KR102243217B1 - Method and apparatus fo encoding audio signal - Google Patents

Abstract

According to an embodiment of the present invention, there is provided an audio signal encoding method and apparatus for improving sound quality of an encoded signal using bits remaining in a specific audio frame.
In the audio signal encoding apparatus according to an embodiment of the present invention, in performing bit allocation, when the total number of bits used for a specific audio frame is smaller than the maximum number of bits available per frame, that is, when bits remain, global A corrected quantized signal is output by adjusting a gain or restoring a signal masked by a masking threshold. The audio signal encoding apparatus according to an embodiment of the present invention may improve sound quality of the encoded audio signal by encoding the corrected quantized signal.

Description

Audio signal encoding method and apparatus {METHOD AND APPARATUS FO ENCODING AUDIO SIGNAL}

The present invention relates to a method and apparatus for encoding an audio signal. More specifically, it relates to an audio signal encoding method and apparatus for improving sound quality of an encoded signal using bits remaining in a specific frame.

In encoding an audio signal, a frame length, which is a basic unit of encoding, must be short to secure a short latency time, and a frame length must be long because sufficient frequency resolution is required to secure high sound quality. Therefore, it is difficult to satisfy the short delay time and high sound quality at the same time.

In the case of the prior art, in order to simultaneously satisfy the requirements for delay time and sound quality, a method of encoding an audio signal to have a delay time or sound quality within an allowable range is used by adjusting the length of the frame according to the intended application. do. Alternatively, a method of giving up the perfect reconstruction of the audio signal and using a specific type of window function is used.

Meanwhile, in the case of a perceptual audio coding method, an audio signal is quantized using a masking threshold derived from a psychoacoustic model, and bit allocation is performed for the quantized signal. By doing so, both delay time and sound quality requirements can be satisfied.

In the case of the perceptual sound encoding apparatus, the amount of quantization noise that the listener cannot perceive is determined according to an audio signal and a frequency band. In encoding an audio signal, the perceptual acoustic encoding apparatus determines a quantization step in consideration of the magnitude of the quantization noise. In addition, in performing bit allocation, the perceptual audio encoding apparatus does not allow the total number of bits used for a specific audio frame to exceed the maximum number of bits that can be used per frame. The maximum number of bits available per frame is determined by the output bit rate and is applied to all frames.

Meanwhile, when the perceptual sound encoding apparatus performs bit allocation, when the total number of bits used for a specific audio frame is less than the maximum number of bits available per frame determined by the output bit rate, an output bitstream There will be a bit left in it. Since information on the audio signal is not included in the remaining bits, when more information is encoded using the remaining bits, sound quality of the encoded audio signal can be improved. Accordingly, there is a need for a method of utilizing the remaining bits in encoding an audio signal of a specific frame.

An embodiment of the present invention provides an audio signal encoding method and apparatus for improving sound quality by using the remaining bits when there are remaining bits in encoding an audio signal of a specific frame.

An audio signal encoding method according to an embodiment of the present invention includes: converting an audio signal into a first frequency domain signal; Generating a first quantized signal by quantizing the first frequency domain signal by applying a global gain; Calculating a first number of used bits used by the first quantized signal; Generating a second quantized signal by correcting the first quantized signal when the number of first used bits is smaller than the number of pre-allocated frame bits for the frame of the audio signal; And encoding the second quantized signal.

In an audio signal encoding method according to an embodiment of the present invention, generating the second quantized signal includes: adjusting the global gain; And generating the second quantized signal by quantizing the first frequency domain signal by applying the adjusted global gain.

In the audio signal encoding method according to an embodiment of the present invention, generating the second quantized signal comprises reducing the global gain based on a difference between the first number of used bits and the number of frame bits. ; And generating the second quantized signal by quantizing the first frequency domain signal by applying the reduced global gain.

In an audio signal encoding method according to an embodiment of the present invention, the generating of the second quantized signal includes: reducing and updating the global gain by a predetermined value; Quantizing the first frequency domain signal by applying the updated global gain; Calculating a second number of used bits used by the quantized signal by applying the updated global gain; Repeating the steps of updating the global gain, quantizing by applying the updated global gain, and calculating the second number of used bits until the second number of used bits exceeds the number of frame bits. ; And generating the second quantized signal by quantizing the first frequency domain signal by applying a global gain updated in a previous iteration when the second number of used bits exceeds the number of frame bits.

In an audio signal encoding method according to an embodiment of the present invention, the generating of the second quantized signal includes: reducing and updating the global gain by a predetermined value; Quantizing the first frequency domain signal by applying the updated global gain; Calculating a second number of used bits used by the quantized signal by applying the updated global gain; And repeating the steps of updating the global gain, quantizing by applying the updated global gain, and calculating the second number of used bits until the updated global gain becomes less than or equal to a predetermined gain. Can include.

In an audio signal encoding method according to an embodiment of the present invention, when the number of first used bits is greater than the number of frame bits previously allocated for the frame of the audio signal, the difference between the number of first used bits and the number of frame bits On the basis of, increasing the global gain; And generating the second quantized signal by quantizing the first frequency domain signal by applying the increased global gain.

An audio signal encoding method according to an embodiment of the present invention includes: when the number of first used bits is greater than the number of pre-allocated frame bits for a frame of the audio signal, increasing and updating the global gain by a predetermined value; Quantizing the first frequency domain signal by applying the updated global gain; Calculating a second number of used bits used by the quantized signal by applying the updated global gain; And updating the global gain until the second number of used bits is less than or equal to the number of frame bits, quantizing by applying the updated global gain, and calculating the second number of used bits. It may further include a step of repeating.

In the audio signal encoding method according to an embodiment of the present invention, the generating of the first quantized signal comprises applying a masking threshold determined based on a psychoacoustic model, Masking at least one of the bands; And generating the first quantized signal by quantizing the masked first frequency domain signal, and generating the second quantized signal comprises: in the first quantized signal, the masked at least one band And generating the second quantized signal by adding a quantized signal for at least one of the bands.

