KR101430221B1 - Frequency band extending apparatus, frequency band extending method, player apparatus, playing method, program and recording medium - Google Patents

Abstract

A playback apparatus for playing back an input signal after band expansion of an input signal includes an extension controller for determining an extension start band of the input signal in accordance with the information related to the input signal and an extension controller for dividing the input signal into a plurality of sub- And a divider.

The frequency bands are divided by the band divider so that the input signal is band- And the signal is expanded based on a plurality of subband signals on the lower side of the extension start band among the plurality of subband signals.

Description

TECHNICAL FIELD [0001] The present invention relates to a frequency band expanding apparatus, a frequency band extending method, a reproducing apparatus, a reproducing method, a program,

The present invention relates to a frequency band enlarging device and a frequency band enlarging device, a reproducing device and a reproducing method, a program for causing a computer to execute reproduction processing, and a program for executing a reproducing process, which can reproduce encoded encoded data with higher quality after erasing a signal of a high frequency band And a recording medium having recorded thereon.

Recently, a music delivery service that provides encoded data such as MP3 (International Standard ISO / IEC 11172-3, MPEG Audio Layer 3) has been spreading. Most of these services distribute encoded data with reduced bit rate so that it does not take time to download the data.

Encoded data of a low bit rate is often encoded by deleting the components of a signal belonging to a high frequency band of about 15 kHz or more, which is difficult to hear in the human ear. As a result, the file size of the resultant data becomes small. However, if the high frequency band signal is deleted, there is a problem that the actual feeling provided by the original signal is lost and the sound becomes dull.

To cope with such a problem, in the encoding method such as HE-AAC (International Standard ISO / IEC 14496-3, High Efficiency MPEG4 AAC), a signal component belonging to a high frequency band of about 15 kHz or more Thereby reproducing a high-frequency component close to the original signal. In recent years, a high-frequency component close to the original signal is obtained by a post-processing band extension technique or the like which inputs a signal obtained by decoding the encoded data by deleting the component of the signal belonging to the high frequency band, Is reproduced.

For example, in Japanese Patent Application Laid-Open No. Hei. According to the technique disclosed in JP 2004-184472 (Patent Document 1), a high frequency band is generated by mixing an input signal and a local oscillation signal, and a high frequency band is generated by filtering by a passband characteristic corresponding to an encoding scheme or a genre of music The band is compensated by adding the high frequency component and the input signal. Japanese Patent Application Laid-open No. Hei. In the technique disclosed in JP 2002-175092 (Patent Document 2), an input signal is converted into a frequency domain by Fourier transformation to give a frequency band signal component, an envelope of a high frequency band is estimated from a frequency spectrum of a low frequency band , The gain of the frequency spectrum of the low frequency band is adjusted to follow the envelope.

However, in the technique disclosed in Patent Document 1, there is a limitation in the type of the pass band characteristic of the high pass filter to be preliminarily learned, and flexibility in the gain adjustment in the high frequency band can not be obtained. Further, in the technique disclosed in Patent Document 2, although the input signal is Fourier transformed to obtain the adjusted amplitude in the frequency domain, and the resultant signal is inverse Fourier transformed to obtain the time domain signal, the technique is based on the Fourier transform field There is a problem that time-domain aliasing occurs.

Further, in Japanese Patent No. 3538122 (Patent Document 3), this problem is avoided by using a band-dividing filter. 10 is a block diagram of a conventional playback apparatus proposed in Patent Document 3. [ In this technique, an input PCM (Pulse-Code Modulation) signal is decomposed into a plurality of subband signals in the band division unit 101. [ Then, the envelope estimating unit 102 estimates a frame-by-frame frequency envelope, and a high-frequency band generating unit 103 generates a subband signal belonging to a high-frequency band. Finally, the band-extended subband signal is supplied to the band synthesizer 104, and the band-extended PCM signal is output.

