KR20090064857A - Method for processing an audio, and apparatus for implementing the same - Google Patents

Abstract

An audio processing method and a device thereof are provided to store an audio by dividing the audio into a voice section and a non-voice section on a frequency domain, thereby saving computing resources. A receiver(110) receives an audio bit stream. An extractor(120) extracts audio attribute information from the audio bit stream. A voice section identifier(130) identifies whether each section of an audio is a voice section or a non-voice section, based on the audio attribute information. If allocation differences on each channel are small, the corresponding sections are identified as voice sections. If the allocation differences are large, the corresponding sections are identified as non-voice sections.

Description

TECHNICAL FOR PROCESSING AN AUDIO, AND APPARATUS FOR IMPLEMENTING THE SAME}

The present invention relates to an audio processing method and apparatus, and more particularly, to an audio processing method and apparatus capable of playing or storing audio received through a broadcast signal.

In general, players receive digital broadcasts such as DMB and DAB, and not only play video or audio, but also record, record, and store in a player. As described above, in playing or storing audio, a user may want to select and extract only a voice section or a music section (non-speech section) from the audio signal. , ZCR (Zero Crossing Rate), Energy, Pitch, Mel-Cepstrum and the like have been proposed. However, these methods must decode the audio bitstream to generate a PCM signal, and then go through a process of extracting a speech or music section from the PCM signal. That is, since both the decoding process and the interval extraction process have to be performed, there is a problem in that computing resources are wasted.

The present invention has been made to solve the above problems, and audio processing capable of separating audio into speech and non-voice sections in the frequency domain without having to generate or analyze a PCM signal after receiving the audio bitstream. Its purpose is to provide a method and apparatus.

Another object of the present invention is to provide an audio processing method and apparatus capable of separating received audio into a voice section and a non-voice section using audio attribute information included in a bitstream.

In order to achieve the above object, an audio processing method according to the present invention comprises the steps of: receiving an audio bitstream; Extracting audio attribute information from the audio bitstream; And identifying whether each section of the audio is a voice section or a non-voice section based on the audio attribute information.

According to the present invention, the audio attribute information may include one or more of an allocation, a scale factor, and a sample value.

According to the present invention, the identifying may be performed for each frequency band.

According to the present invention, the audio property information includes a channel-specific location or a channel-specific sample value, and the identifying includes: a section having a relatively small channel-to-channel difference or a channel-specific difference of the sample value. May be identified as a voice interval, and a section having a relatively large difference between channels of the location or the sample value may be identified as a non-voice interval.

According to the present invention, the audio property information includes scale factor information for each channel, and the identifying may include identifying a section having a small deviation for each channel of the scale factor as a voice interval, and for each channel deviation of the scale factor. The larger interval may be identified as a non-voice interval.

According to the present invention, the audio attribute information includes time-scale scale factor information, and the identifying includes identifying a section having a large temporal deviation of the scale factor as a speech section and a section having a small temporal deviation of the scale factor. Can be identified as a non-voice interval.

According to the present invention, the audio attribute information includes a sample value for each channel, and the identifying may include identifying a section having a small difference between channels of the sample value as a voice section and a section having a small scale factor. It can be identified by the voice interval.

According to the present invention, the method may further include performing one or more of reproduction and storage on one of the audio section and the non-voice section of the audio based on the identification result.

According to the invention, the reproduction can be performed using a synthetic subband filter.

According to another aspect of the invention, the receiver for receiving an audio bitstream; An extraction unit for extracting audio attribute information from the audio bitstream; And a voice section identification unit for identifying whether each section of the audio section is a voice section or a non-voice section based on the audio attribute information.

According to an aspect of the present invention, since the audio can be divided and stored in a voice interval and a non-voice interval in the frequency domain without receiving an audio bitstream and generating and analyzing a PCM signal, the computing resource is saved. have.

According to another aspect of the present invention, since audio may be divided into a voice section and a non-voice section using attribute information included in the audio bitstream, there is also an effect of saving computing resources.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Prior to this, terms or words used in the specification and claims should not be construed as having a conventional or dictionary meaning, and the inventors should properly explain the concept of terms in order to best explain their own invention. Based on the principle that can be defined, it should be interpreted as meaning and concept corresponding to the technical idea of the present invention. Therefore, the embodiments described in the specification and the drawings shown in the drawings are only the most preferred embodiment of the present invention and do not represent all of the technical idea of the present invention, various modifications that can be replaced at the time of the present application It should be understood that there may be equivalents and variations.