In an audio signal encoding method according to an embodiment of the present invention, generating the second quantized signal comprises: comparing energy for each band masked by the masking threshold and a masking threshold for each band; Selecting at least one of the masked at least one band based on the comparison result; And generating the second quantized signal by adding a quantized signal for the selected at least one band to the first quantized signal.

In the audio signal encoding method according to an embodiment of the present invention, generating the second quantized signal comprises: when a difference between the first number of used bits and the number of frame bits is greater than or equal to a predetermined value, the first quantized signal In, it may include generating the second quantized signal by adding a quantized signal for the at least one masked band.

Meanwhile, an audio signal encoding apparatus according to an embodiment of the present invention includes: a frequency converter configured to convert an audio signal into a first frequency domain signal; A quantization unit that generates a first quantized signal by quantizing the first frequency domain signal by applying a global gain, a bit number calculation unit that calculates a first number of used bits used by the first quantized signal, and the first A bit number adjustment quantization unit including a correction unit for generating a second quantization signal by correcting the first quantized signal when the number of used bits is less than the number of frame bits pre-allocated for the frame of the audio signal; And an encoding unit encoding the second quantized signal.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit, when the number of first used bits is smaller than the number of frame bits, adjusts the global gain and applies the adjusted global gain to The second quantized signal may be generated by quantizing one frequency domain signal.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit, when the number of first used bits is smaller than the number of frame bits, based on a difference between the number of first used bits and the number of frame bits, The second quantized signal may be generated by reducing a global gain and quantizing the first frequency domain signal by applying the reduced global gain.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit, when the number of first used bits is smaller than the number of frame bits, decreases the global gain by a predetermined value and updates the updated global gain. An operation of quantizing the first frequency domain signal by applying a gain, and an operation of calculating a second number of used bits used by the quantized signal by applying the updated global gain, the second number of used bits in the frame Repeat until the number of bits is exceeded, and when the number of second used bits exceeds the number of frame bits, the second quantized signal is quantized by applying a global gain updated in the previous iteration to quantize the first frequency domain signal. Can be generated.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit includes an operation of reducing the global gain by a predetermined value and updating it, and an operation of quantizing the first frequency domain signal by applying the updated global gain. , And calculating the number of second used bits used by the quantized signal by applying the updated global gain may be repeated until the updated global gain becomes less than or equal to a predetermined gain.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit, when the number of first used bits is greater than the number of pre-allocated frame bits for the frame of the audio signal, the first number of used bits and the number of The second quantized signal may be generated by increasing the global gain based on the difference in the number of frame bits, and quantizing the first frequency domain signal by applying the increased global gain.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit increases the global gain by a predetermined value when the number of first used bits is greater than the number of pre-allocated frame bits for the frame of the audio signal. An operation of performing an update to perform an update, an operation of quantizing the first frequency domain signal by applying the updated global gain, and an operation of calculating a second number of used bits used by the quantized signal by applying the updated global gain. , It may be repeated until the second number of used bits is less than or equal to the number of frame bits. In the audio signal encoding apparatus according to an embodiment of the present invention, the quantization unit is determined based on the psychoacoustic model. By applying a masking threshold, masking at least one of a plurality of bands included in the first frequency domain signal, quantizing the masked first frequency domain signal to generate the first quantized signal, and the correction The unit may generate the second quantized signal by adding a quantized signal for at least one of the masked at least one band to the first quantized signal.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit compares energy for each band masked by the masking threshold and a masking threshold for each band, and the masked based on the comparison result. The second quantized signal may be generated by selecting at least one of the at least one band and adding a quantized signal for the selected at least one band to the first quantized signal.

In the audio signal encoding apparatus according to an embodiment of the present invention, the correction unit, when a difference between the first number of used bits and the number of frame bits is greater than or equal to a predetermined value, the at least one masked by the first quantized signal The second quantized signal may be generated by adding a quantized signal for a band.

Meanwhile, in a computer-readable recording medium recording a program for executing an audio signal encoding method in a computer according to an embodiment of the present invention, the method includes: converting an audio signal into a first frequency domain signal; Generating a first quantized signal by quantizing the first frequency domain signal by applying a global gain; Calculating a first number of used bits used by the first quantized signal; And generating a second quantized signal by correcting the first quantized signal when the number of first used bits is smaller than the number of pre-allocated frame bits for the frame of the audio signal. And encoding the second quantized signal.

1 is a block diagram illustrating an audio signal encoding apparatus to which the present invention can be applied.
2 is a block diagram illustrating an audio signal encoding apparatus according to an embodiment of the present invention.
3 is a flowchart illustrating a method of encoding an audio signal according to an embodiment of the present invention.
4 is a flowchart illustrating a step of correcting a quantized signal according to the first embodiment of the present invention.
5 is a flowchart illustrating a step of correcting a quantized signal according to a second embodiment of the present invention.
6 is a flowchart illustrating a step of correcting a quantized signal according to a third embodiment of the present invention.
7 is a flowchart illustrating a step of correcting a quantized signal according to a fourth embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily implement the present invention. However, the present invention may be implemented in various different forms and is not limited to the embodiments described herein. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" with another part, this includes not only "directly connected" but also "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, it means that other components may be further included rather than excluding other components unless specifically stated to the contrary.

In addition, in the present invention, the following terms may be interpreted according to the following criteria, and even terms not described may be interpreted according to the following purpose. Information is a term including all values, parameters, coefficients, elements, etc., and the meaning may be interpreted differently in some cases, and the present invention is not limited thereto. .

Meanwhile, an audio signal is a concept that is distinguished from a video signal in a broad sense, and may mean a signal that can be identified by hearing during reproduction. An audio signal is, by definition, a concept that is distinguished from a speech signal, and refers to a signal having no or little speech characteristics. The audio signal in the present invention should be interpreted in a broad sense, and when used separately from an audio signal, it can be understood as a narrow audio signal.

Meanwhile, a frame refers to a data unit for encoding or decoding an audio signal, and is not limited to a specific number of samples or a specific time.