However, the technique proposed in Patent Document 3 causes three problems. First, it is a problem that the generation of the high-frequency band signal is not performed in accordance with the temporal fluctuation of the input signal within one frame as a result of the processing performed on a constant number of frame basis. Secondly, when an extremely large signal is input, there is a possibility that the signal of the high frequency band corresponding to it is also calculated very large, and the output of the band combining filter may overflow. Third, a signal obtained by decoding the encode data is input, and a band start frequency band for band extension can not be known in a post-processing band extension technique in which a high frequency band is supplemented.

SUMMARY OF THE INVENTION In view of the above problems, it is an object of the present invention to provide a frequency band enlarging device and a frequency band enlarging method, a reproducing device and reproducing method, a reproducing method and a reproducing method capable of deleting components of a signal belonging to a high frequency band and reproducing the encoded encoded data with higher quality. And a recording medium on which the program is recorded.

In one embodiment of the present invention, a frequency band extending apparatus for enlarging a frequency band of an input signal is provided. The frequency band extending apparatus includes expansion control means for determining an extension start band of the input signal and band dividing means for dividing the input signal into a plurality of subband signals, The frequency band is expanded based on a plurality of subband signals on the low-frequency side lower than the extension start frequency band among a plurality of subband signals divided into bands.

In another embodiment of the present invention, there is provided a frequency band enlarging method for enlarging a frequency band of an input signal. The frequency band extension method includes an extension control step of determining an extension start band for an input signal according to information related to an input signal, a band division step of dividing the input signal into a plurality of subband signals, When the input signal is band- And expanding the frequency band based on the plurality of subband signals on the low-frequency side lower than the expansion start frequency band among the plurality of subband signals.

In another embodiment of the present invention, there is provided a playback apparatus for playing back an input signal after bandwidth expansion of an input signal. The reproduction apparatus includes expansion control means for determining an extension start band for an input signal in accordance with information related to the input signal and band dividing means for dividing the input signal into a plurality of subband signals, The input signal is band divided And the frequency band is expanded based on a plurality of subband signals on the lower side of the extension start band among the plurality of subband signals.

In another embodiment of the present invention, a playback method is provided for playing back an input signal after bandwidth expansion of the input signal. This reproduction method includes an extension control step of determining an extension start band for an input signal in accordance with information related to the input signal and a band dividing step of dividing the input signal into a plurality of subband signals, The input signal is band divided And expanding the frequency band based on a plurality of subband signals on the lower side of the extended start band among the plurality of subband signals.

In another embodiment of the present invention, there is provided a program for causing a computer to execute processing for expanding an input signal and then reproducing an input signal. The program includes an extension control step of determining an extension start band of an input signal in accordance with information related to the input signal and a band dividing step of dividing the input signal into a plurality of subband signals, The input signal is band divided The frequency band expanding step of expanding the frequency band based on a plurality of subband signals on the lower side of the extended start band among the plurality of subband signals.

In another embodiment of the present invention, there is provided a recording medium on which a program for causing a computer to execute processing for expanding an input signal and then reproducing an input signal is provided. The program includes an extension control step of determining an extension start band for the input signal and a band dividing step of dividing the input signal into a plurality of subband signals, The frequency band expanding step of expanding the frequency band based on a plurality of subband signals on the lower side of the extended start band among the plurality of subband signals.

According to the embodiment of the present invention, the extension start band of the input signal can be determined, and the frequency band is expanded based on the plurality of subband signals on the lower side than the extension start band. Therefore, the input signal can be reproduced with high sound quality.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. This embodiment allows the input signal to be reproduced with high sound quality.

1 is a block diagram showing a configuration of a frequency band extending apparatus 10 according to the present embodiment. The frequency band extending apparatus 10 includes an extension control unit 11, a band dividing unit 12, a time dividing unit 13, a envelope estimating unit 14, a band compensating unit 15, a high frequency band generating unit 16, A phase adjusting unit 17, and a band synthesizing unit 18.