1 is a view showing the configuration of an audio processing apparatus according to an embodiment of the present invention, Figure 2 is a view showing the procedure of the audio processing method according to an embodiment of the present invention. First, referring to FIG. 1, an audio processing apparatus 100 according to an exemplary embodiment of the present invention includes a receiver 110, an extractor 120, a voice segment identification unit 130, a playback unit 140, and a storage unit. And 150.

1 and 2, first, the receiver 110 receives an audio bitstream through a broadcast signal (S120). The extraction unit 120 parses the received audio bitstream and extracts audio attribute information (step S120). The audio attribute information may include channel-specific frequency (α), channel-specific frequency scale factor (β), and channel-specific frequency sample value (γ). 3 is a diagram illustrating an example of a bitstream structure received in the present invention. Referring to FIG. 3, each channel-specific frequency α (allocation [ch] [sb]), channel-specific frequency scale factor β, scalefactor [ch] [sb] [], and channel-specific frequency It can be seen that the sample value γ (sample [ch] [sb] []) is included. Here, as shown in the figure (if (allocation [ch] [sb]! = 0 {}), the location α may be used to extract the spacing factor β and the sample value γ. .

Referring back to FIGS. 1 and 2, the speech section identifying unit 130 identifies whether the corresponding section of the audio corresponds to the speech section or the non-voice section by using the audio attribute information extracted in step S120 ( Step S130 to step S160). In detail, when the difference between channels of the location α is small (YES in step S130), the corresponding section is identified as a voice section (step S170), and vice versa, a non-voice section is identified. (Step S180).

4 is a diagram illustrating an example of a difference for each channel of an allocation. Here, the channel-specific difference in allocation indicates, for example, the difference in allocation between the left channel and the right channel. Referring to FIG. 4, the x-axis is a play time [sec], and the y-axis is a LR difference (allocation value) between the left and right positions, and the voice section SP and the music section MU (pop) , MU (Latin Music), MU (Bossa Nova), MU (String Performance), MU (Pop-Guitar Accompaniment), MU (Acapella)). This is because the difference between channels is smaller than that of MUs, because the voices of singers, moderators and hosts are generally similar in the left and right channels, while in music, the similarity tends to be inferior in the left and right channels. In operation S130, whether the corresponding audio section is a voice section or a non-voice section may be identified based on the channel-specific difference of the location extracted from the bitstream.

On the other hand, the voice interval identification unit 130 determines whether the channel-specific difference of the scale factor β extracted in step S120 is small, and if the channel-specific difference of the scale factor β is small (YES in step S140). In step S170, the corresponding audio section is identified as a voice section.

5 is a diagram illustrating an example of difference between channels of a scale factor. The x axis is also the play time [sec] and the y axis is the Mean of LR SCF difference. The scale factor is a value for reflecting the size of the largest component in units of bands and is also included in the MUSICAM (MPEG-1 Audio Layer II, MP2) syntax of the digital audio broadcasting standard (DAB). If the left and right channel signals are similar to each other, the scale factor also has a similar value for each channel. On the contrary, if the left and right channel signals have a large difference, the scale factor also has a large difference for each channel. Referring to FIG. 5, it can be seen that the left and right difference of the scale factor of the voice section SP is lower than that of any kind of music section MP. 5 is a root mean square (RMS) in the PCM domain.

On the other hand, the voice section identification unit 130 determines whether the temporal deviation of the scale factor β extracted in step S120 is large (step S150), and if the temporal deviation is large (Yes in step S150), the corresponding audio The section is identified as a voice section (step S170). In the opposite case, it is identified as a non-voice interval (step S180).

6 is a diagram illustrating an example of temporal deviation of a scale factor. The x-axis is the play time [sec]) and the y-axis is the scale factor. In FIG. 6A, the '63 -scale factor index 'and FIG. 6B are the '63 -scale factor index' Fig. 6C shows the interval of the scale factor temporal deviation. Here, the scale factor index is an indicator indicating the value of the scale factor SCF, and is defined as follows in the case of MUSICAM.

SCF =

SCF =

SCF: Scale Factor

index: scale factor index

When the scale factor index has a value of 0 to 62, the scale factor SCF has a value of 2-0.00000120155435. In other words, the smaller the scale factor index is, the more inversely proportional the scale factor SCF becomes. Accordingly, since the '63 -scale factor index 'value in FIG. 6A and (B) may be proportional to the scale factor (SCF), when the '63 -scale factor index' is large The scale factor (SCF) value can also be said to be large.