The audio signal encoding method and apparatus according to the present invention may further be an audio signal processing apparatus and method to which the apparatus and method is applied.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram illustrating an audio signal encoding apparatus to which the present invention can be applied.

Referring to FIG. 1, an audio signal encoding apparatus to which the present invention can be applied includes a frequency converter 210, a quantization unit 120, an encoding unit 130, and a psychoacoustic model unit 140.

After receiving the input audio signal, the frequency converter 110 performs frequency conversion on the input audio signal to generate a frequency domain signal.

The psychoacoustic model unit 140 calculates a masking threshold by reflecting a person's auditory characteristics. The psychoacoustic model unit 140 calculates a masking threshold by applying a masking effect to the input audio signal.

The masking effect is based on psychoacoustic theory, and uses the characteristic that the human auditory structure does not recognize it well because small signals adjacent to a large signal are covered by a large signal. For example, in a noisy space, such as at a bus stop where a noisy bus passes, you will not hear the sound of conversation that would be heard in a quiet space.

The masking threshold may mean a threshold that a listener can hear. According to the masking effect, an audio signal located below the masking threshold cannot be heard by the listener.

The quantization unit 120 quantizes the frequency domain signal transformed by the frequency converter 110 by applying the masking threshold calculated by the psychoacoustic model 140. The quantization unit 120 allocates bits for the quantized signal.

For example, the quantization unit 120 may allocate a large number of bits to a frequency band in which noise is easily heard due to a low masking threshold, and may allocate a small number of bits to a frequency band with a high masking threshold. In addition, the quantization unit 120 may quantize only the remaining signals and perform bit allocation, except for a frequency band that cannot be heard by a user located below the masking threshold.

The encoder 130 outputs a bitstream through processes such as noiseless coding and bitstream packing on the quantized audio signal.

When the audio signal encoding apparatus shown in FIG. 1 performs bit allocation, the quantization unit 120 prevents the total number of bits used for a specific audio frame from exceeding the maximum number of bits that can be used per frame. The maximum number of bits available per frame is determined by the output bit rate and is applied to all frames.

At this time, when the total number of bits used for a specific audio frame is smaller than the maximum number of bits available per frame, there are bits remaining in the output bitstream. Since information on the audio signal is not included in the remaining bits, when more information is encoded using the remaining bits, sound quality of the encoded audio signal can be improved.

Accordingly, according to the method and apparatus for encoding an audio signal according to an embodiment of the present invention, sound quality can be improved by encoding more information on an audio signal using the remaining bits.

Hereinafter, an audio signal encoding apparatus according to an embodiment of the present invention will be described in detail with reference to FIG. 2.

2 is a block diagram illustrating an audio signal encoding apparatus according to an embodiment of the present invention.

Referring to FIG. 2, an apparatus 200 for encoding an audio signal according to an embodiment of the present invention includes a frequency converter 210, a quantization unit 220 for adjusting the number of bits, and an encoding unit 230. In addition, the audio signal encoding apparatus 200 according to an embodiment of the present invention may further include a psychoacoustic model unit 240.

The frequency conversion unit 210, the encoding unit 230, and the psychoacoustic model unit 240 of FIG. 2 are the frequency converter 110, the encoding unit 130, and the psychoacoustic model unit 140 of FIG. 1 Since it corresponds to, the overlapping description will be omitted.

The frequency converter 210 converts the input audio signal into a first frequency domain signal. The frequency transformation may use FFT (Fast Fourier Transform), MDCT (Modified Discrete Transform), wavelet packet transform (WPT), Frequency varying Modulated Lapped Transform (FV-MLT), and similar methods, but are not limited thereto. Does not.

The number of bits control quantization unit 220 quantizes the frequency domain signal converted by the frequency conversion unit 210, corrects the quantized signal, and outputs the quantized signal. The number of bits adjustment quantization unit 220 includes a quantization unit 222, a bit number calculation unit 224, and a correction unit 226.

The quantization unit 222 generates a first quantized signal by applying a global gain to quantize the first frequency domain signal. The global gain refers to a quantization scale factor value applied to the entire band included in the frequency domain signal in quantizing the frequency domain signal. The scale factor means a quantization step size.

The initial value of the global gain used by the audio signal encoding apparatus according to an embodiment of the present invention may be set by a user input or may be a predetermined value according to an application. The application may mean an application program used to encode an audio signal. The application may determine the initial value of the global gain with an experimentally optimized value in consideration of audio quality and the like.

Also, the quantization unit 222 may mask at least one of a plurality of bands included in the first frequency domain signal by applying a masking threshold determined by the psychoacoustic model unit 240. The quantization unit 222 may generate a first quantized signal by quantizing the first frequency domain signal masked by the masking threshold.

The psychoacoustic model unit 240 may determine a masking threshold by applying a masking effect to the input audio signal.

The masking threshold may mean a threshold that a listener can hear. According to the masking effect, an audio signal located below the masking threshold cannot be heard by the listener.

For example, in applying a psychoacoustic model, a signal with the largest energy exists in the middle in a plurality of frequency scale factor bands included in one window in which an audio signal is divided, and This will be described with reference to the case where there are several signals of much smaller size around. In this case, the largest signal becomes a masker, and a masking curve is drawn based on this masker. The small signal obscured by this masking curve can be a masked signal or a maskee. Excluding this masked signal, leaving only the remaining signals as valid signals is called masking.

The psychoacoustic model models the human auditory system using various algorithms. Various known psychoacoustic models can be used with embodiments of the present invention.

The quantization unit 222 may determine, for example, that a user cannot hear a frequency band whose energy is lower than the masking threshold, and mask a frequency band determined that the user cannot hear. That is, the quantization unit 222 may quantize only the remaining signals, except for a frequency band having an energy lower than the masking threshold, and perform bit allocation.

The number of bits calculation unit 224 calculates the number of first used bits used by the first quantized signal quantized by the quantization unit 222.