The extension control unit 11 is supplied with side information such as an encoding type, a sampling rate, a bit rate, etc. related to an input signal, determines an extension start frequency band based on the side information, (12). Optionally, the side information may be a preset value or an arbitrary user specified value depending on the type of encoding scheme of the input signal.

The band dividing section 12 divides the input signal into a plurality of subband signals. As a result, the band dividing section 12 supplies a plurality of subband signals (hereinafter, referred to as low-band subband signals) on the lower side of the generated start frequency band to the band combining section 18 do. In addition, the band dividing section 12 divides a plurality of subband signals (hereinafter, referred to as extended low band subband signals) closer to the extension start frequency band into a plurality of low frequency subband signals into a time division section 13 and a high frequency band generation (16).

The time division unit 13 performs temporal transient detection of the extended low-band subband signal, groups the extended low-band subband signals in the time direction, generates average sample power for each group of the extended low-band subband signals, And supplies the average sample power to the envelope estimating unit 14. [

The envelope estimating unit 14 obtains the group power per group from the sum of the average sample powers generated in the time classifying unit 13 and calculates the average value of the group power of the entire extended low band subband signal. Then, the envelope of the subband on the high-frequency side is estimated from the average value of the group power as the starting point, and the estimated envelope is supplied to the band compensating unit 15. [

The band compensating unit 15 calculates the gain adjustment value from the envelope of the high-frequency subband and the envelope of the low-frequency subband to the high-frequency subband signal and outputs the calculated gain adjustment value to the high- And supplies it to the band generating section 16. [

The high frequency band generating section 16 multiplies the gain adjustment value of the high frequency side subband signal by the extended low frequency subband signal to generate a high frequency side subband signal and supplies the generated high frequency side subband signal to the phase adjusting section 17).

The phase adjusting unit 17 changes the phase of the high-frequency side subband signal generated by the high frequency band generating unit 16 and supplies the subband signal whose phase is changed on the high frequency side to the band combining unit 18. [

The band combining section 18 performs band combining on the high-frequency side subband signal supplied from the phase adjusting section 17 and the low-frequency subband signal supplied from the band dividing section, and outputs a resultant band-widened signal.

By using the side information related to the input signal in this way, the expansion start frequency band for band extension can be determined with high accuracy. In addition, since the subband signal on the high-frequency side is generated based on the extended low-frequency subband signal close to the extension start frequency band, the frequency band can be expanded to a higher quality. In addition, overflow can be prevented by changing the phase of the generated high-frequency subband signal.

Hereinafter, each component of the above-described frequency band expanding apparatus will be described in detail.

The extension control unit

The extension control unit 11 determines the extension start frequency band based on the side information related to the input signal. This side information includes a type of encoding scheme, a sampling rate, a bit rate, and the like. Optionally, the side information may be a preset value or an arbitrary user specified value depending on the type of encoding scheme of the input signal.

Generally, the frequency band of the input signal correlates with various side information such as the type of encoding scheme, the sampling rate, and the bit rate. Therefore, in the present embodiment, the frequency band of the input signal is estimated using this side information, and the extended start frequency band sb for compensating the frequency band is determined. The determined extended start frequency band sb is supplied to the band division unit 12. [

2 is a diagram showing the relationship between the side information and the extended start frequency band sb. As shown in FIG. 2, when the frequency band belonging to the input signal is divided into 16 subbands, the extended start frequency band sb (sb is an arbitrary arbitrary range from 0 to 15) according to the encoding scheme, sampling rate and bit rate Is determined. For example, when the encoding method of the side information is B, the sampling rate is 44100 Hz, and the bit rate is 64-96 kbps, the extension start frequency band sb is determined to be 9. The elements for determining side information include stereo / mono, CBR / VBR, and the like.

treason Partition

The band dividing section 12 divides the input signal x (n) into 16 subband signals x (ib, n) (where ib = 0 to 15, (Sb-1) which is one subband of the extended start frequency band (hereinafter referred to as sb) among the 16 subband signals x (ib, n) The subband signals x (ib, n) belonging to the subbands from sb-4 to sb-1 are supplied to the band synthesizer 18, ) To the time division section 13 and the high frequency band generation section 16. [

