JP2015227912A - Audio coding device and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To code an audio signal of a certain information amount into data having a less information amount which is hardly buried in noise.SOLUTION: An audio coding device (232) uses a masking threshold on the basis of audibility property for reducing an information amount, and codes an audio signal. The audio coding device comprises: a first processing part (239, 240, 242) which corrects the masking threshold (302) according to a surrounding noise level (304) received from a terminal, at the time, an input signal (350) which has larger difference between it and the masking threshold, than a certain threshold, becomes larger than corrected masking threshold (306, 310), corrects at least one of the masking threshold (302) and the input signal (350); and a second processing part (236, 246) for coding the input signal (350, 351) which is larger than the corrected masking threshold (306, 310), after at least one of the masking threshold and the input signal has been corrected.

Description

  The present invention relates to encoding audio signals.

  Known recording devices can receive and record audio and video streams of broadcast programs according to user operations. A known recording device with a copyright protection function further converts the recorded audio and video streams to a low transmission rate in response to a transmission request from a portable information terminal by a user via a mobile communication network and the Internet. Can be transmitted to the portable information terminal. The recording device with a copyright protection function may be adapted to, for example, the function DTCP + (Digital Transmission Content Protection Plus) of the copyright protection standard. The user can receive and play back the low-rate audio and video stream from the recording apparatus with a copyright protection function on the portable information terminal.

  In a known signal encoding method, a transmission mode determination unit detects environmental noise included in the background of a voice or musical sound signal in an input signal, and is a communication terminal that is a counterpart communication terminal according to the level of the environmental noise A transmission mode for controlling a transmission bit rate of a signal transmitted from the apparatus is determined. In addition, the signal decoding unit decodes the encoded information transmitted from the communication terminal device via the transmission path, and outputs the obtained signal as an output signal. At this time, the signal decoding unit compares the transmission mode information included in the encoded information output from the transmission path with the transmission mode information obtained from the transmission mode determination unit in consideration of the transmission delay, Detect transmission errors. Thus, by controlling the transmission bit rate on the transmission side in consideration of the usage environment on the reception side, it is possible to efficiently encode a voice or musical sound signal while maintaining the quality.

  In a known communication terminal, a background noise detection unit detects a noise level of background noise. Further, the control unit determines the decoding level based on the noise level. The control unit causes the compression / decompression processing unit to perform AMR (Adaptive Multi-Rate) decoding at the decoding level thus determined. As a result, it is possible to prevent the output of sound with excessive quality and to improve efficiency.

  In a certain known communication terminal device, the lower layer information that notifies the upper layer of the information of the lower layer to the conventional communication terminal device configured to hide the lower layer information such as the radio wave intensity and the network congestion from the application. A new layer management unit is introduced. As a result, the connection control unit that provides the service generation environment of the application can directly recognize the information of the lower layer, and various situation determinations and control changes based on the information of the lower layer can be performed at the application level. This also provides a communication terminal that allows the user to continue video and audio communication with a seamless feeling even when the communication status changes in a mobile environment.

JP 2005-241761 A JP 2003-218781 A JP-T-2004-266330

E. Kurniawati, CT Lau, B. Premkumar, J. Absar, S. George, “New Implementation Techniques of an Efficient MPEG Advanced Audio Coder”, [Search March 7, 2014] Internet <http: // www. img.lx.it.pt/~fp/cav/ano2008_2009/Trabalhos_MEEC_2009/Artigo_MEEC_11/pagina%20web/pagina%20web/referencias/artigo3.pdf>

  When audio stream data recorded at a certain rate on the recording device is re-encoded and transmitted at a low rate on the transmission side and reproduced on the reception side, the reproduced sound on the reproduction side is received on the reception side. It is difficult to hear for the listener because of being buried in the ambient noise, and the listening sound quality may be deteriorated.

  The inventors can improve the effective listening sound quality by converting the audio stream data recorded at a certain rate into the audio stream data in a form that is hard to be buried in ambient noise during reproduction. Recognized that it was possible.

  In one aspect, an object of the present invention is to encode an audio signal having a certain amount of information in a form that is less likely to be buried in noise and with a small amount of information.

  According to the embodiment of the present invention, there is provided an audio encoding device that encodes an audio signal using a masking threshold based on auditory characteristics so as to reduce the amount of information. The audio encoding device corrects the masking threshold according to the ambient noise level received from the terminal. At this time, an input signal whose difference from the masking threshold is larger than a certain threshold is larger than the corrected masking threshold. A first processing unit that corrects at least one of the masking threshold and the input signal, and a second processor that encodes the input signal after at least one of the corrections that is greater than the corrected masking threshold. And a processing unit.

  According to the embodiment of the present invention, an audio signal having a certain amount of information can be encoded in a form that is less likely to be buried in noise and with a small amount of information.

FIG. 1 shows an example of a communication system according to an embodiment. FIG. 2 shows an example of a schematic configuration of the transcoder device. FIG. 3 shows an example of a schematic configuration of the portable information terminal. 4A to 4C illustrate examples of frequency spectra of decoded audio stream data processed by the acoustic encoding unit. FIG. 5 shows an example of a flowchart of processing for generating and transmitting noise power information related to a sound including ambient noise, which is executed by the portable information terminal. FIGS. 6A and 6B are examples of a flowchart of a process for performing audio coding of decoded video stream data according to the noise power level of the portable information terminal, which is executed by the acoustic coding unit of the transcoder device. Is shown. (Explained in Fig. 6A) FIG. 7 shows an example of a more specific flowchart of step 620 in FIG. 6B. 8A to 8C show examples of correction functions of three elements. FIG. 9 is a diagram for explaining an example of the peak nature of the spectrum of an audio signal. FIG. 10 shows an example of another schematic configuration of the portable information terminal according to a modification of the above-described embodiment. FIG. 11 is a modification of the flowchart of FIG. 5 and shows an example of a flowchart of processing for generating and transmitting noise power and a signal-to-noise ratio, which is executed by the portable information terminal. FIG. 12 is a modification of the flowchart of FIG. 7 and shows an example of another flowchart of the process for correcting the composite masking threshold value with elements and three elements. FIG. 13 is a modification of the transcoder apparatus of FIG. 2 and shows another example of a schematic configuration of the transcoder apparatus. FIGS. 14A and 14B are flowcharts of a process for auditory encoding of decoded video stream data according to the noise power level and signal-to-noise ratio of the portable information terminal, which is executed by the acoustic encoding unit. An example is shown. (Explained in Fig. 14A)

The objects and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims.
It is understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not intended to limit the invention.

  Non-limiting embodiments of the present invention will be described with reference to the drawings. In the drawings, similar components and elements have the same reference numerals.

  When audio (acoustic) stream data recorded on a recording device at a certain rate is re-encoded and transmitted at a low rate on the transmission side device and played back on the reception side terminal, The reproduced sound is sometimes buried in the ambient noise of the terminal and is difficult for the listener to hear. On the other hand, with respect to audio stream data that is encoded and transmitted at a low rate by the transmission-side apparatus, the reception-side terminal increases the gain of the signal in the frequency partial band that is difficult to hear by ambient noise at the reception-side terminal. As a result, the quantization error of encoding increases and the sound quality of the reproduced sound decreases.

  On the other hand, as a human auditory characteristic, i.e., a human perceptual characteristic of a sound, a sound larger than a certain threshold power level at a certain frequency is smaller than a certain threshold power level at a nearby frequency depending on the sound level. There is an effect of making other sounds inaudible, that is, a masking effect. The masking threshold represents such a threshold power level distributed over the audible frequency band, and represents a limit value (upper limit) of spectral power that cannot be perceived by a human due to the masking effect. On the other hand, perceptual coding or auditory coding based on auditory characteristics uses a masking effect to encode only a subband audio signal that can be perceived according to the spectrum and level of the audio signal, and a masking threshold. The following partial band signals are not encoded. Auditory coding increases the effective coding efficiency of audio signals.