The correction unit 226 corrects the first quantized signal when the number of first used bits is smaller than the number of frame bits previously allocated for the frame of the audio signal. The correction unit 226 may correct the first quantized signal by adjusting a global gain or restoring a signal masked by a masking threshold when the number of first used bits is smaller than the number of frame bits, that is, bits remain. . The correction unit 226 may generate and output a second quantized signal by correcting the first quantized signal.

In addition, the correction unit 226 may generate a second quantized signal by correcting the first quantized signal even when the number of first used bits is greater than the number of pre-allocated frame bits for the frame of the audio signal. The correction unit 226 may correct the first quantized signal by adjusting the global gain when the first number of used bits is greater than the number of frame bits, that is, when the number of bits is insufficient.

Further, when the number of first used bits is the same as the number of frame bits, the correction unit 226 may output the first quantized signal as the second quantized signal without additional correction.

Related to a specific method of improving sound quality of an encoded audio signal by outputting a corrected quantized signal when the number of first used bits is smaller than the number of frame bits, the audio signal encoding apparatus 200 according to an embodiment of the present invention Therefore, it will be described in detail with reference to FIG. 3 below.

3 is a flowchart illustrating a method of encoding an audio signal according to an embodiment of the present invention.

Referring to FIG. 3, an audio signal encoding method according to an embodiment of the present invention includes steps processed by the audio signal encoding apparatus 200 illustrated in FIG. 2. Accordingly, it can be seen that the contents described above with respect to the audio signal encoding apparatus 200 illustrated in FIG. 2 are also applied to the audio signal encoding method of FIG. 3 even if omitted below.

In step S310, the audio signal encoding apparatus 200 according to an embodiment of the present invention converts the audio signal into a first frequency domain signal. The frequency transformation may use FFT (Fast Fourier Transform), MDCT (Modified Discrete Transform), wavelet packet transform (WPT), Frequency varying Modulated Lapped Transform (FV-MLT), and similar methods, but are not limited thereto. Does not.

In step S320, the audio signal encoding apparatus 200 generates a first quantized signal by quantizing the first frequency domain signal by applying a global gain. The audio signal encoding apparatus 200 uses a global gain as a quantization scale factor commonly used for all frequency bands of the first frequency domain signal. In addition, the audio signal encoding apparatus 200 may allocate and quantize necessary bits by adjusting a quantization scale factor for each band for each frequency band based on a global gain.

In step S330, the audio signal encoding apparatus 200 calculates the number of first used bits used by the first quantized signal.

In step S340, the audio signal encoding apparatus 200 determines whether the number of first used bits is smaller than the number of pre-allocated frame bits for the frame of the audio signal.

When the number of first used bits is not smaller than the number of frame bits, the audio signal encoding apparatus 200 may encode the first quantized signal quantized in step S320. On the other hand, when the number of first used bits is smaller than the number of frame bits, the audio signal encoding apparatus 200 performs step S350.

In step S350, the audio signal encoding apparatus 200 generates a second quantized signal by correcting the first quantized signal when the number of first used bits is less than the number of frame bits. The audio signal encoding apparatus 200 may output a corrected quantized signal by adjusting a global gain or restoring a signal masked by a masking threshold when the first number of used bits is smaller than the number of frame bits, that is, bits remain. I can.

According to the first embodiment of the present invention, the audio signal encoding apparatus 200 may generate a second quantized signal by adjusting the global gain and quantizing the first frequency domain signal by applying the adjusted global gain.

In this case, the audio signal encoding apparatus 200 reduces the global gain based on the difference between the number of first used bits and the number of frame bits, and quantizes the first frequency domain signal by applying the reduced global gain to the second It can generate a quantized signal.

In addition, the audio signal encoding apparatus 200 may update the global gain by decreasing it by a predetermined value, and quantize the first frequency domain signal by applying the updated global gain. For example, the audio signal encoding apparatus 200 may update a global gain, that is, a quantization step size applied to the entire band by decreasing by one. The audio signal encoding apparatus 200 may calculate the number of second used bits used by the quantized signal by applying the updated global gain. The audio signal encoding apparatus 200 may generate a second quantized signal by repeating the steps of updating the above-described global gain, quantizing by applying the updated global gain, and calculating the second number of used bits. .

When the global gain is reduced, the quantization error for the entire frequency band is reduced. That is, since the second quantized signal to which the reduced global gain is applied uses a larger number of bits than the first quantized signal to which the existing global gain is applied, the sound quality of the encoded audio signal is improved. In connection with the first embodiment of the present invention, it will be described in more detail with reference to FIG. 4 later.

Meanwhile, according to the second embodiment of the present invention, the audio signal encoding apparatus 200 may generate a second quantized signal by restoring a signal masked by a masking threshold.

The audio signal encoding apparatus 200 according to the second embodiment of the present invention quantizes the first frequency domain signal by applying a masking threshold determined based on the psychoacoustic model in quantizing the first frequency domain signal in step S320. can do. The audio signal encoding apparatus 200 according to the second embodiment may mask at least one of a plurality of bands included in the first frequency domain signal by using the masking threshold in step S320. The audio signal encoding apparatus 200 may generate a first quantized signal by quantizing the masked first frequency domain signal.

The audio signal encoding apparatus 200 may correct the first quantized signal by adding a quantized signal for at least one of the bands masked by the masking threshold to the first quantized signal. The audio signal encoding apparatus 200 may output the corrected first quantized signal as a second quantized signal.

In this case, the audio signal encoding apparatus 200 may compare energy for each frequency band included in the first frequency domain signal with a masking threshold, and generate a second quantized signal based on the comparison result. More specifically, the audio signal encoding apparatus 200 may compare energy for each band of the frequency bands masked by the masking threshold and a masking threshold for the corresponding band. The audio signal encoding apparatus 200 may select at least one band from among at least one masked band based on the comparison result.