Although the input signal x (n) is described as being divided into 16 subband signals x (ib, n) in this embodiment, the number of subbands in which the input signal is divided is not limited.

time Classification section

The time classifying section 13 classifies the signals as different groups and compensates for the high frequency band for each group every time it detects transitions in the time direction such as rising or falling of the sound. By this arrangement, , And sound quality deterioration can be prevented even for a natural sound signal having a difference in gain or frequency characteristics. On the other hand, in the technique proposed in Patent Document 3, frame processing is performed, and processing is performed on a frame-by-frame basis to supplement the high frequency band. That is, the sound signal of the natural system is not separated, but the sound signal is compensated without considering the time variation. This causes sound quality deterioration.

ㆍ Time division and power envelope calculation

(X (ib, n) on the low-band side from sb-4 to sb-1 supplied from the band dividing section 12 as an input. Each subband signal x (ib, n) is divided into 16, and it is a unit called a slot. The average sample power per sample (power (ib, islot)) is calculated for each slot, and the number of samples in each slot is set to eight.

ㆍ Grouping by transient detection

the average sample power power (ib, islot) is compared in the time direction (before and after the time axis) with respect to the entire 16 slots in the low-band sideband from sb-4 to sb-1, The transient detection is performed. The term transient detection refers to the detection of a location where the average sample power exhibits large variations in the time direction.

The ratio of the average sample power (power (ib, islot-1)) of the slot being searched and the slot before the slot is calculated. When the ratio is 16 times or more, (= 0.0625) times or less, it is judged as a fall, and the slot from the slot in which the transient was detected in the past is formed into one group from the slot next to the slot in which the current transient is detected.

It is assumed that rising and falling are detected in all the subbands on the low-band side from sb-4 to sb-1 when a rise or a fall is detected in any subband ib.

As a result, grouping is performed following the time variation, so that it is possible to generate a high-frequency band component that is closer to the natural sound signal, so that high-quality sound can be calculated.

In the present embodiment, the subband signals x (ib, n) on the low-band side from sb-4 to sb-1 supplied from the band dividing unit are divided into 16 in the time direction, However, the number of division in the time direction is not limited. Although one slot includes eight samples, the number of divisions in the time direction and the number of samples in one slot are not limited. When the ratio for transient detection is 16 times or more, it is judged as an increase. When the ratio is 1/16 (= 0.0625) times or less, it is judged as a fall. The threshold value for detection can be changed.

Determination of subbands for transient detection

In the grouping of the encode signals accompanied by the code degradation, the accuracy of the time variation depends on the deterioration state of the low-band side subband signal to be subjected to transient detection. Fig. 3 is a diagram showing the frequency amplitude characteristic when the code degradation is vigorous. As shown in Fig. 3, when the code degradation ("a") is intense, it is seen as a hole on the frequency axis, and the time classifying section 13 interprets the hole as the attenuation state of the signal, There is a problem of detecting an erroneous transient even in a place where it does not exist. As a result, the accuracy of the grouping is lowered, so that the sound quality is deteriorated, and the amount of calculation is increased by such transient detection.

As to these problems, in this embodiment, the maximum value of the average sample power is compared for each subband, and it is judged whether or not the subband is necessary for the transient detection. Then, actual transient detection is performed. In addition, in the case of a signal in which all the subbands are extremely small, no transient detection is performed in order to prevent an increase in the amount of calculation due to the time variation in the range that can not be heard.

4 is a flowchart showing a flow of processing for determining a target subband for transient detection.

In steps S41 to S43, the maximum value power (ib, islot) of the average 16-slot average sample power is searched for each of the four subband signals on the low-band side from sb-4 to sb-1 , And this maximum value is set to the representative value max power (ib) of the subband.