  The inventors select the audio signal to be encoded according to the masking threshold and the ambient noise for each subband, assigning a large number of bits only to the audible subband signal, even at a low rate. Recognized that audio signals can be encoded in a form that is less susceptible to noise. In addition, if the inventors select the audio signal of each partial band according to the level of ambient noise on the receiving side, and perform level correction and encoding, then the inventor will not be buried in ambient noise based on human auditory characteristics. Recognized that audio signals can be encoded into low rate data.

  An object of the embodiment is to adaptively encode or convert an audio signal into encoded data having a form that is less likely to be buried in noise and having a small amount of information according to the ambient noise level. This object is achieved by the embodiment.

  FIG. 1 shows an example of a communication system 2 according to the embodiment.

  In FIG. 1, the communication system 2 includes a portable information terminal 10 and an information processing device 20. In response to a request from the portable information terminal 10 via the communication network 5, the information processing device 20 decodes the encoded audio and video stream in which the broadcast stream is recorded at a certain data rate, and replays it at a low rate. It can be encoded and transmitted to the portable information terminal 10. The communication network 5 may include a mobile communication network including a radio base station or an access point 50, and an IP (Internet Protocol) network such as the Internet.

  The mobile information terminal 10 may be, for example, a mobile phone, a smartphone, a tablet terminal, or a mobile personal computer.

  The information processing apparatus 20 may be, for example, a broadcast recording apparatus or a broadcast recording / reproducing apparatus such as an HDD (Hard Disk Drive) recorder, or a personal computer with a broadcast recording / reproducing function. In FIG. 1, an information processing apparatus 20 includes, for example, a processor 22, a storage device 24, a network interface (NW / IF) 26, an input unit 27, a transcoder device or transcoder unit 200, a broadcast receiver 202, and a drive 28. Contains. However, the information processing device 20 itself may be referred to as a transcoder device or a transcoder device.

  The processor 22 may be a CPU (Central Processing Unit) for a computer. The storage device 24 includes a main storage device and an auxiliary storage device. The main storage device includes a storage device such as a semiconductor memory. Further, the auxiliary storage device includes, for example, a hard disk drive (HDD) and / or a semiconductor memory such as a flash memory. At least a portion of the auxiliary storage device of the storage device 24 may function as a stream data recording unit. The input unit 27 may include, for example, a plurality of keys, a touch pad, a numeric keypad, a keyboard, and / or a touch panel.

  The drive 28 may be for reading software and a recording medium 284, such as an optical disk or magnetic disk, on which audio and video stream data is recorded. The software may include, for example, an OS, a database management system (DBMS), an application program, and the like. The application program may include an application for transmitting the requested stream data to the portable information terminal 10 in response to receiving an operation command or request from the portable information terminal 10.

  The processor 22 may be a dedicated processor implemented as an integrated circuit, for example. Further, the processor 22 may operate in accordance with an application program stored in the storage device 24. The application program may be stored in the recording medium 284, read from the recording medium 284 by the drive 28, and installed in the information processing apparatus 20.

  In the information processing apparatus 20, the processor 22 causes the broadcast receiver 202 to receive, for example, terrestrial and satellite broadcast signals in accordance with user operations, and records the audio and video (AV) stream data in the storage device 24. Can be made. The broadcast signal may carry audio and video stream data copyright-protected by, for example, DTCP (Digital Transmission Content Protection). In the information processing apparatus 20, the processor 22 can receive a stream transmission request from the portable information terminal 10 by a user operation via the communication network 5. Further, in response to the request from the portable information terminal 10, the processor 22 reads the audio and video stream data recorded in the storage device 24, and causes the transcoder device 200 to decode and decode the low-rate stream signal. Data can be re-encoded. The processor 22 can further transmit the re-encoded stream data to the portable information terminal 10 via the communication interface 5 via the network interface 26.

  FIG. 2 shows an example of a schematic configuration of the transcoder device 200.

  In FIG. 2, the transcoder apparatus 200 includes, for example, an information input unit 210, a stream input unit 222, an acoustic decoding unit 226, an acoustic encoding unit 230, an image decoding unit 256, an image encoding unit 260, a multiplexing unit 270, and A stream output unit 280 is included. The acoustic encoding unit 230 includes, for example, a signal conversion unit 236, a masking threshold generation unit 238, a masking threshold correction unit 239, an encoding determination unit 240, a level correction unit 242, a masking correction unit 244, a quantization unit 246, and a multiplexing. Part 248. The acoustic encoding unit 230 may be an encoder that complies with an audio compression encoding standard such as AAC (Advanced Audio Coding), MP3 (MPEG Audio Layer-3), or AC3 (Audio Code number 3). Therefore, the low-rate audio stream data generated by the acoustic encoding unit 230 may be encoded by an audio compression encoding standard such as AAC, MP3, or AC3.

  FIG. 3 shows an example of a schematic configuration of the portable information terminal 10.

  The portable information terminal 10 is related to an audio signal other than telephone communication, for example, a receiving unit 130, a decoding unit 132, an acoustic reproduction unit 136, a speaker 138, a microphone 142, an acoustic input unit 144, a conversion unit 146, and a measurement unit 148. , And a transmission unit 150. The receiver 130 and transmitter 150 may be coupled to a mobile transceiver wireless transceiver (not shown) or may include a mobile transceiver. However, in FIG. 3, other telephone communication units, information processing units, and the like of the portable information terminal 10 are not shown. The decoding unit 132 may be a decoder according to an audio compression coding standard such as AAC, MP3, or AC3.

  In the portable information terminal 10, the receiving unit 130 is coupled to the decoding unit 132, the decoding unit 132 is coupled to the sound reproduction unit 136, and the sound reproduction unit 136 is coupled to the speaker 138. The microphone 142 is coupled to the acoustic input unit 144, the acoustic input unit 144 is coupled to the conversion unit 146, the conversion unit 146 is coupled to the measurement unit 148, and the measurement unit 148 is coupled to the transmission unit 150. The sound reproduction unit 136 may be further coupled to the sound input unit 144.

  The microphone 142 captures sound including noise around the portable information terminal 10 and generates a sound signal including noise, and the sound input unit 144 receives the sound signal, filters it, and amplifies it. Further, the conversion unit 146 converts the time-domain sound signal from the sound input unit 144 into a frequency-domain spectrum. The transformation may be, for example, a fast Fourier transform (FFT). The measurement unit 148 also divides the frequency spectrum into a plurality of partial bands, detects the power of the spectrum of each partial band, and generates noise power of each partial band. The transmission unit 150 transmits noise power information including noise power of each partial band to the information processing device 20. On the other hand, the portable information terminal 10 may transmit noise power information related to ambient noise to the information processing apparatus 20 together with a stream transmission request according to a user operation or as a part of the stream transmission request. Further, the portable information terminal 10 may generate new noise power information periodically (for example, at intervals of 3 or 5 minutes) and transmit the new noise power information to the information processing apparatus 20.

  On the other hand, the receiving unit 130 receives encoded audio stream data from the information processing apparatus 20, and the decoding unit 132 decodes the encoded audio stream data to generate an audio signal. The sound reproducing unit 136 amplifies and reproduces the decoded audio signal and supplies the amplified audio signal to the speaker 138 to generate sound. When the audio reproduction unit 136 reproduces the audio signal and the speaker 138 generates sound, the sound input unit 144 subtracts the audio signal being reproduced by the sound reproduction unit 136 from the sound signal captured by the microphone 142. Then, the sound signal of the noise component may be extracted and generated.