For example, the audio signal encoding apparatus 200 may preferentially select a band having the smallest difference between energy for a frequency band and a masking threshold for the band. The audio signal encoding apparatus 200 may generate a second quantized signal by adding a quantized signal for at least one selected band to the first quantized signal. The second quantized signal includes a quantized signal for at least one frequency band selected from among the frequency bands masked by the masking threshold, thereby improving sound quality of an audio signal encoded using the remaining bits.

Also, the audio signal encoding apparatus 200 may generate a second quantized signal from which the masked signal is reconstructed based on a difference between the first number of used bits and the number of frame bits. When the difference between the first number of used bits and the number of frame bits is greater than or equal to a predetermined value, the audio signal encoding apparatus 200 adds a quantization signal for at least one band masked by a masking threshold to the first quantization signal, 2 Can generate quantized signals.

For example, the first quantized signal is a signal obtained by removing a signal value less than or equal to the masking threshold in a predetermined band, and the removed signal is assumed to be a third quantized signal. The second quantized signal may be a signal obtained by adding a signal for at least one band among the third quantized signals to the first quantized signal.

That is, when the number of bits equal to or greater than the predetermined number of bits remains, the audio signal encoding apparatus 200 may output a second quantized signal that is not masked by the masking threshold, instead of the first quantized signal masked by the masking threshold. .

The second quantized signal output according to the second embodiment of the present invention is encoded using a larger number of bits than the first quantized signal by further including a quantized signal for at least one band masked by a masking threshold. Accordingly, the audio signal encoding apparatus 200 according to the second embodiment of the present invention can improve the sound quality of the audio signal by correcting the quantized signal to use the remaining bits. In connection with the second embodiment of the present invention will be described in more detail with reference to FIG. 5 later.

In step S360, the audio signal encoding apparatus 200 encodes the second quantized signal. For example, the audio signal encoding apparatus 200 may output a bitstream through processes such as noiseless encoding and bitstream packing on the second quantized signal.

Meanwhile, although not shown in FIG. 3, when it is determined in step S340 that the number of first used bits is greater than the number of frame bits, the audio signal encoding apparatus 200 according to an embodiment of the present invention It may further include increasing the global gain based on the difference in the number of frame bits. The audio signal encoding apparatus 200 may generate a second quantized signal by quantizing the first frequency domain signal by applying the increased global gain.

In addition, when it is determined in step S340 that the number of first used bits is greater than the number of frame bits, the audio signal encoding apparatus 200 according to an exemplary embodiment of the present invention may increase the global gain by a predetermined value and update it. The audio signal encoding apparatus 200 may quantize the first frequency domain signal by applying the updated global gain. The audio signal encoding apparatus 200 may calculate the number of second used bits used by the quantized signal by applying the updated global gain. The audio signal encoding apparatus 200 includes: updating a global gain by a predetermined value until the number of second used bits is less than or equal to the number of frame bits, quantizing by applying the updated global gain, and the second The step of calculating the number of used bits can be repeated.

4 is a flowchart illustrating a step of correcting a quantized signal according to the first embodiment of the present invention.

According to the first embodiment of the present invention, the audio signal encoding apparatus 200 may generate a corrected quantized signal by adjusting the global gain and quantizing the first frequency domain signal by applying the adjusted global gain.

In step S410, the audio signal encoding apparatus 200 reduces and updates the global gain by a predetermined value. In this case, the predetermined value by which the global gain is reduced is a predetermined value and may be set by a user's input, may be predetermined according to an application, or may be a predetermined value according to the number of frame bits. For example, the audio signal encoding apparatus 200 may update the global gain by decreasing it by one. Alternatively, the predetermined value by which the global gain is reduced may be a value determined based on a difference between the first number of used bits and the number of frame bits. In step S420, the audio signal encoding apparatus 200 quantizes the first frequency domain signal by applying the updated global gain. In step S430, the audio signal encoding apparatus 200 calculates the number of second used bits used by the quantized signal by applying the updated global gain.

In step S440, the audio signal encoding apparatus 200 determines whether the number of second used bits is greater than the number of frame bits. That is, the audio signal encoding apparatus 200 quantizes the audio signal by applying the updated global gain to determine whether or not a bit to be allocated to the audio signal is insufficient.

In step S450, the audio signal encoding apparatus 200, when the second number of used bits is greater than the number of frame bits, that is, when the number of bits to be allocated becomes insufficient, the quantized first frequency domain by applying the global gain updated in the previous iteration. The signal can be output as a second quantized signal.

In step S460, the audio signal encoding apparatus 200 determines whether the updated global gain or the number of second used bits satisfies the repetition termination condition. When the updated global gain or the second number of used bits does not satisfy the repetition termination condition, the audio signal encoding apparatus 200 can output a second quantized signal having a suitable global gain and a suitable second number of used bits. Until, it is possible to repeat the preceding steps S410 to S450.

As an example, the repetition termination condition may include a case where the updated global gain becomes less than or equal to a predetermined gain. The predetermined gain may be a value input by a user, a value predetermined according to an application, or a value calculated according to a frame.

The audio signal encoding apparatus 200 may terminate the repetition so that the global gain is no longer reduced when the updated global gain becomes less than or equal to the predetermined gain. When the global gain is reduced, the quantization error for all frequency bands is reduced. However, as the global gain continues to decrease, the amount of computation of the audio signal encoding apparatus 200 increases. Accordingly, the audio signal encoding apparatus 200 may preset the minimum value of the updated global gain. The audio signal encoding apparatus 200 may preset an experimentally optimized value as the minimum value of the updated global gain.

When the updated global gain is equal to or less than a predetermined gain, the audio signal encoding apparatus 200 may apply the updated global gain and output the quantized signal as the second quantized signal.

As another example, the repetition end condition may include a case where the number of second used bits is the same as the number of frame bits. When the total number of bits used by the quantized signal by applying the updated global gain is the same as the maximum number of bits that can be used per frame, the audio signal encoding apparatus 200 uses all the allocated bits per frame. It can be judged that the signal has the best sound quality.

Accordingly, when the number of second used bits is equal to the number of frame bits, the audio signal encoding apparatus 200 may apply the updated global gain to output the quantized signal as the second quantized signal.