In step S44, the maximum value among the four representative values (max power (ib) (ib = sb-4, sb-3, sb-2, sb-1) obtained by the four subbands on the low- And the rest are set to the child subbands cb (0), cb (1), cb (2)). The representative value of the parent band is set as max power (pb) (step S45).

If it is determined in step S46 that the representative value max power (pb) of the parent subband is equal to or higher than -80 [dBFs] based on the 16-bit full scale reference, the process proceeds to step S47 .

On the other hand, when the representative value (max power (pb)) of the parent subbands is equal to a level lower than -80 [dBFs] based on the 16-bit full-scale criterion, the time Direction grouping is not performed. This means that there is no subband (step S48). As a result, the transient detection of the small signal is not performed, thereby preventing the unnecessary increase of the calculation amount.

If it is determined in step S47 that max power (ib) is equal to or larger than -80 [dBFs], and if any child subband cb (m) is set to the representative value max power (pb) of the parent subband pb (Max power (cb (m))) is less than 0.0015625 times, the process proceeds to step S49, and no transient detection is performed on this subband.

On the other hand, when the max power (ib) is equal to or larger than -80 [dBFs] and the representative value max (max) of any child subbands cb If power (cb (m)) is a value of 0.0015625 times or more, the process proceeds to step S50, and transient detection is performed on this subband. The subband for transient detection also includes the parent subband pb.

As a result, it is possible to reproduce a high quality sound by preventing the error detection of the time variation due to the code deterioration and reproducing the temporal envelope close to the natural acoustic signal. The average sample power power (ib, islot) of the four subbands ib on the low-band side from sb-4 to sb-1 generated in the time classifying section 13 is supplied to the envelope estimating section 14. [

In the present embodiment, the transient detection is not performed on subbands having a value for the transient detection smaller than 0.0015625 times, but the threshold value for transient detection is changed to 0.0015625 corresponding to the number of band divisions, It is also good.

Envelope Estimation part

The envelope estimating unit 14 first obtains the group power for each group from the sum of the average sample powers (powr (ib, islot)) generated in the time classifying unit 13 and calculates the group power from sb-4 to sb- The average value of the group power of the subband on the low-band side is calculated. Subsequently, the envelope of the subband in the high-frequency side from sb to 15 is estimated by the external insertion based on the primary straight line from the average value of the group power of the subband on the low-frequency side as a starting point. When the primary straight line of the envelope is represented by ax + b, the reference point (b) is obtained by calculating the envelope reference value by the weighted average to be described later, and the slope (a) is obtained by the envelope slope value a_lev described later.

ㆍ Calculation of group power

The envelope estimating unit 14 receives the average sample power power (ib, islot) of the four subbands ib on the low-band side from sb-4 to sb-1 supplied from the time classifying unit 13 . Calculates the sum of the average sample power power (ib, islot) as the number of slots nslot (ig) existing in each group in each subband ib for each group, and ig represents the current group , And sets the sum as a group power tpow (ig, ib) that is a maximum of 16 groups.

ㆍ Calculation of envelope reference value by weighted average

The average value of all the subband signals on the low-band side from sb-4 to sb-1 is obtained from the group power tpow (ig, ib) in each group obtained by the equation (4). Here, by allocating a larger weight to subbands close to sb using a weighted average to obtain an average value, the present embodiment can more smoothly connect the envelope on the low-band side to the envelope on the high-band side.

5 (a) and 5 (b) are schematic diagrams showing the envelope reference values based on different averaging methods. Here, as shown in Figs. 5 (a) and 5 (b), the group power tpow (sb-1, ig) of the subband sb- And the differences that result from different averaging methods when small.

Using an average value having a uniform weight is calculated later from the average value due to the influence of the magnitude of the group power tpow (ib, i g) of the remaining three subbands far from sb as shown in Fig. 5 (a) The reference point b is calculated as a large value. As a result, the subband sb-1 and the subband sb are not smoothly connected and sound quality deterioration occurs.

On the other hand, in the present embodiment, as shown in FIG. 5B, a larger weight is assigned to subbands close to sb, and the average is calculated, so that the frequency envelope can be smoothly connected.