  In FIG. 2, in response to the stream transmission request from the user's portable information terminal 10, the processor 22 reads the audio and video stream data of the requested program from the storage device 24 and supplies it to the transcoder device 200. Re-encode at a lower rate. Further, the processor 22 receives noise power information from the portable information terminal 10 and supplies it to the transcoder device 200.

  In the transcoder device 200, the information input unit 210 inputs noise power information and supplies the noise power level (FIG. 4A, 304) of each partial band to the acoustic encoding unit 230. In this case, the noise power information including the noise power is supplied to the masking threshold correction unit 239 and the encoding determination unit 240. On the other hand, the stream input unit 222 separates the audio and video stream data extracted from the storage device 24, supplies the audio stream data to the acoustic decoding unit 226, and converts the video stream data into the image decoding unit. 256.

  The image decoding unit 256 decodes the encoded video stream data and supplies the decoded video stream data to the image encoding unit 260. The image encoding unit 260 re-encodes the decoded video stream data at a low rate, and supplies the encoded video stream data to the multiplexing unit 270.

  On the other hand, the acoustic decoding unit 226 decodes the encoded audio stream data and supplies the decoded audio stream data to the acoustic encoding unit 230. In the acoustic encoding unit 230, the signal conversion unit 236 converts the received decoded audio stream data in the time domain into a frequency domain spectrum, divides the frequency spectrum into a plurality of partial bands, and converts the spectrum of each partial band. Generate power. The signal conversion unit 236 further supplies the spectrum of each partial band to the masking threshold value generation unit 238, the encoding determination unit 240, and the level correction unit 242. The signal conversion unit 236 may be, for example, a converter for MDCT (modified discrete cosine transform).

  Next, how to re-encode the decoded audio stream data in the acoustic encoding unit 230 according to the embodiment will be schematically described.

  4A to 4C show examples of frequency spectra of decoded audio stream data processed by the acoustic encoding unit 230. FIG. 4A to 4C correct the initial masking threshold 302, select the power 350, 351 of the audio signal to be encoded with the corrected masking threshold 310, and select the power 332, 334, 338, 340 to be level corrected. It is for explaining the method.

  4A-4C, the power (power) 330 (bar graph) of each subband of the decoded audio signal and the initial masking threshold (profile) 302 (power) for the power 330 of each subband of the audio signal with respect to frequency f. (Broken line) is shown. Here, the power 330 of each partial band of the audio signal in FIGS. 4A to 4C represents the total power of each partial band obtained by dividing the entire audible frequency band. FIG. 4A shows the noise power level 304 (thin broken line) of each partial band with respect to the frequency f of ambient noise captured by the portable information terminal 10. In this case, the audible frequency band may be a frequency range of 20 Hz to 24 kHz, for example.

  The masking threshold value generation unit 238 generates an initial masking threshold value 302 and masks the distribution of each power 330 of the decoded audio signal of the block in a certain time interval with respect to each partial band based on the masking characteristic based on the auditory characteristic. The threshold value correction unit 239 is supplied. On the other hand, the information input unit 210 supplies the noise power level 304 over a plurality of partial bands included in the noise power information received from the portable information terminal 10 to the masking threshold correction unit 239 and the encoding determination unit 240.

  Next, the masking threshold correction unit 239 corrects each partial band by combining the initial masking threshold 302 with the noise power level 304 to form a combined masking threshold 306 (thick one-dot chain line). For example, the combination compares the masking threshold 302 and the noise power level 304 for each subband and replaces the masking threshold 302 with the noise power level 304 if the noise power level 304 exceeds the power of the masking threshold 302. Is done by doing. In other words, the combination may be a process of selecting the larger power level of the masking threshold 302 and the noise power level 304 for each partial band to form the combined masking threshold 306.

  In FIG. 4A, comparing each power 330 of the decoded audio signal with the composite masking threshold 306, for example, the power 336 for a certain partial band is greater than the composite masking threshold 306. Here, it is assumed that the power 336 in the partial band larger than the initial masking threshold 302 is audio-encoded and transmitted to the portable information terminal 10. In this case, the power 336 in the partial band sufficiently larger than the composite masking threshold 306 can be effectively heard in the portable information terminal 10. However, another sub-band power 331 that is slightly smaller than the composite masking threshold 306 is only slightly audible and has low audible playback quality. Also, if only the partial band power 336 greater than the combined masking threshold 306 is aurally encoded, the remaining unused portion of the transmission rate or transmission capacity is wasted. Therefore, the transcoder 200 selects the power of another partial band that contributes to the improvement of the auditory reproduction sound quality within the range of the transmission rate, performs level correction or amplification, and additionally performs auditory encoding and transmission. Can do. Thereby, the transcoder device 200 can improve the effective auditory reproduction sound quality of the audio signal in the portable information terminal 10.

Next, in FIG. 4B, the masking threshold value correction unit 239 uses the power of the combined masking threshold value 306 over all the partial bands, and the estimated margin transmission capacity range as indicated by a downward arrow with the masking threshold value 302 as a lower limit, depending on the transmission rate. The correction is made by moving it uniformly low within. Thereby, a masking threshold value 308 indicated by a thick broken line is obtained. This correction is a correction by a first element f ₁ described later.

Next, the masking threshold value correction unit 239 sets the masking threshold value 308 by a predetermined amount or lower than each power 350 with the masking threshold value 302 as a lower limit for and in the vicinity of each power 350 of the audio signal sufficiently larger than the masking threshold value 302. Correct so that At this time, as such a partial band to be corrected, for example, a partial band in which the gain ratio of the signal-to-masking threshold with respect to the audio signal power 330 and the initial masking threshold 302 is larger than the gain ratio threshold is selected. May be. The predetermined amount may be determined according to a gain ratio of the signal to masking threshold. Here, the gain ratio threshold value may be one or more values such as 1.0 to 1.2, for example. This correction is a correction according to the second element f ₂ to be described later.

The masking threshold correction unit 239 also sets the masking threshold 302 as the lower limit and the masking threshold 308 by a predetermined amount or lower than the power 351 for the power 351 of the locally high-peak audio signal in each partial band and adjacent partial bands. Correct as follows. At this time, as such a partial band to be corrected, for example, the average gain ratio between the power 330 of the audio signal of a certain partial band larger than the masking threshold 302 and the power 330 of the audio signal of the adjacent partial band A partial band whose value is greater than the gain ratio threshold is selected. The predetermined amount may be determined according to the peak power 351. Here, the gain ratio threshold may be 2.0, 2.5, or 3, for example. This correction is a correction according to the third element f ₃ to be described later.

  In this way, the masking threshold correction unit 239 generates corrected masking thresholds 310 (thick broken line 308 and thick solid line 310 in FIG. 4B, thick solid line 310 in FIG. 4C) over all partial bands. Next, the masking threshold correction unit 239 supplies the correction masking threshold 310 of each partial band to the encoding determination unit 240 and the masking correction unit 244.

  Next, the encoding determination unit 240 compares the power 330 of the decoded audio signal with the correction masking threshold 310 for each partial band in FIG. 4B. The encoding determination unit 240 selects a partial band having a power larger than the correction masking threshold 310, and determines the data 350 and 351 of the audio signal in the selected partial band as an encoding target. On the other hand, the encoding determination unit 240 determines that the data of the power 352 in the partial band of the audio signal having power equal to or less than the correction masking threshold 310 is not to be encoded.