In the case of the voice interval, since there is a short interval close to the silence between the phonograms, such a calculation has a relatively small size, so that the scale factor is also reduced to express this. Accordingly, in the case of the voice interval, since the variation of the scale factor is large to express the low magnitude generated in the short interval, that is, the variation is large. Referring to FIG. 6B, it can be seen that the variation of the scale factor 63-scale factor index in the speech section SP is much larger than the deviation in the music section MU. Referring to FIG. 6C, it can be seen that the interval in the voice section SP is much larger than the interval in the music section.

Finally, the voice interval identification unit 130 determines whether the channel difference of the sample value γ extracted in step S120 is large, and if the channel difference is large (YES in step S160), the corresponding audio. The section is identified as a voice section (step S170), and vice versa, the section is identified as a non-voice section (step S180).

7 is a diagram illustrating an example of difference between sample value channels. The x-axis is the play time [sec] and the y-axis is the average of the left and right differences of the dequantized sample values. Referring to FIG. 7, similar to the channel-to-channel difference of the location α and the channel-to-channel difference of the scale factor β, the channel-to-channel difference of the sample value γ is also very low in the voice interval SP and is non-voice. It can be seen that the interval MU is relatively high. The difference between the channel-specific difference of the allocation α and the sample value γ may have a proportional relationship. The relationship between the channel-specific difference and the channel-specific difference of the sample value is shown in FIG. 8.

As such, the voice section identification unit 130 identifies whether the corresponding audio section is a voice section or a non-voice section through steps S130 to S160. When audio attribute information exists for each frequency band, the identification process described in steps S130 to S160 is performed. It may also be performed for each frequency band. That is, it is possible to determine which band of which audio section is a voice section or a non-voice section. In addition, the steps S130 to S160 need not be all performed, and any one or more of four steps may be performed.

When the voice section identifier 130 completes the identification of the voice section and the non-voice section as described above, one of the voice section and the non-voice section is reproduced or stored as a PCM signal through the playback unit 140 according to a user's selection. It is stored in the unit 150 (steps S175 and S185). At this time, the reproduction unit 140 may reproduce the PCM signal using the synthesized subband filter. For example, only the background music may be excluded and only the background music may be stored, or the music section may be deleted and only the voice may be stored.

As described above, although the present invention has been described by way of limited embodiments and drawings, the present invention is not limited thereto and is intended by those skilled in the art to which the present invention pertains. Of course, various modifications and variations are possible within the scope of equivalents of the claims to be described.

The present invention can be applied to a broadcast receiver, an audio player, and the like.

1 is a block diagram of an audio processing apparatus according to an embodiment of the present invention.

2 is a flowchart of an audio processing method according to an embodiment of the present invention.

3 is a view showing an example of a bitstream structure received in the present invention.

4 is a diagram illustrating an example of channel-to-channel difference of location;

5 is a diagram illustrating an example of difference between channels of a scale factor.

6 is a diagram illustrating an example of temporal deviation of a scale factor.

7 is a diagram illustrating an example of difference between sample value channels.

8 is a diagram illustrating a relationship between a channel-specific difference of location and a channel-to-channel difference of sample values.

Claims (10)

Receiving an audio bitstream; Extracting audio attribute information from the audio bitstream; And, And identifying whether each section of the audio is a voice section or a non-voice section based on the audio attribute information. The method of claim 1, The audio attribute information includes at least one of an allocation, a scale factor, and a sample value. The method of claim 1, The identifying may be performed for each frequency band. The method of claim 1, The audio attribute information includes channel-specific location or channel-specific sample value, The identifying may include identifying, as a voice interval, a section having a relatively small channel-to-channel difference or a channel-specific difference of the sample value as a voice section, and a section having a relatively large channel-to-channel difference of the location or sample value. Audio processing method characterized in that it identifies by a non-voice interval. The method of claim 1, The audio attribute information includes scale factor information for each channel. The identifying may include identifying a section having a small channel deviation of the scale factor as a voice section and identifying a section having a large channel variation of the scale factor as a non-voice section. The method of claim 1, The audio attribute information includes scale factor information for each time. The identifying may include identifying a section having a large temporal deviation of the scale factor as a voice section and identifying a section having a small temporal deviation of the scale factor as a non-voice section. The method of claim 1, The audio attribute information includes a sample value for each channel, The identifying may include identifying a section having a small difference between channels of the sample value as a voice section and identifying a section having the small scale factor as a non-voice section. The method of claim 1, And based on the identification result, performing one or more of reproduction and storage of one of a voice section and a non-voice section of the audio. The method of claim 8, And said reproducing is performed using a synthetic subband filter. A receiver for receiving an audio bitstream; An extraction unit for extracting audio attribute information from the audio bitstream; And, And a speech section identifying unit for identifying whether each section of the audio section is a speech section or a non-voice section based on the audio attribute information.

Images