6 is a flowchart showing the flow of processing for calculating the envelope reference value by the weighted average. In step S61 to S63, the group power tpow (ig, ib) of each of the four subbands on the low-band side from sb-4 to sb-1 is calculated. Then, in order to obtain the weighted average value w_avg (ig) at a ratio of 8: 4: 2: 1, for example, to the group power tpow (ig, ib) in order from the subband closest to sb Step S64) The weighted average is performed (Step S65).

Then, the group power of the subband sb is estimated using the weighted average value w_avg (ig) obtained from the four subband signals on the low-band side from sb-4 to sb-1. This value is the same as the reference value (b) and is called the envelope reference value (fenv (ig)). In the present embodiment, this value is determined by multiplying by the user-specified envelope reference adjustment value b_lev. That is, the envelope reference value is not determined uniquely, but the user can control the envelope reference adjustment function.

In the embodiment of the present invention, the envelope reference adjustment value b_lev is in the range of 0.25 to 1.0, and can be set by the user at will. In the embodiment of the present invention, the envelope reference adjustment value b_lev is set to 0.5 as a recommended value based on the frequency envelope obtained by statistically analyzing general music data. However, the range of the envelope reference adjustment value b_lev may be changed corresponding to the number of band divisions and the extension start frequency band sb.

ㆍ Limitation of envelope standard value

The envelope reference value fenv (ig) becomes an extremely large value according to the weighted average value w_avg (ig) or the extended strength e_lev, and the band synthesized output signal Is likely to overflow. In view of this situation, in the embodiment of the present invention, overflow of the output signal can be prevented by applying a limiter so as not to receive an extremely large value for the envelope reference value fenv (ig).

7 is a flowchart showing a flow of processing for limiting the envelope reference value. 8 (a) and 8 (b) are schematic diagrams showing how the envelope reference value is limited.

If it is determined in step S71 that the envelope reference value fenv (ig) is greater than the threshold value -6 [dBFs] (= 16384 ^ 2 * nslot (ig)), the process proceeds to step S72 Forcibly attenuates to the same level as the threshold value, as shown in Fig. 8 (b).

On the other hand, if it is determined in step S71 that the envelope reference value fenv (ig) is less than or equal to the threshold value -6 [dBFs] (= 16384 ^ 2 * nslot (ig)), the process proceeds to step S73 And the envelope reference value fenv (ig) is directly used as shown in Fig. 8 (a).

In the embodiment of the present invention, the threshold for limiting the envelope reference value fenv (ig) is set to -6 [dBFs]. Alternatively, the threshold value may be changed corresponding to the number of band divisions, the extension start frequency band sb, and the like.

ㆍ Determination of the envelope on the high side

The envelope (env (ib, ig) of the high-frequency side subband from sb to 15 is calculated by multiplying the envelope reference value fenv (ig) by the slope (slope) (a). The slope (a) is determined by the inclination value a_lev. In the embodiment of the present invention, the inclination is not uniquely determined, but a user-controllable inclination adjustment function is provided.

(a _- lev : inclined angle)

In the embodiment of the present invention, the inclination value a_lev is in the range of 0.25 to 1.0, and the user can set it arbitrarily. In the embodiment of the present invention, the envelope slope value a_lev is set to a recommended value of 0.5 based on the frequency envelope obtained by statistically analyzing general music data. The range of the envelope slope value a_lev is, The starting frequency band sb, or the like.

The envelope env (ib, iig) in the low-band side subband has the same meaning as the group power tpow (ib, iig), and the group power of the low-band extended band is set as the low- .

(Env (ib, iig) of the low-band side subband signal from sb-4 to sb-1 supplied from the time classifying section 13 obtained by the above-described processing and the envelope (Env (ib, ig)) of the subband of the subband is supplied to the band complementer 15.

treason Complementary part

The band compensating unit 15 adjusts the gain of the low-band sideband signal from sb-4 to sb-1 and compensates the high-band side subband signal of sb to 15. The mapping pattern of each pair of subbands is uniquely determined by sb.