According to the first embodiment of the present invention, the audio signal encoding apparatus 200 may use a method of adjusting the global gain by repeating an operation of decreasing and updating the global gain by a predetermined value, as shown in FIG. 4. have. Meanwhile, the audio signal encoding apparatus 200 may further improve the global gain adjustment speed by using a method of reducing the global gain based on the number of remaining bits.

Based on the number of remaining bits, that is, the difference between the number of first used bits and the number of frame bits, the global gain to be adjusted may be calculated through the following method.

First, the audio signal encoding apparatus 200 according to an embodiment of the present invention, as the global gain increases or decreases by a predetermined value, increases or decreases the number of bits used by the quantized signal, that is, the entropy change rate. Can be estimated. For example, as the global gain, that is, the quantization step size applied to the entire band increases by 1 or decreases by 1, an increase or decrease rate of the number of bits used by the quantized signal may be estimated.

In estimating an entropy change rate according to a change in a global gain, the audio signal encoding apparatus 200 may estimate a rate of change of the number of bits per spectral data when the global gain changes by one. Accordingly, in order to estimate the change in the number of bits used as the global gain changes by one, the audio signal encoding apparatus 200 must consider the total number of frequency data, that is, the frame size.

Hereinafter, as the global gain increases by 1, -3/16 bits per frequency data decreases, and the case where the entropy change rate according to the change of the global gain is estimated will be described as an example. However, the present invention is not limited thereto.

If it is estimated that the rate of change of entropy according to the change of the global gain decreases as the global gain increases by 1, 3/16 bits decreases, and if the frame size is 128 bits, 3/16 * 128 = 24 bits decreases as the global gain decreases by 1. It can be seen that it is needed additionally.

For example, if the number of frame bits is 600 and the number of first used bits is 580, the difference between the number of first used bits and the number of frame bits is 20 bits. That is, it can be seen that 20 bits remain. In this case, the audio signal encoding apparatus 200 may reduce the global gain by 1 by looking at the difference between the first number of used bits and the number of frame bits as approximately 24 bits.

As another example, when the number of frame bits is 600, if the number of first used bits is 550, the difference between the number of first used bits and the number of frame bits is 50. In this case, the audio signal encoding apparatus 200 may reduce the global gain by 2 by looking at the difference between the first number of used bits and the number of frame bits as approximately 48 bits.

As described above, the audio signal encoding apparatus 200 according to the first embodiment of the present invention may reduce the global gain according to the difference between the first number of used bits and the number of frame bits. However, how much the global gain is to be reduced according to the difference between the number of first used bits and the number of frame bits is not limited to the above calculation formula.

5 is a flowchart illustrating a step of correcting a quantized signal according to a second embodiment of the present invention.

According to the second embodiment of the present invention, the audio signal encoding apparatus 200 may generate a corrected quantized signal by restoring a signal masked by a masking threshold. According to the second embodiment of the present invention, when the number of bits used in a specific frame is less than the maximum number of bits that can be used for one frame, the number of bits can be adjusted by using an unmasked original frequency domain signal. .

In step S510, the audio signal encoding apparatus 200 selects at least one band from among at least one band masked by the masking threshold.

For example, the audio signal encoding apparatus 200 may compare energy of each frequency band with a masking threshold, and select at least one band based on the comparison result. The audio signal encoding apparatus 200 may preferentially select a band having the smallest difference between the energy of each band for at least one frequency band masked by the masking threshold and the masking threshold for the corresponding band.

As another example, the audio signal encoding apparatus 200 may select all bands masked by the masking threshold when the difference between the number of first used bits and the number of frame bits is equal to or greater than a predetermined number of bits.

In step S520, the audio signal encoding apparatus 200 may generate a second quantized signal by adding a quantized signal for at least one band selected in step S510 to the first quantized signal.

The second quantized signal output according to the second embodiment of the present invention is encoded using a larger number of bits than the first quantized signal by further including a quantized signal for at least one band masked by a masking threshold. Accordingly, the audio signal encoding apparatus 200 according to the second embodiment of the present invention can improve the sound quality of the audio signal by correcting the quantized signal to use the remaining bits.

As described above, according to the method and apparatus for encoding an audio signal according to an embodiment of the present invention, an audio signal to be encoded by correcting a quantized signal using the remaining bits included in each frame and encoding the corrected quantized signal Sound quality can be improved.

According to an audio signal encoding method and apparatus according to an embodiment of the present invention, a global gain is adjusted according to a characteristic of each frame signal and the number of remaining bits, and an unmasked original frequency domain signal is used based on the number of remaining bits. You can create high-quality coded audio signals. Accordingly, since the perceptual acoustic encoding apparatus uses the low-delay audio encoding method, a problem such as sound quality deterioration caused by insufficient number of frame bits can be solved through the present invention.

On the other hand, according to another embodiment of the present invention, the audio signal encoding apparatus 200 corrects the audio signal when bits remain in the encoded audio signal as well as when the bits are insufficient. You can adjust the number of beats. Accordingly, in performing bit allocation, the audio signal encoding apparatus 200 may prevent the total number of bits used for a specific audio frame from exceeding the maximum number of bits available per frame.

6 is a flowchart illustrating a step of correcting a quantized signal according to a third embodiment of the present invention.

Referring to FIG. 6, an audio signal encoding method according to an embodiment of the present invention includes steps processed by the audio signal encoding apparatus 200 illustrated in FIG. 2. Accordingly, it can be seen that the contents described above with respect to the audio signal encoding apparatus 200 illustrated in FIG. 2 are also applied to the audio signal encoding method of FIG. 6 even if omitted below.

Steps S610 to S640 of FIG. 6 correspond to steps S310 to S340 of FIG. 3. Accordingly, it can be seen that even though the contents are omitted hereinafter, the contents described above with respect to the audio signal encoding method shown in FIG. 3 also apply to the audio signal encoding method of FIG. 6.

