fluctuation rate. Can be reduced.

  Further, in the encoding process of the present embodiment, as described above, when correcting the masking threshold, the autocorrelation value corre is calculated and the power value held in the spectrum power value holding unit 19b is updated. However, the encoding process according to the present embodiment is not limited to this. Regardless of whether or not the masking threshold value is corrected, the power value held by the spectrum power value holding unit 19b every time step S10 in FIG. 3A is performed. It may be updated.

  Note that the encoding apparatus 1 of the present embodiment can be realized by a computer and a program executed by the computer, as in the first embodiment. At this time, the computer that operates as the encoding apparatus 1 may have a hardware configuration as shown in FIG. Moreover, the program should just be the content which makes a computer perform the encoding process shown to FIG. 3A, FIG. 19A, and FIG. 19B.

The following additional notes are further disclosed with respect to the embodiments including the examples described above.
(Appendix 1)
An encoding device that converts an audio signal into a frequency spectrum for each frame and quantizes and encodes the frequency spectrum,
A threshold generating unit that generates an initial masking threshold when quantizing the frequency spectrum based on the frequency spectrum;
A threshold correction unit that corrects the initial masking threshold based on whether or not to allow loss of a band due to quantization and the amount of bits given to quantization of the frequency spectrum;
Based on the degree of change in the frequency spectrum in a plurality of consecutive frames, the first masking threshold value corrected under conditions that allow band loss due to quantization, and correction under conditions that do not allow band loss due to quantization A threshold value determining unit that determines which of the second masking threshold values to be used for quantization;
An encoding device comprising:
(Appendix 2)
The threshold determination unit extracts a band that is missing when the frequency spectrum is quantized using the first masking threshold, and changes the missing band in a plurality of the frames as a degree of change in the frequency spectrum. Determining a rate and determining which of the first masking threshold and the second masking threshold to use for quantization based on the variation rate;
The encoding device according to appendix 1, wherein
(Appendix 3)
A quantization unit that quantizes the frequency spectrum using the first masking threshold;
The threshold value determination unit extracts a band missing due to quantization by the quantization unit, obtains a variation rate of the missing band in the plurality of frames as a degree of change of the frequency spectrum, and based on the variation rate Determining whether to use the first masking threshold for quantization;
The encoding device according to appendix 1, wherein
(Appendix 4)
An encoding unit for encoding the quantized frequency spectrum value;
When the threshold value determining unit determines to use the first masking threshold value for quantization, the encoding unit causes the encoding unit to encode the value of the frequency spectrum quantized by the quantization unit.
The encoding apparatus according to Supplementary Note 3, wherein
(Appendix 5)
The threshold determination unit calculates a fluctuation rate of the missing band in the plurality of frames;
A determination unit that determines whether to use the first masking threshold for quantization based on the calculated variation rate,
The encoding device according to any one of appendices 2 to 4, characterized in that:
(Appendix 6)
The calculation unit calculates the fluctuation rate of the missing band from the number of bands that are missing in the current encoding target frame and the number of bands that are missing in other frames among the missing bands.
The encoding apparatus according to supplementary note 5, wherein:
(Appendix 7)
The calculation unit, when the band missing in the current encoding target frame and the band before and after it are missing in other frames, coefficient the missing band as a band missing in other frames,
The encoding apparatus according to supplementary note 5, wherein:
(Appendix 8)
The threshold value determination unit calculates an autocorrelation value from the frequency spectrum in the plurality of frames as the degree of change in the frequency spectrum;
A determination unit that determines whether to correct the initial masking threshold based on the autocorrelation value based on a condition that allows a missing band due to quantization or a condition that does not allow a missing band due to quantization; ,
The encoding device according to appendix 1, wherein
(Appendix 9)
The threshold correction unit corrects the initial masking threshold under a condition that allows a missing band due to quantization, and a first corrector corrects the initial masking threshold under a condition that does not allow a missing band due to quantization. 2. The encoding device according to appendix 1, wherein the encoding device includes two correction units.
(Appendix 10)
The threshold generation unit generates the initial masking threshold based on an auditory psychological model.
The encoding device according to appendix 1, wherein
(Appendix 11)
Computer
When the initial masking threshold generated based on the frequency spectrum of the audio signal does not satisfy the condition for quantizing the frequency spectrum,
Based on the degree of change in the frequency spectrum in a plurality of consecutive frames of the audio signal, the first masking threshold corrected under the condition that allows band loss due to quantization and the band loss due to quantization are not allowed A process of determining which of the second masking threshold corrected by the condition is used for quantization;
The encoding method characterized by performing these.
(Appendix 12)
The computer is
After correcting the initial masking threshold to the first masking threshold,
Extracting a band that is missing when the frequency spectrum is quantized using the first masking threshold;
Calculating the rate of change of the missing band in the plurality of frames as the degree of change in the frequency spectrum;
Determining whether to use the first masking threshold for quantization based on the calculated variation rate;
The encoding method according to supplementary note 11, wherein
(Appendix 13)
The computer is
Correcting the initial masking threshold to the first masking threshold;
After quantizing the frequency spectrum with the first masking threshold,
Extracting a missing band in the quantized frequency spectrum;
Calculating a rate of change of the missing band in the plurality of frames as the degree of change in the frequency spectrum;
Determining whether to use the first masking threshold for quantization based on the calculated fluctuation rate correlation;
The encoding method according to supplementary note 11, wherein
(Appendix 14)
The computer is
If it is determined that the first masking threshold is used for quantization, the value of the frequency spectrum quantized using the first masking threshold before the determination is encoded.
The encoding method according to supplementary note 13, characterized by:
(Appendix 15)
The computer is
A process of correcting the initial masking threshold to the first masking threshold and a process of correcting the initial masking threshold to the second masking threshold are performed in parallel.
The encoding method according to any one of appendices 11 to 14, characterized in that:
(Appendix 16)
The computer is
Calculating an autocorrelation value of the frequency spectrum in a plurality of the frames;
Determining whether to correct the initial masking threshold to the first masking threshold or the second masking threshold based on the autocorrelation value;
The encoding method according to supplementary note 11, wherein
(Appendix 17)
Generate an initial masking threshold for quantizing the frequency spectrum of the audio signal;
If the generated initial masking threshold does not satisfy the condition for quantizing the frequency spectrum, the band loss due to quantization is based on the degree of change of the frequency spectrum in a plurality of consecutive frames of the audio signal. To determine which one of the first masking threshold value corrected under the condition that allows the quantization and the second masking threshold value corrected under the condition that does not allow the loss of the band due to the quantization are used for the quantization,
Quantizing the frequency spectrum using any of the initial masking threshold, the first masking threshold, and the second masking threshold;
A process of encoding the quantized frequency spectrum value;
A program that causes a computer to execute.