The envelope env (ib, iig) for each of the high-side subbands from sb to 15 supplied from the envelope estimating unit 14 is divided and the sb- (Square root) of the division obtained by the envelope env (sb_map (ib), ig) for each of the subbands sb_map (ib) on the low- (gain (ib, ig)) of subbands on the high-frequency side from sb to 15 is obtained.

The gain adjustment coefficient gain (ib, ig) of the subbands on the high-frequency side from sb to 15 obtained by the equation (12) is supplied to the high-frequency band generation unit 16. [

The high-

The high frequency band generating section 16 receives the subband signals x (ib, n) on the low side from sb-4 to sb-1 supplied from the band dividing section 12, (Ib, ig) of the subbands on the high-frequency side from sb to 15 supplied from the adder 15 are input. (Ib (ib), n) on the low-band side from sb-4 to sb-1 serving as the source signal, , ig) to obtain the subband signal x (ib, n) on the high-frequency side from sb to 15.

Then, the subband signal x (i b, n) on the high-frequency side from sb to 15 obtained in the equation (13) is supplied to the phase adjusting section 17. [

The phase-

However, the subband signals x (ib, n) on the high-frequency side from sb to 15 supplied from the band complementer 15 are the four subband signals x (ib, n) on the low-band side from sb-4 to sb- (sb_map (ib), n). Therefore, the peak of the time-domain signal appears at the same timing in the subband signal on the low-frequency side and the subband signal on the high-frequency side. If all of the subbands are added together by band synthesis at a position having a peak at the same timing, an overflow of the band synthesis output signal may occur.

In view of the above, the phase adjusting section 17 changes the peaks of the subband signal on the low-frequency side and the subband signal on the high-frequency side and then inputs the subband signal to the band combining section 18 in order to prevent overflow.

9 is a schematic diagram showing the phase adjustment of the subband signal on the low-frequency side and the subband signal on the high-frequency side. Here, the subband signal x (ib, n) on the high-frequency side from sb to 15 is moved backward by four samples along the time axis. That is, in the present embodiment, the subband signal x (ib, n) is delayed in the time direction within a range that can not be audited by using the backward time masking characteristic observed in the human auditory system.

Here, although the delay is performed by four samples, the delay of four samples can be changed corresponding to the number of band divisions, the extension start frequency band (sb), the sampling frequency, and the like.

The phase adjusting unit 17 supplies the band combining unit 18 with the subband signals x (ib, n) on the high-range side from sb to 15 obtained by sample-shifting.

treason Synthesizing unit

Band synthesizer 18 multiplies the subband signals x (ib, n) on the high-frequency side from sb to 15 supplied from the phase adjusting section 17 and the subband signals x (ib, n) supplied from the band dividing section 12 Band synthesized output signal y (n) by band-narrowing the subband signal x (ib, n) on the low-band side of the filter bank.

As described above, sb is determined according to the side information in the embodiment of the present invention, and then the frequency band is expanded based on a plurality of subband signals on the lower side than sb. Therefore, a signal from which a signal component belonging to a high frequency band is deleted can be reproduced with high sound quality. Further, the code deterioration of the low-band side subband signal from sb-4 to sb-1 is detected, and the transient detection of the low band side subband signal from sb-4 to sb-1 is performed according to the deterioration detection result . Therefore, it is possible to prevent the amount of calculation of the transient detection from being increased. In addition, the subbands on the low-band side from sb-4 to sb-1 can be weighted close to the high-frequency side to average the frequency envelopes, so that the frequency envelopes on the low and high frequencies can be smoothly connected. In addition, over-flow of the band synthesized output signal can be prevented by applying the limiter to the envelope reference value calculated from the low-band side sub-band signals sb-4 to sb-1 to perform band synthesis. In addition, it is possible to change the phase of a plurality of subband signals belonging to subbands sb to 15 corresponding to a plurality of subband signals on the low side from 0 to sb-1, have.