  Next, the coding determination unit 240 performs level correction or amplification so that the power 330 that is equal to or lower than the combined masking threshold 306 among the powers 350 and 351 of the partial band determined as the encoding target is greater than the combined masking threshold 306. To determine the amplification factor. At this time, the level of the power 336 sufficiently larger than the composite masking threshold 306 among the powers 350 and 351 of the partial band determined as the encoding target may be maintained as it is, or the amplification factor may be determined to be 1. Next, the encoding determination unit 240 notifies the level correction unit 242 of the partial bands of power 350 and 351 determined as encoding targets and the amplification factors of the respective partial bands.

  The level correction unit 242 determines the partial band of the power 330 of the audio signal of each partial band received from the signal conversion unit 236 according to the partial bands of the power 350 and 351 to be encoded received from the encoding determination unit 240 and the amplification factor. The power 350, 351 of the current is amplified. FIG. 4C shows audio signal powers 332, 334, 338 and 340 (thick bar-shaped broken lines) amplified to be greater than the combined masking threshold 306. Thereby, the sub-band power 350, 351 is amplified as the power 332, 334, 338, 340. On the other hand, the level correction unit 242 deletes or amplifies the power 352 of the audio signal in the partial band not to be encoded with an amplification factor of zero. Thereby, the amount of information of the audio signal to be encoded or transmitted is reduced.

  The power 332 to 340 of the audio signal after the level correction by the level correction unit 242 may not be completely quantized within a set transmission rate or transmission capacity based on the correction masking threshold 310. The masking correction unit 244 can further correct the correction masking threshold 310 so as to maintain the sound quality as much as possible so that the power of the encoded audio signal satisfies the number of bits of the transmission rate.

  Next, the quantization unit 246 quantizes the powers 332 to 340 of the audio signal corrected by the level correction unit 242 for each partial band. The multiplexing unit 248 multiplexes the quantized data of each partial band, and supplies it to the multiplexing unit 270 as encoded audio stream data.

  The multiplexing unit 270 multiplexes the encoded audio stream data with the video stream data from the image encoding unit 260 to generate audio and video stream data encoded at a low rate. And supplied to the stream output unit 280.

  The stream output unit 280 outputs encoded audio and video stream data. The processor 20 transmits the audio and video stream data to the portable information terminal 10 via the network interface 26 and the communication network 5.

  Next, a more specific operation of the portable information terminal 10 will be described based on a flowchart.

  FIG. 5 shows an example of a flowchart of processing for generating and transmitting noise power information related to a sound including ambient noise, which is executed by the portable information terminal 10.

  Referring to FIG. 5, in step 502, the acoustic input unit 144 of the portable information terminal 10 records a sound signal representing a sound captured by the microphone 142 in a storage device (not shown). In this case, the sound signal is captured before reproduction of the audio signal and recorded as noise representing ambient noise. The sound input unit 144 subtracts the reproduction signal from the sound reproduction unit 136 from the sound signal captured by the microphone 142 when the reception audio signal is reproduced by the sound reproduction unit 136 and sound is emitted from the speaker 138. A noise component may be obtained and recorded as noise representing ambient noise.

In step 504, the transform unit 146 transforms the recorded sound signal x _in (n) including noise in the time domain into a spectrum X (f) in the frequency domain, for example, by fast Fourier transform (FFT). Here, x _in (n) represents an input signal of each sample n, and X (f) represents an input spectrum of each frequency bin (section) f.

The conversion unit 146 further converts the frequency spectrum X (f) into, for example, a frequency power spectrum S (f) represented by the following expression.
S (f) = | X (f) ² |

ここで、ノイズ電力ｎｏｉｓｅ＿ｐｏｗ（ｓｆｂ）は、各部分帯域ｓｆｂ毎のグループ化された電力の総和を表し、ｓｆｂは部分帯域の番号を表し、ｓｆｂｏｆｆｓｅｔ（ｎｏ）は部分帯域のオフセット番号ｎｏを表す。それによって、測定部１４８は、各部分帯域ｓｆｂ毎にグループ化されたノイズ電力を生成する。 In step 506, the measurement unit 148 divides the frequency power spectrum S (f) into a plurality of partial bands sfb, and obtains noise power noise_pow (sfb) of each partial band sfb expressed by the following equation, for example.

Here, the noise power noise_pow (sfb) represents the sum of the grouped power for each partial band sfb, sfb represents the number of the partial band, and sfboffset (no) represents the offset number no of the partial band. Thereby, the measurement unit 148 generates noise power grouped for each partial band sfb.

  In step 510, the transmission unit 150 transmits noise power information including the noise power noise_pow (sfb) of each partial band sfb to the information processing apparatus 20 via the communication network 5 together with the stream transmission request or periodically. On the other hand, the information processing apparatus 20 receives the noise power information, and the transcoder apparatus 200 of the information processing apparatus 20 corrects the initial masking threshold 302 based on the noise power noise_pow (sfb).

  Next, a more specific operation of the transcoder device 200 will be described based on a flowchart.

  FIGS. 6A and 6B show processes for performing audio coding of the decoded video stream data according to the noise power level of the portable information terminal 10, which is executed by the acoustic coding unit 230 of the transcoder device 200. An example of a flowchart is shown.

ここで、ｋは周波数ビンを表し、ＮはＭＤＣＴ変換の窓（例えば、窓長２０４８または２５６サンプル）を表し、ｎ_０＝（Ｎ／２＋１）／２である。 Referring to FIG. 6A, in step 606, the signal conversion unit 236 converts the time-domain decoded audio signal x _in decoded by the acoustic decoding unit 226 into a frequency-domain spectrum X (f). For example, the signal conversion unit 236 performs a conversion MDCT (modified discrete cosine transform) on the decoded audio signal x _in to obtain, for example, a frequency spectrum mdct (k) of the following equation.

Here, k represents a frequency bin, N represents an MDCT transform window (for example, window length 2048 or 256 samples), and n ₀ = (N / 2 + 1) / 2.

ここで、ｋは、例えば、１０２４個のビン番号０〜１０２３を表していてもよい。また、部分帯域ｓｆｂは、例えば、５２個の部分帯域番号０〜５１を表していてもよい。 In step 608, the masking threshold value generation unit 238 divides the frequency band of the plurality of frequency bins k into a plurality of partial bands sfb, and for example, represents each partial band sfb of the spectrum mdct (k) represented by the following equation: The frequency spectrum power mdct_pow (sfb) is calculated.

Here, k may represent, for example, 1024 bin numbers 0 to 1023. The partial band sfb may represent, for example, 52 partial band numbers 0 to 51.

  In step 610, the masking threshold value generator 238 calculates or obtains an initial masking threshold value 302 over the band of the audible frequency f according to the auditory characteristic based on the distribution characteristics of the power of each partial band sfb. The masking threshold 302 is expressed by, for example, allowed_pow (sfb) for each partial band sfb.

At this time, each masking threshold value of each input signal x _in may be obtained, and a small value or a large value among the masking threshold values of each partial band sfb may be selected as the masking threshold value. Further, as the masking threshold value of each input signal x _in , for example, the power in the minimum audible range of each partial band sfb may be used simply. For example, “New Implementation Techniques of an Efficient MPEG Advanced Audio Coder” (Non-Patent Document 1) is known as a document relating to a known method for calculating a more accurate masking threshold.