DESCRIPTION OF SYMBOLS 1 Encoding apparatus 10 Block switching part 11 MDCT process part 12 Masking threshold value generation part 13 Auditory characteristic calculation part 14 Masking threshold value correction part 141 1st correction part 142 2nd correction part 143 Correction threshold value determination part 143a Missing SFB fluctuation rate calculation Unit 143b missing SFB information holding unit 143c determining unit 144 correction threshold determining unit 144a quantizing unit 144b missing MDCT fluctuation rate calculating unit 144c missing MDCT information holding unit 144d determining unit 15 quantizing unit 16 encoding unit 17 multiplexing unit 18 storage unit 19 correction threshold value determination unit 19a autocorrelation value calculation unit 19b spectrum power value holding unit 19c determination unit

Claims (8)

An encoding device that converts an audio signal into a frequency spectrum for each frame and quantizes and encodes the frequency spectrum,
A threshold generating unit that generates an initial masking threshold when quantizing the frequency spectrum based on the frequency spectrum;
A threshold correction unit that corrects the initial masking threshold based on whether or not to allow loss of a band due to quantization and the amount of bits given to quantization of the frequency spectrum;
Based on the degree of change in the frequency spectrum in a plurality of consecutive frames, the first masking threshold value corrected under conditions that allow band loss due to quantization, and correction under conditions that do not allow band loss due to quantization A threshold value determining unit that determines which of the second masking threshold values to be used for quantization;
An encoding device comprising: The threshold determination unit extracts a band that is missing when the frequency spectrum is quantized using the first masking threshold, and changes the missing band in a plurality of the frames as a degree of change in the frequency spectrum. Determining a rate and determining which of the first masking threshold and the second masking threshold to use for quantization based on the variation rate;
The encoding apparatus according to claim 1. A quantization unit that quantizes the frequency spectrum using the first masking threshold;
The threshold value determination unit extracts a band missing due to quantization by the quantization unit, obtains a variation rate of the missing band in the plurality of frames as a degree of change of the frequency spectrum, and based on the variation rate Determining whether to use the first masking threshold for quantization;
The encoding apparatus according to claim 1. The threshold determination unit calculates a fluctuation rate of the missing band in the plurality of frames;
A determination unit that determines whether to use the first masking threshold for quantization based on the calculated variation rate,
The encoding apparatus according to claim 2 or 3, characterized in that The calculation unit calculates the fluctuation rate of the missing band from the number of bands that are missing in the current encoding target frame and the number of bands that are missing in other frames among the missing bands.
The encoding apparatus according to claim 4. The threshold value determination unit calculates an autocorrelation value from the frequency spectrum in the plurality of frames,
A determination unit that determines whether to correct the initial masking threshold based on the autocorrelation value based on a condition that allows a missing band due to quantization or a condition that does not allow a missing band due to quantization; ,
The encoding apparatus according to claim 1. Computer
When the initial masking threshold generated based on the frequency spectrum of the audio signal does not satisfy the condition for quantizing the frequency spectrum,
Based on the degree of change in the frequency spectrum in a plurality of consecutive frames of the audio signal, the first masking threshold corrected under the condition that allows band loss due to quantization and the band loss due to quantization are not allowed A process of determining which of the second masking threshold corrected by the condition is used for quantization;
The encoding method characterized by performing these. Generate an initial masking threshold for quantizing the frequency spectrum of the audio signal;
If the generated initial masking threshold does not satisfy the condition for quantizing the frequency spectrum, the band loss due to quantization is based on the degree of change of the frequency spectrum in a plurality of consecutive frames of the audio signal. To determine which one of the first masking threshold value corrected under the condition that allows the quantization and the second masking threshold value corrected under the condition that does not allow the loss of the band due to the quantization are used for the quantization,
Quantizing the frequency spectrum using any of the initial masking threshold, the first masking threshold, and the second masking threshold;
A process of encoding the quantized frequency spectrum value;
A program that causes a computer to execute.

Images