It is needless to say that the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present invention. In the embodiment of the present invention, a frequency band expanding apparatus for processing a signal after a decoding process has been described as an example. Alternatively, the present invention is also applicable to a reproducing apparatus provided with decoding means. Although the hardware configuration has been described in the above embodiment, the present invention is not limited thereto. The present invention can also be realized by a CPU (Central Processing Unit) executing a computer program in an arbitrary process. In this case, the computer program may be provided to be recorded on a recording medium, and may be provided by transmission over the Internet or other transmission medium.

The present invention includes contents related to Japanese Patent Application (JP 2006-304501) filed on November 9, 2006 by the Japanese Patent Office.

1 is a block diagram showing a configuration of a frequency band extending apparatus according to an embodiment of the present invention.

2 is a diagram showing the relationship between the side information and the extension start frequency band ("sb").

Fig. 3 is a diagram showing the frequency amplitude characteristic when the code degradation is vigorous.

4 is a flowchart showing a flow of a process for determining a subband to be detected for a transient.

5 (a) and 5 (b) are schematic diagrams showing the envelope reference values based on different averaging techniques, respectively.

6 is a flowchart showing a flow of processing for calculating the envelope reference value by the weighted average.

7 is a flowchart showing a flow of processing for limiting the envelope reference value.

8 (a) and 8 (b) are schematic views showing that the envelope reference value is limited, respectively.

9 is a schematic diagram showing the phase adjustment of the subband signal on the low-frequency side and the subband signal on the high-frequency side.

10 is a block diagram showing a configuration of a frequency band extending apparatus of the prior art.

DESCRIPTION OF THE REFERENCE NUMERALS

10: Frequency band extending device 11: Extension control part

12: band dividing section 13: time dividing section

14: envelope estimation unit 15: band complementary unit

16: high frequency band generating section 17: phase adjusting section

18:

Claims (18)

The frequency band of the input signal is enlarged, An extension control means for determining an extension start band of the input signal, And band dividing means for dividing the input signal into a plurality of subband signals, The frequency band is expanded by the band dividing means on the basis of a plurality of subband signals on the low-frequency side lower than the extension start frequency band among a plurality of subband signals into which the input signal is band- Band synthesizing means for synthesizing a plurality of subband signals on the low-frequency side and a plurality of subband signals on the extension start frequency band that are lower than the extended start frequency band, Wherein said band synthesizing means synthesizes a plurality of subband signals on the lower side of said extended start band and a plurality of subband signals of said extended start frequency band by changing the phase. delete delete delete delete delete delete delete delete delete delete delete delete delete delete 1. A frequency band expanding apparatus for expanding a frequency band of an input signal, Extension control means for determining an extension start band of the input signal in accordance with information related to the input signal; Band dividing means for dividing the input signal into a plurality of subband signals, A transitional period in which a predetermined number of subband signals continuous to the extended start band are transiently detected in a time direction among a plurality of subband signals on a lower side than the extended start band in which the input signal is band- Transient detection) Group dividing means for dividing the predetermined number of subband signals into a plurality of groups in the time direction on the basis of the transient detection result in the transient detecting means; Power average value calculating means for calculating an average value of the powers of the predetermined number of subband signals based on an average value of group powers of the group in which the predetermined number of subband signals are divided by the group dividing means; Envelop estimation means for extrapolating an envelope straight line of a plurality of subband signals of the extension start frequency band or more starting from an average value calculated by the power average value calculation means; Band complementing means for supplementing a plurality of subband signals over the extension start band on the basis of the envelope straight line, And band synthesizing means for synthesizing a plurality of subband signals on the lower side of the extended start band and a plurality of subband signals on the extended start band or more, Wherein the frequency band is enlarged by the band dividing means based on a plurality of subband signals on the low-band side lower than the extension start band among a plurality of subband signals into which the input signal is band-divided. delete delete

Images