In step 612, the masking threshold correction unit 239 combines the initial masking threshold 302 with the noise power level 304 (noise_pow (sfb)) generated by the portable information terminal 10 to form a combined masking threshold 306. The combined masking threshold 306 is expressed by, for example, the following new_allowed_pow (sfb) for each partial band sfb.
If allowed_pow (sfb) ≧ noise_pow (sfb),
new_allowed_pow (sfb) = allowed_pow (sfb)
It becomes.
If allowed_pow (sfb) <noise_pow (sfb),
new_allowed_pow (sfb) = noise_pow (sfb)
It becomes.

  The composite masking threshold value 306 represents a power level threshold value of the audio signal that is effective for auditory reproduction quality in the presence of ambient noise in the portable information terminal 10. That is, in the partial band where the masking threshold 302 is higher than the noise power level 304, when the power of the audio signal is less than or equal to the combined masking threshold 306, the person hardly hears the audio signal due to the auditory characteristics. In addition, in a partial band where the noise power level 304 is higher than the masking threshold 302, when the power of the audio signal is equal to or lower than the combined masking threshold 306, a person hardly hears the audio signal due to ambient noise. Therefore, an audio signal having a composite masking threshold value 306 or less in a certain partial band may not be encoded and transmitted. On the other hand, audio signals larger than the composite masking threshold 306 can be heard in a certain partial band. Accordingly, if the audio signal of the partial band that contributes to maintaining or improving the sound quality is level-corrected so as to be larger than the synthesis masking threshold 306 and the level-corrected audio signal is encoded and transmitted, the decoded and reproduced audio signal is transmitted. The signal can be heard with high sound quality.

Referring to FIG. 6B, in step 620, the masking threshold correction unit 239 further corrects the combined masking threshold 306 with three elements or factors f _{1 to} f ₃ to generate a corrected masking threshold 310. The first element f ₁ is a correction amount for the combined masking threshold 306 according to the transmission data rate or the bit rate. The second factor f ₂ is the correction amount of the combined masking threshold 306 according to the signal to masking threshold ratio or gain ratio for the audio signal 330 and the initial masking threshold 302. The third element f ₃ is a correction amount of the synthetic masking threshold 306 corresponding to the peak of the power of the audio signal. Correction according to an element _f 1 ~f ₃ for each sub-band, the synthesis masking threshold 306 by pulling respect elements _f 1 ~f _3, and generates a corrected masking threshold 310.

  By using the corrected masking threshold value 310 obtained by correcting the composite masking threshold value 306 and lowering the level, in audio coding, it is possible to increase the partial band of the audio signal that contributes to the improvement of auditory reproduction sound quality as an encoding target. .

Next, correction by the _three elements f _{1 to} f ₃ for increasing the number of encoding targets effective for maintaining or improving the auditory reproduction sound quality will be described more specifically.

  FIG. 7 shows an example of a more specific flowchart of step 620 in FIG. 6B.

8A to 8C show examples of _three- element correction functions f ₁ , f ₂ and f ₃ .
FIG. 9 is a diagram for explaining an example of the peak nature of the spectrum of an audio signal.

Referring to FIG. 7, in step 626, the masking threshold correction unit 239 calculates a correction amount f ₁ corresponding to the transmission bit rate. The correction amount f ₁ is expressed as a function f ₁ (bit rate) of the transmission bit rate bit rate, for example, as shown in FIG. 8A. Therefore, as the transmission bit rate increases, the correction amount f ₁ generally increases in a certain bit rate range higher than the threshold value BR _TH . Thereby, it is possible to increase the partial band to be encoded within the range of the estimated marginal number of bits according to the transmission bit rate.

  In step 628, the masking threshold correction unit 239 calculates the gain ratio R of the power (330) mdct_pow (sfb) of the audio signal and the initial masking threshold (302) allowed_pow (sfb) in each partial band sfb. Here, the gain ratio is represented by R = mdct_pow (sfb) / allowed_pow (sfb).

In step 630, the masking threshold correction unit 239 calculates the correction amount f ₂ based on the ratio R. The correction amount f ₂ is expressed as a function f ₂ (R) of the ratio R, for example, as shown in FIG. 8B. Therefore, according to the gain ratio R of the audio signal power mdct_pow (sfb) and the masking threshold allowed_pow (sfb) is increased, generally the correction amount _{f 2} is increased in the range of the threshold value _{R TH} higher certain ratio. Accordingly, a partial band having a large gain ratio between the power 330 of the audio signal and the masking threshold 302 is easily selected as an encoding target.

In step 632, the masking threshold correction unit 239 calculates the peak power of the audio signal in each partial band sfb. The peak peak (sfb_x) of the audio signal power mdct_pow (sfb) in a certain partial band sfb_x can be calculated based on the gain ratio or the average of the gain difference with the power of the adjacent partial band. For example, in FIG. 9, the peak peak (sfb_x) of a certain partial band sfb_x with respect to the adjacent partial bands sfb_x-1 and sfb_x + 1 is expressed by the following equation based on the average value of the gain ratio.
peak (sfb_x)
= {Mdct_pow (sfb_x) / mdct_pow (sfb_x-1)
+ Mdct_pow (sfb_x) / mdct_pow (sfb_x + 1)} / 2

In step 634, the masking threshold correction portion 239 calculates the correction amount _{f 3} based on the peak properties peak (sfb_x). The correction amount f ₃ is expressed as a function f ₂ (peak) of the peak peak (sfb_x), for example, as shown in FIG. 8C. Thus, in accordance with the peak of peak increases, generally the correction amount _{f 3} increases at a certain peak of the range from the threshold value _{P TH.} This makes it easier to select a partial band of an audio signal having a high peak as compared with an adjacent partial band as an encoding target.

In step 636, the masking threshold correction unit 239 corrects the composite masking threshold 306 based on the composite correction amount α (sfb). The correction masking threshold 310 is expressed by the following equation, for example.
Correction masking threshold 310 = composite masking threshold 306 × correction amount α (sfb)
Here, the combined correction amount α (sfb) by the _three elements f _{1 to} f ₃ is expressed by the following equation, for example.
α (sfb) = − (f ₁ × f ₂ × f ₃ )

Therefore, the corrected masking threshold 310 obtained by correcting the composite masking threshold 306, that is, new_allowed_pow (sfb) is expressed by the following equation, for example.
new_allowed_pow (sfb)
= New_allowed_pow (sfb) × α (sfb)
However, the corrected masking threshold 310 is never smaller than the initial masking threshold 302, and therefore satisfies the following equation.
Left side new_allowed_pow (sfb) ≧ allowed_pow (sfb)
Next, the masking threshold correction unit 239 supplies the correction masking threshold 310 of each partial band to the masking correction unit 244.

  Referring to FIG. 6B again, in step 640, the encoding determination unit 240 compares the power 330 of the audio signal with the correction masking threshold 310 for each partial band sfb. Next, the encoding determination unit 240 determines the powers 350 and 351 of the audio signal in the partial band sfb having a power larger than the correction masking threshold 310 as an encoding target. That is, for each partial band sfb, the power mdct_pow (sfb) of the audio signal is selected as an encoding target when it is larger (>) than the correction masking threshold new_allowed_pow (sfb). On the other hand, the encoding determination unit 240 determines that the power 352 of the audio signal in the partial band having power equal to or less than the correction masking threshold 310 is not to be encoded.

  In step 642, the encoding determination unit 240 sets an amplification factor γ for level correction or amplification so that the power 350 and 351 of each subband larger than the correction masking threshold 310 is larger than the combined masking threshold 306. decide. On the other hand, the encoding determination unit 240 determines the amplification factor γ of the power 336 of the audio signal in each partial band sufficiently larger than the synthesis masking threshold 306 as 1 so as not to amplify. Next, the level correction unit 242 corrects and amplifies the power 350 and 351 of the partial band of the audio signal according to each amplification factor γ.