In step S610, the audio signal encoding apparatus 200 according to an embodiment of the present invention converts the audio signal into a frequency domain signal. In step S620, the audio signal encoding apparatus 200 quantizes the frequency domain signal by applying a global gain. In step S630, the audio signal encoding apparatus 200 calculates the number of used bits used by the quantized signal. In step S640, the audio signal encoding apparatus 200 determines whether the number of used bits is smaller than the number of pre-allocated frame bits for the frame of the audio signal.

In step S350, when the number of used bits is smaller than the number of frame bits, the audio signal encoding apparatus 200 determines whether at least one of the applied global gain and the number of used bits satisfies the repetition end condition.

For example, the audio signal encoding apparatus 200 may determine that the repetition end condition is satisfied when the applied global gain is less than or equal to a predetermined gain, or when the difference between the number of used bits and the number of frame bits is less than or equal to the predetermined number of bits. The predetermined gain and the predetermined number of bits related to the repetition end condition may be a value input by a user, a predetermined value according to an application, or a value calculated according to a frame.

When the applied global gain or the number of used bits satisfies the repetition termination condition, the audio signal encoding apparatus 200 encodes the quantized signal in step S620. (S670)

When the applied global gain or the number of used bits does not satisfy the repetition termination condition, the audio signal encoding apparatus 200 reduces the global gain by a predetermined value and updates it. (S655) As an example, the audio signal encoding apparatus 200 includes: , Based on the difference between the number of used bits and the number of frame bits, it is possible to determine how much to reduce the global gain. As another example, the audio signal encoding apparatus 200 may update the global gain by decreasing it by one. After updating the global gain, the audio signal encoding apparatus 200 returns to step S620 and quantizes the frequency domain signal by applying the updated global gain.

In step S660, when the number of used bits is not smaller than the number of frame bits, the audio signal encoding apparatus 200 determines whether at least one of the applied global gain and the number of used bits satisfies the repetition termination condition.

For example, the audio signal encoding apparatus 200, when the applied global gain is equal to or greater than a predetermined gain, when the difference between the number of used bits and the number of frame bits is less than or equal to the predetermined number of bits, or when the number of used bits and the number of frame bits are the same, It can be determined that the repetition end condition is satisfied. The predetermined gain and the predetermined number of bits related to the repetition end condition may be a value input by a user, a predetermined value according to an application, or a value calculated according to a frame.

When at least one of the applied global gain and the number of bits used satisfies the repetition termination condition, the audio signal encoding apparatus 200 encodes the quantized signal in step S620. (S670)

When the applied global gain or the number of used bits does not satisfy the repetition end condition, the audio signal encoding apparatus 200 increases the global gain by a predetermined value and updates it. (S665) As an example, the audio signal encoding apparatus 200 includes: , Based on the difference between the number of used bits and the number of frame bits, it is possible to determine how much to increase the global gain. As another example, the audio signal encoding apparatus 200 may increase the global gain by 1 and update it. After updating the global gain, the audio signal encoding apparatus 200 returns to step S620 and quantizes the frequency domain signal by applying the updated global gain.

7 is a flowchart illustrating a step of correcting a quantized signal according to a fourth embodiment of the present invention.

In step S610, the audio signal encoding apparatus 200 according to an embodiment of the present invention converts the audio signal into a frequency domain signal.

In operation S710, the audio signal encoding apparatus 200 may mask at least one of a plurality of bands included in the frequency domain signal by applying a masking threshold determined based on the psychoacoustic model.

In step S620, the audio signal encoding apparatus 200 may quantize only the remaining frequency domain signals except for the frequency band masked by the masking threshold. In step S630, the audio signal encoding apparatus 200 may calculate the number of used bits used by the quantized signal. In step S640, the audio signal encoding apparatus 200 may determine whether the number of used bits is smaller than the number of frame bits previously allocated for the frame of the audio signal.

In operation S720, when the number of used bits is less than the number of frame bits, the audio signal encoding apparatus 200 may determine whether at least one of the applied global gain and the number of used bits satisfies the repetition end condition.

For example, the audio signal encoding apparatus 200 may determine that the repetition end condition is satisfied when the applied global gain is less than or equal to a predetermined gain, or when the difference between the number of used bits and the number of frame bits is less than or equal to the predetermined number of bits. The predetermined gain and the predetermined number of bits related to the repetition end condition may be a value input by a user, a predetermined value according to an application, or a value calculated according to a frame.

When the applied global gain or the number of used bits satisfies the repetition termination condition, the audio signal encoding apparatus 200 may encode the quantized signal in step S620. (S670)

When the applied global gain or the number of used bits does not satisfy the repetition termination condition, the audio signal encoding apparatus 200 in step S723 may select at least one band from among at least one band masked by the masking threshold in step S710. .

For example, the audio signal encoding apparatus 200 may compare energy of each frequency band with a masking threshold, and select at least one band based on the comparison result. The audio signal encoding apparatus 200 may preferentially select a band having the smallest difference between the energy of each band for at least one frequency band masked by the masking threshold and the masking threshold for the corresponding band.

As another example, the audio signal encoding apparatus 200 may select all bands masked by the masking threshold when the difference between the number of used bits and the number of frame bits is greater than or equal to a predetermined number of bits.

In step S725, the audio signal encoding apparatus 200 may quantize the frequency domain signal to which the frequency domain signal for the selected band is added. That is, the audio signal encoding apparatus 200 may generate a corrected quantized signal by restoring a frequency domain signal corresponding to a selected band from among the frequency domain signals masked in step S710.

Meanwhile, in step S660, when the number of used bits is not smaller than the number of frame bits, the audio signal encoding apparatus 200 may determine whether at least one of the applied global gain and the number of used bits satisfies the repetition termination condition. When at least one of the applied global gain and the number of bits used satisfies the repetition termination condition, the audio signal encoding apparatus 200 may encode the quantized signal in step S620. (S670)

When the applied global gain or the number of used bits does not satisfy the repetition termination condition, the audio signal encoding apparatus 200 may increase the global gain by a predetermined value and update it (S665).