More specifically, the power mdct_pow (sfb) of the audio signal is amplified so as not to be buried in ambient noise, for example, according to the following equation.
For the partial band sfb,
mdct_pow (sfb) = mdct_pow (sfb) × gain (sfb)
Here, the gain gain (sfb) may be expressed by the following equation, for example.
gain (sfb) = (noise_pow (sfb) / mdct_pow (sfb)) × γ,
Here, for example, the coefficient γ may be 1.2.
In this gain gain (sfb) expression, the power of the audio signal is level-corrected to the same level as the noise power by the ratio of noise_pow (sfb) / mdct_pow (sfb), and further amplified by a coefficient γ to be larger than the noise power. The
Further, the amplification of the power of each bin in the group power of the partial band sfb is expressed by the following equation.
mdct_pow (k) = mdct_pow (k) × gain (sfb)

  In step 644, the masking correction unit 244 operates in cooperation with the quantization unit 246 to correct the corrected masking threshold 310 received from the masking threshold correction unit 239 so as to match the transmission rate. On the other hand, the quantization unit 246 quantizes the power of the audio signal of each partial band or frequency bin whose level is corrected by the level correction unit 242 based on the masking threshold corrected by the masking correction unit 244.

  In step 646, the multiplexing unit 248 multiplexes the plurality of subband audio stream data generated by the quantization unit 246 to generate audio stream data that is audio-coded. In this way, the audio signal decoded by the acoustic decoding unit 226 is adaptively converted into low-rate data in a form that is hard to be buried in noise by the acoustic encoding unit 230 according to the ambient noise level of the portable information terminal 10. Is encoded or converted.

Next, by correcting the composite masking threshold 306 based on another factor f _SN in addition to the _three factors f ₁ -f ₃ in addition to the _three factors f ₁ -f 3 according to a variation of the above embodiment, A method for generating the correction masking threshold 310 will be described. In this case, the reproduction audio signal in the portable information terminal 10 and the signal-to-noise (SN) ratio related to the ambient noise are used to correct the combined masking threshold 306 to generate the corrected masking threshold 310.

  FIG. 10 shows an example of another schematic configuration of the portable information terminal 10 according to a modification of the above-described embodiment.

  In FIG. 10, the measurement unit 148 is coupled to the decoding unit 132 and receives the decoded audio signal or its power spectrum from the decoding unit 132. Other coupling relationships in the portable information terminal 10 are the same as those in FIG.

  In the portable information terminal 10 of FIG. 10, the receiving unit 130 receives the encoded audio stream data from the information processing apparatus 20, and the decoding unit 132 generates the audio signal by decoding the encoded audio stream data. To do. The sound reproducing unit 136 amplifies and reproduces the audio signal and supplies the amplified signal to the speaker 138 to generate sound.

  On the other hand, the conversion unit 146 converts the sound signal captured by the microphone 142 into a frequency spectrum. In this case, the captured sound signal includes a noise signal component of ambient noise and a reproduced audio signal component generated by the speaker 138. The measurement unit 148 calculates the frequency spectrum of each partial band of the decoded audio signal. In addition, the measurement unit 148 calculates the spectrum of the partial band of the captured sound signal. Next, the measurement unit 148 may calculate the spectrum of the noise signal component by subtracting the spectrum of the audio signal component from the spectrum of the captured sound signal for each partial band. Next, the measurement unit 148 calculates a signal-to-noise ratio based on each spectrum of the audio signal component and the noise signal component. Next, the transmission unit 150 transmits noise power information including noise power and a signal-to-noise ratio to the information processing apparatus 20 together with the stream transmission request or periodically.

  As an alternative, the portable information terminal 10 does not generate the decoded audio signal at the speaker 138, and the measurement unit 148 is based on the noise signal spectrum of the captured ambient noise and the calculated spectrum of the decoded audio signal. The signal-to-noise ratio may be calculated. At this time, the spectrum of the decoded audio signal may be calculated by correcting the spectrum of the audio signal based on the acoustic characteristics when the decoded audio signal is reproduced by the speaker 138.

  In this case, referring to FIG. 2 again, in the transcoder device 200, the information input unit 210 inputs noise power information including noise power and a signal-to-noise ratio and supplies the noise power information to the acoustic encoding unit 230. The noise power information including the noise power and the signal-to-noise ratio is supplied to the masking threshold correction unit 239 and the encoding determination unit 240. The masking threshold value correction unit 239 of the transcoder device 200 of the information processing device 20 further corrects the combined masking threshold value 306 based on the signal-to-noise ratio in the portable information terminal 10.

  FIG. 11 is a modification of the flowchart of FIG. 5 and shows an example of a flowchart of processing for generating and transmitting noise power and a signal-to-noise ratio, which is executed by the portable information terminal 10.

  Referring to FIG. 11, in step 502, the acoustic input unit 144 of the portable information terminal 10 records the sound signal captured by the microphone 142 in the storage device. In this case, the sound signal captured before the reproduction of the audio signal is recorded as noise, and the sound signal including the reproduced sound signal component during reproduction of the decoded audio signal and the noise component of the ambient sound is a mixed sound signal. Is recorded in the storage device.

Step 504 is similar to that of FIG. That is, the converting unit 146 converts the sound signal x _in (n) _in the time domain into the spectrum X (f) in the frequency domain, and further converts it into the frequency power spectrum S (f).

In step 506, the measurement unit 148 divides the frequency power spectrum S (f) into a plurality of partial bands sfb, and the noise power noise_pow (sfb) and the reproduced sound signal of each partial band sfb, as in the case of FIG. The power signal_pow (sfb) is obtained. Here, the reproduction sound signal power signal_pow (sfb) represents the power of the decoded audio signal during reproduction. In this case, the reproduction sound signal power signal_pow (sfb) may be obtained as a calculated estimated value based on the audio signal decoded by the decoding unit 132. At this time, the reproduction sound signal power signal_pow (sfb) may be calculated by correcting acoustic characteristics of the portable information terminal 10 such as the speaker 138 and the microphone 142. Further, the reproduced sound signal power signal_pow (sfb) is obtained by subtracting the noise power noise_pow (sfb) captured at a different time from the power sound_pow (sfb) of the mixed sound signal captured at a certain time. It may be sought. Further, the noise power noise_pow (sfb) may be obtained by subtracting the reproduced sound signal power signal_pow (sfb) obtained by calculation from the power sound_pow (sfb) of the captured mixed sound signal. The noise power noise_pow (sfb) and the reproduction sound signal power signal_pow (sfb) of each partial band sfb are expressed by the following equations, for example.

In step 508, the measurement unit 148 calculates the signal-to-noise ratio SN_ratio (sfb) of each partial band based on the noise power noise_pow (sfb) and the reproduction sound signal power signal_pow (sfb) of each partial band. The signal-to-noise ratio SN_ratio (sfb) is expressed by the following equation, for example.
SN_ratio (sfb)
= Signal_pow (sfb) / noise_pow (sfb)

  In step 510, the transmission unit 150 obtains noise power information including the noise power noise_pow (sfb) and the signal-to-noise (SN) ratio SN_ratio (sfb) of each partial band sfb as in the case of FIG. Send to. In this case, the transcoder device 200 of the information processing device 20 corrects the initial masking threshold 302 based on the noise power noise_pow (sfb), and further combines the masking threshold based on the signal-to-noise (SN) ratio SN_ratio (sfb). 306 is corrected.

  As an alternative, the signal power signal_pow (sfb) and the signal-to-noise ratio in each partial band may be calculated by the masking threshold correction unit 239 of the transcoder device 200 based on the noise power transmitted from the portable information terminal 10. Good. In this case, the signal-to-noise ratio may be calculated based on the signal obtained by delaying the audio signal encoded by the acoustic encoding unit 230 and the noise power received from the portable information terminal 10.

  The flowcharts of FIGS. 6A and 6B are similarly executed by the acoustic encoding unit 230 of the transcoder apparatus 200 with respect to the flowchart executed by the portable information terminal 10 of FIG.

FIG. 12 is a modification of the flowchart of FIG. 7 and shows an example of another flowchart of the process for correcting the composite masking threshold 306 with the element f _SN and the three elements f _{1 to} f ₃ .

Referring to FIG. 12, in step 622, the masking threshold value correction unit 239 determines that the signal-to-noise (S / N) ratio (gain ratio) in the portable information terminal 10 is less than 1, for example (<1). The correction amount f _SN is calculated according to the (S / N) ratio. The correction amount f _SN is a correction amount represented by the following equation for correcting the composite masking threshold 306 in accordance with the signal-to-noise (S / N) ratio.
f _SN = f _SNM = 1 / SN_ratio (sfb)
In this case, the masking threshold correction unit 239 corrects as follows, as will be described later.
new_allowed_pow (sfb)
= New_allowed_pow (sfb) × f _SNM
= New_allowed_pow (sfb) / SN_ratio (sfb)

As an alternative, the correction amount f _SN may be a correction amount represented by the following equation.
f _SN = f _SNA = new_allowed_pow (sfb) × (1 / SN_ratio (sfb) −1)
In this case, the masking threshold correction unit 239 corrects as follows, as will be described later.
The corrected composite masking threshold (306) new_allowed_pow (sfb) is expressed by the following equation, for example.
new_allowed_pow (sfb)
= New_allowed_pow (sfb) × f _SNA
= New_allowed_pow (sfb) × 1 / SN_ratio (sfb)

  As described above, when the signal-to-noise ratio of the audio signal is less than 1, the composite masking threshold 306 is raised, so that the level correction amount of the audio signal in the partial band that contributes to the improvement of the auditory reproduction sound quality is buried in the noise. Can be enlarged to be difficult.

Steps 626 to 634 are the same as those in FIG.
In step 637, the masking threshold correcting portion 239 corrects the synthesized masking threshold 306 based on the correction amount _{f SN} and combined correction amount α (sfb). The correction masking threshold 310 is expressed by the following equation, for example.
Correction masking threshold 310 = synthesis masking threshold 306 × correction amount f _SNM × correction amount α (sfb), or correction masking threshold 310 = synthesis masking threshold 306 × correction amount f _SNA × correction amount α (sfb)
Here, as described above, the combined correction amount α (sfb) = − (f ₁ × f ₂ × f ₃ ). In this case, the correction amount f ₁ corresponding to the transmission data rate may tend to increase as the correction amount f _SN increases.

Accordingly, the corrected masking threshold 310 obtained by further correcting the composite masking threshold 306, that is, new_allowed_pow (sfb) is expressed by the following equation, for example.
new_allowed_pow (sfb)
= New_allowed_pow (sfb) × correction amount _{f SNM} × _α (sfb), or = new_allowed_pow (sfb) + correction amount _{f SNA} × _α (sfb)
However, the corrected masking threshold 310 is never smaller than the initial masking threshold 302, and therefore satisfies the following equation.
Left side new_allowed_pow (sfb) ≧ allowed_pow (sfb)
Next, the masking threshold correction unit 239 supplies the correction masking threshold 310 of each partial band to the masking correction unit 244.

  Next, the time domain analysis of the audio signal converted into the frequency domain in the transcoder device 200 according to the signal-to-noise ratio of the reproduced audio signal of the portable information terminal 10 according to still another modification of the above-described embodiment. A method for controlling the length of the window will be described.

  As for audio signal encoding, the analysis length is shortened for a signal that shows a sharp change in the time domain, such as an attack sound, and pre-echo noise is suppressed. However, since the encoding efficiency per unit time decreases, the length of the analysis length is appropriately adjusted according to the input sound. For example, in AAC or the like, switching is performed between two types of windows: a short window of 128 points and a long window of 1024 points.

  In the following case, the length of the analysis window in the time domain of the audio signal converted into the frequency domain in the transcoder 200 is adjusted or switched according to the signal-to-noise ratio. When the signal-to-noise ratio is low, pre-echo noise is hardly perceived even if the analysis window is lengthened, so there is no problem even if the analysis window is lengthened. Increasing the analysis length increases the number of bits that can be used per partial band per unit time, improving coding efficiency, or reducing the amount of bits used, and that bit in another unit time (frame). By using numbers, the sound quality can be improved.

  FIG. 13 is a modification of the transcoder device 200 of FIG. 2, and illustrates another example of a schematic configuration of the transcoder device 200.

  In this case, the transcoder device 200 includes, for example, an acoustic encoding unit 232 instead of the acoustic encoding unit 230 of FIG. Similar to the acoustic encoding unit 230 in FIG. 2, the acoustic encoding unit 232 includes a signal conversion unit 236, a masking threshold value generation unit 238, a masking threshold value correction unit 239, an encoding determination unit 240, and a level correction unit 242. . The acoustic encoding unit 232 includes a masking correction unit 244, a quantization unit 246, and a multiplexing unit 248, similarly to the acoustic encoding unit 230 of FIG. In this case, the acoustic encoding unit 232 further includes a block switching control unit 232 and a block switching unit 234.

  Even if the block switching control unit 232 receives the signal-to-noise ratio of each partial band from the information input unit 210 or based on the noise power received from the information input unit 210 and the power of the audio signal of the signal conversion unit 236, The signal-to-noise ratio of the band may be calculated. Here, the signal-to-noise ratio is the signal-to-noise ratio of the reproduced audio signal in the portable information terminal 10.

The block switching control unit 232 causes the block switching unit 234 to switch the block length of the input data of the signal conversion unit 236 supplied from the acoustic decoding unit 226 according to the signal-to-noise (SN) ratio in the portable information terminal 10. . For example, when the signal-to-noise (SN) ratio of x partial bands or more is equal to or less than the threshold SN _th (SN_ratio (sfb) ≦ SN _th ), the block switching control unit 232 forcibly increases the length of the block. (For example, 1024 points). Accordingly, the audio encoding unit 232 has encoded audio data per unit time when the signal-to-noise (SN) ratio is lower than the threshold SN _th and the audio signal is difficult to hear, compared to when the block length is short. The number of bits can be increased. As a result, a larger number of bits can be assigned to one partial band, and a larger number of bits can be allocated to a partial band audio signal that contributes to an effective improvement in auditory sound quality.

On the other hand, when the signal-to-noise ratio (SN_ratio (sfb)) is higher (>) than the threshold SN _th , the signal conversion unit 236 uses the signal conversion unit 236 to convert the input data according to the characteristics of the input signal as usual. Convert to long or short block frequency domain data. In this case, for example, the window length may be 2048, the long block may be 1024 points, and the short block may be 128 points. The signal conversion unit 236 converts a signal showing a steep change in the time domain, such as an attack sound, into short block frequency domain data. Thereby, temporal resolution is improved and pre-echo reduction is improved.

  FIGS. 14A and 14B are flowcharts of processing for audio-coding the decoded video stream data according to the noise power and signal-to-noise ratio of the portable information terminal 10 executed by the acoustic encoding unit 232. An example is shown. The flowchart of FIGS. 14A and 14B is a variation of the flowchart of FIGS. 6A and 6B.