As described above, according to the audio signal encoding method and apparatus according to an embodiment of the present invention, not only when bits remain in the encoded audio signal, but also when bits are insufficient, the encoded audio signal is corrected by correcting the audio signal. You can adjust the number of beats. Accordingly, according to the audio signal encoding method and apparatus according to an embodiment of the present invention, the audio signal is encoded using an appropriate number of bits for each frame, thereby improving sound quality of an encoded audio signal.

An embodiment of the present invention may also be implemented in the form of a recording medium including instructions executable by a computer, such as a program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer, and includes both volatile and nonvolatile media, removable and non-removable media. Further, the computer-readable medium may include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Communication media typically includes computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave, or other transmission mechanism, and includes any information delivery media.

The above description of the present invention is for illustrative purposes only, and those of ordinary skill in the art to which the present invention pertains will be able to understand that other specific forms can be easily modified without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and are not limiting. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as being distributed may also be implemented in a combined form.

The scope of the present invention is indicated by the claims to be described later rather than the detailed description, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention. do.

Claims (21)

Converting the audio signal into a first frequency domain signal;
Generating a first quantized signal by quantizing the first frequency domain signal by applying a global gain;
Calculating a first number of used bits used by the first quantized signal;
Generating a second quantized signal by correcting the first quantized signal when the number of first used bits is less than the number of pre-allocated frame bits for the frame of the audio signal; And
Encoding the second quantized signal,
Generating the first quantized signal,
Masking at least one of the plurality of bands included in the first frequency domain signal by applying a masking threshold determined based on the psychoacoustic model; And
Generating the first quantized signal by quantizing the masked first frequency domain signal,
Generating the second quantized signal,
And generating the second quantized signal by adding a quantized signal for at least one of the masked at least one band to the first quantized signal. delete delete delete delete The method of claim 1,
Increasing the global gain based on a difference between the first number of used bits and the number of frame bits when the number of first used bits is greater than the number of frame bits previously allocated for a frame of the audio signal; And
And generating the second quantized signal by quantizing the first frequency domain signal by applying the increased global gain. The method of claim 1,
Updating the global gain by increasing the global gain by a predetermined value when the number of first used bits is greater than the number of frame bits previously allocated for the frame of the audio signal;
Quantizing the first frequency domain signal by applying the updated global gain;
Calculating a second number of used bits used by the quantized signal by applying the updated global gain; And
Repeating the steps of updating the global gain, quantizing by applying the updated global gain, and calculating the second number of used bits until the second number of used bits becomes less than or equal to the number of frame bits. An audio signal encoding method further comprising the step of performing. delete The method of claim 1,
Generating the second quantized signal,
Comparing energy for each band masked by the masking threshold and a masking threshold for each band;
Selecting at least one of the masked at least one band based on the comparison result; And
And generating the second quantized signal by adding a quantized signal for the selected at least one band to the first quantized signal. The method of claim 1,
Generating the second quantized signal,
When the difference between the first number of used bits and the number of frame bits is greater than or equal to a predetermined value, generating the second quantized signal by adding a quantized signal for the masked at least one band to the first quantized signal. Audio signal encoding method comprising a. A frequency converter converting the audio signal into a first frequency domain signal;
A quantization unit that generates a first quantized signal by quantizing the first frequency domain signal by applying a global gain, a bit number calculation unit that calculates a first number of used bits used by the first quantized signal, and the first A bit number adjustment quantization unit including a correction unit for generating a second quantization signal by correcting the first quantized signal when the number of used bits is less than the number of frame bits pre-allocated for the frame of the audio signal; And
Including an encoding unit for encoding the second quantized signal,
The quantization unit,
The first quantization is performed by applying a masking threshold determined based on the psychoacoustic model, masking at least one of a plurality of bands included in the first frequency domain signal, and quantizing the masked first frequency domain signal. Generate a signal,
The correction unit,
And generating the second quantized signal by adding a quantized signal for at least one of the masked at least one band to the first quantized signal. delete delete delete delete The method of claim 11,
The correction unit,
When the number of first used bits is larger than the number of frame bits previously allocated for the frame of the audio signal, based on the difference between the number of first used bits and the number of frame bits, the global gain is increased, and the increased And generating the second quantized signal by quantizing the first frequency domain signal by applying a global gain. The method of claim 11,
The correction unit,
When the number of first used bits is greater than the number of pre-allocated frame bits for the frame of the audio signal, an operation of increasing and updating the global gain by a predetermined value, and the first frequency domain signal by applying the updated global gain. The operation of quantizing and calculating the number of second used bits used by the quantized signal by applying the updated global gain, until the number of second used bits becomes less than or equal to the number of frame bits, An audio signal encoding apparatus, characterized in that it repeats. delete The method of claim 11,
The correction unit,
Compare the energy for each band masked by the masking threshold and a masking threshold for each band, select at least one band from among the masked at least one band based on the comparison result, and the first quantized signal And generating the second quantized signal by adding a quantized signal for the at least one selected band. The method of claim 11,
The correction unit,
When the difference between the first number of used bits and the number of frame bits is greater than or equal to a predetermined value, the second quantized signal is generated by adding a quantized signal for the masked at least one band to the first quantized signal. Audio signal encoding device. In a computer-readable recording medium recording a program for executing an audio signal encoding method on a computer,
The above method,
Converting the audio signal into a first frequency domain signal;
Generating a first quantized signal by quantizing the first frequency domain signal by applying a global gain;
Calculating a first number of used bits used by the first quantized signal; And
Generating a second quantized signal by correcting the first quantized signal when the number of first used bits is less than the number of pre-allocated frame bits for the frame of the audio signal; And
Encoding the second quantized signal,
Generating the first quantized signal,
Masking at least one of the plurality of bands included in the first frequency domain signal by applying a masking threshold determined based on the psychoacoustic model; And
Generating the first quantized signal by quantizing the masked first frequency domain signal,
Generating the second quantized signal,
And generating the second quantized signal by adding a quantized signal for at least one of the masked at least one band to the first quantized signal. .

Images