Referring to FIG. 14A, in step 602, the block switching control unit 232 determines whether or not the signal-to-noise ratio at the portable information terminal 10 is greater than the threshold (SN_ratio (sfb)> SN _th ) for each partial band. . In step 604, the block switching control unit 232 forces the block according to the signal-to-noise ratio of each partial band when the signal-to-noise ratio of the number threshold of x number or more is larger than the noise ratio threshold. Control to be longer. In other cases, the block switching control unit 232 causes the signal conversion unit 236 to control the block length in a normal form.

  Steps 606-612 and steps 620-644 are similar to those of FIGS. 6A and 6B.

  In the above-described embodiment, the power 332, 334, 338, 340 of the audio signal partial band is corrected by the level correction unit 242 of the acoustic encoding units 230, 232. As an alternative, the audio encoding units 230 and 232 encode the audio signal without amplifying the power, and the portable information terminal 10 receives the encoded audio data received and decoded according to the user's reproduction operation. You may amplify so that it may become larger than the noise power level of noise. However, in this case, the quantization error tends to increase.

  In the above-described embodiment, the masking threshold value 302 is corrected to the masking threshold value 306 by the masking threshold value correction unit 239 of the acoustic encoding units 230 and 232, and further corrected to the masking threshold value 310. Used at 244. As an alternative, the encoding determination unit 240 selects powers 350 and 531 of each partial band whose difference from the correction masking threshold 302 is greater than a certain gain threshold, and the level correction unit 242 uses the powers 350 and 351 as noise. The level may be corrected so as to be higher than the power level 306. In this case, the masking correction unit 244 corrects the masking threshold 306 instead of the correction masking threshold 310 for the quantization in the quantization unit 246, and the quantization unit 246 uses the level-corrected power of each partial band. The quantization may be performed according to the masking correction unit 246.

  Thus, according to the embodiment, the acoustic encoding units 230 and 232 select and increase the power of the partial band that contributes to the maintenance or improvement of the auditory sound quality within the allowable range of the transmission rate, and the auditory encoding. Therefore, the auditory reproduction sound quality of the audio signal in the portable information terminal 10 can be improved. In addition, according to the embodiment, the acoustic encoding units 230 and 232 selectively encode information effective for ensuring auditory reproduction sound quality in accordance with ambient noise in the portable information terminal 10, which is effective. The amount of transmission information can be reduced in a form that is advantageous for ensuring a good reproduction sound quality.

  All examples and conditional expressions given here are intended to help the reader understand the inventions and concepts that have contributed to the promotion of technology, such examples and It is understood that the present invention is not limited to the conditions, and that the organization of such examples in the specification is not related to the superiority or inferiority of the present invention. Although embodiments of the present invention have been described in detail, it will be understood that various changes, substitutions and variations can be made thereto without departing from the spirit and scope of the invention.

Regarding the embodiment including the above examples, the following additional notes are further disclosed.
(Supplementary note 1) An audio encoding device that encodes an audio signal using a masking threshold based on auditory characteristics so as to reduce the amount of information,
The masking threshold is corrected according to the ambient noise level received from the terminal, and at this time, the masking threshold is set such that an input signal whose difference from the masking threshold is larger than a certain threshold is larger than the corrected masking threshold. And a first processing unit for correcting at least one of the input signals;
A second processing unit that encodes the input signal after the at least one correction is performed, which is greater than the corrected masking threshold;
An audio encoding device.
(Supplementary note 2) The audio encoding device according to supplementary note 1, wherein the first processing unit corrects the masking threshold according to a signal-to-noise ratio in the terminal.
(Supplementary note 3) The supplementary note 1 or 2, wherein the first processing unit further corrects the masking threshold according to a transmission rate of the encoded audio signal to the terminal. The audio encoding device described.
(Supplementary note 4) The input signal according to any one of Supplementary notes 1 to 3, wherein the input signal is larger than the masking threshold and a gain difference between the input signal and the masking threshold is larger than the certain threshold. Audio encoding device.
(Supplementary Note 5) The first processing unit has a gain difference between an input signal larger than the masking threshold and another input signal of a partial band adjacent to the partial band of the certain input signal, or The audio encoding device according to any one of appendices 1 to 4, wherein the audio coding apparatus corrects the masking threshold so as to be smaller than the certain input signal when the gain threshold is larger than a certain gain threshold.
(Additional remark 6) The said 2nd process part further changes the length of the block of the said input signal encoded by the said 2nd process part according to the signal-to-noise ratio in the said terminal, It is characterized by the above-mentioned. The audio encoding device according to any one of appendices 1 to 5.
(Supplementary note 7) The audio encoding device according to any one of supplementary notes 1 to 6, wherein the at least one correction is performed for each partial band of the input signal.
(Supplementary note 8) A method of encoding an audio signal using a masking threshold based on auditory characteristics so as to reduce the amount of information,
The masking threshold is corrected according to the ambient noise level received from the terminal, and at this time, the masking threshold is set such that an input signal whose difference from the masking threshold is larger than a certain threshold is larger than the corrected masking threshold. And correcting at least one of the input signals,
A method in which an information processing apparatus executes a process of encoding the input signal after at least one of the corrections is performed, which is larger than the corrected masking threshold.

DESCRIPTION OF SYMBOLS 10 Mobile communication terminal 20 Information processing apparatus 200 Transcoder apparatus 210 Information input part 226 Acoustic decoding part 230, 232 Acoustic encoding part 232 Block switching control part 234 Block switching part 236 Signal conversion part 238 Masking threshold value 239 Masking threshold value correction part 240 Code Determining unit 242 Level correcting unit 244 Masking correcting unit 246 Quantizing unit 256 Image decoding unit 260 Image encoding unit 270 Multiplexing unit 280 Stream output unit

Claims (7)

An audio encoding device that encodes an audio signal using a masking threshold based on auditory characteristics so as to reduce the amount of information,
The masking threshold is corrected according to the ambient noise level received from the terminal, and at this time, the masking threshold is set such that an input signal whose difference from the masking threshold is larger than a certain threshold is larger than the corrected masking threshold. And a first processing unit for correcting at least one of the input signals;
A second processing unit that encodes the input signal after the at least one correction is performed, which is greater than the corrected masking threshold;
An audio encoding device.   The audio encoding apparatus according to claim 1, wherein the first processing unit corrects the masking threshold according to a signal-to-noise ratio in the terminal.   The audio according to claim 1 or 2, wherein the first processing unit further corrects the masking threshold according to a transmission rate of the encoded audio signal to the terminal. Encoding device.   The audio coding according to any one of claims 1 to 3, wherein the input signal is larger than the masking threshold and a gain difference between the input signal and the masking threshold is larger than the certain threshold. apparatus.   The first processing unit has a gain difference between an input signal larger than the masking threshold and another input signal of a partial band adjacent to the partial band of the certain input signal less than a certain gain threshold. 5. The audio encoding device according to claim 1, wherein when the value is larger, the masking threshold is corrected to be smaller than the certain input signal.   The second processing unit further changes a length of a block of the input signal encoded by the second processing unit according to a signal-to-noise ratio in the terminal. The audio encoding device according to any one of 1 to 5. A method of encoding an audio signal using a masking threshold based on auditory characteristics so as to reduce the amount of information,
The masking threshold is corrected according to the ambient noise level received from the terminal, and at this time, the masking threshold is set such that an input signal whose difference from the masking threshold is larger than a certain threshold is larger than the corrected masking threshold. And correcting at least one of the input signals,
A method in which an information processing apparatus executes a process of encoding the input signal after at least one of the corrections is performed, which is larger than the corrected masking threshold.

Images