JP2816052B2 - Audio data compression device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for recording audio data on an optical disk, and more particularly to a method for compressing audio data.

[0002]

2. Description of the Related Art A magazine "Nikkei Electronics" published by Nikkei McGraw-Hill, Inc., December 9, 1991, pp. 155 to 168, entitled "Starting Commercialization of Minidiscs, Pursuing the Leading DCC" A new audio system called a system has been disclosed.

Hereinafter, a normal recording operation of the mini disc system will be described with reference to FIG. First, the audio signal of two channels is transmitted to the A / A
Frequency 4 same as compact disc by D converter
It is digitized by being sampled at 4.1 KHz and converted to audio data. The generated audio data is input to the data compression means 2. In the data compression means 2, ATRAC (Adaptive Transform A
coustic coding) and data compression.

The specific operation of the data compression means 2 employing this data compression method will be described below. The audio data input to the data compression means 2 is about 11.6.
The signal is divided in units of m seconds (one sound block), and is first input to the band division filter 21. Band division filter 2
5 is shown in FIG. The audio data input to the band division filter 21 is divided into block data of three bands. Note that the low band of the three bands is DC to 5 KH.
z, 5 kHz to 10 KHz for the middle band, and 10 KHz or higher for the high band. The generation order of the low / medium / high block data will be described with reference to FIG. First, the audio data input to the band division filter 21 is constituted by a digital filter. First two-part filter 21
1 is divided into high frequency block data and middle / low frequency block data. The middle / low-frequency block data is further divided into middle-frequency and low-frequency block data by a second two-division filter 212. The previously generated high-frequency block data is supplied to a second two-partition filter 212.
Is input to the delay circuit 213 in order to match the timing of the high-frequency block data with the delay time of the middle / low-frequency block data caused by the data division processing by the.

[0005] Each block data generated by dividing audio data into three bands as described above is a first DC data.
T unit 221, second DCT unit 222, third DCT unit 2
23 performs a predetermined DCT transform to generate spell vector data. DCT conversion is one of the methods for converting time axis data to frequency axis data, and is often used when audio data or image data is converted to spell spectrum data.

The generated spell spectrum data of each band is input to the frequency data processing means 23. Frequency data selection means 23 in frequency data processing means 23 of FIG.
In step 1, data compression processing is performed based on the minimum audibility characteristic and the auditory masking effect. When the sound becomes lower than a certain level, it becomes inaudible to human ears,
The level is different depending on the frequency.
Using this characteristic, the frequency data selection unit 231 first extracts only the frequency components of the spell spectrum data generated by the three DCT units that reach a level at which a human can hear, and extracts the frequencies that do not reach the level at which they can be heard. Reduce ingredients. Furthermore, the data is reduced using the auditory masking effect. The auditory masking effect is that when a loud sound and a loud sound are generated at close frequencies, the loud sound is covered by the loud sound and becomes inaudible. By using this effect, the data amount of the spell spectrum data is further compressed by reducing the frequency components that have reached the level that can be heard by the person mentioned above and that cannot be heard due to the large frequency components of the adjacent levels. I do.

The encoding means 232 performs an encoding process on the spectrum data generated by the above-mentioned compression by synthesizing the spectrum data of each block and information on each data amount, and performs about 5% of the audio data amount.
Generate one-third compressed data. The compressed data thus generated by the frequency data processing unit 23 is
The data is sent to the memory control unit 3 and sequentially stored in the memory 4 according to the address specified by the memory control unit 3. The amount of data in the memory 4 is detected by the memory control unit 3, and the information is sent to the recording command unit 82 in the system controller 8. When detecting that the storage data amount has reached a predetermined recording unit (1 cluster = 176 sound blocks), the recording command unit 82 sends a recording command to the memory control unit 3. As a result, the memory control unit 3 reads the storage data of the recording unit from the memory 4 and applies the CD-ROM encoder 5 to the read compressed data.
And a predetermined encoding process by the CD-DA encoder 6 to generate compressed data for recording. The compressed data for recording is supplied to the magnetic head 7 and is recorded by the magnetic head 7 at a predetermined track position of the disk D irradiated by an optical pickup (not shown).

[0008]

When the frequency spectrum of a human voice is examined, the level around 1 KHz is high, and the level gradually becomes lower in a higher frequency range. Therefore, when recording only human voice, frequency data of about 1 KHz is indispensable, and data in a higher frequency range than that is less important. For example, when recording only the voice of a person such as a conference or a lecture, the high-frequency components may sometimes become noise.

The present invention has been made in view of the above, and when recording only human voice, a frequency component sufficiently higher than the frequency component of the voice is not recorded, thereby reducing noise. It is an object of the present invention to realize an audio data recording method for performing reduced and clear recording by adding only a small configuration to an apparatus adopting the above-described audio data compression method called the ATRAC method.

[0010]

In order to solve the above-mentioned problems, the present invention drives only the frequency converter that generates spell spectrum data included in an audio signal band among a plurality of the frequency converters. A first feature is that a band limiting unit is provided, and a second feature is that a band limiting unit that causes the band dividing unit to generate only block data included in the audio signal band is provided. A third feature is that a band limiting means for instructing the frequency data processing means to perform only the compression processing on the spell spectrum data included in the data is provided.

[0011]

The band of audio data to be recorded is limited by the band limiting means, and only the necessary band is recorded.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the embodiments shown in FIGS. In the embodiments described later, the present invention is applied to the above-described conventional technology.
The same components as those of the related art are denoted by the same reference numerals, and redundant description is omitted.

A feature of the present invention resides in that, when only a person's voice is recorded exclusively, such as in a conference or a lecture, only block data in a band including the voice is recorded. <First Embodiment> First, in a first embodiment shown in FIG. 1, an operator operates a recording key and a recording band limit key included in the operation unit 9. In conjunction with this operation, the band limiting means 81 constituted by software in the system controller 8 operates, and sends a band limiting signal to the second DCT unit 222 and the third DCT unit 223. However, the band division filter 21 operates normally to generate each block data, and the corresponding block data is supplied to the three DCT units 221, 222, and 223, respectively. But the second
The operation of the DCT unit 222 and the third DCT unit 223 is stopped. As a result, the first DC for inputting low-frequency data
Only the T unit 221 supplies the spell spectrum data to the frequency data processing unit 23. Therefore, the above-described frequency data processing unit 23 processes only the input low frequency spell spectrum data.

The band-limited signal is also supplied to a frequency data processing means 23. When this band limited signal is input, the frequency data processing means 23
A compression process of multiplexing the band information indicating that no other spectrum data exists in the spell spectrum data generated by 21 is performed to generate compressed data. This low-frequency band data is audio data of 5 KHz or less as described above. Considering that the frequency distribution of human voice is highest at around 1 KHz, it is sufficient when recording only human voice. Bandwidth. Therefore, clear recording with less noise is possible by not recording the middle and high frequency components containing many noise components unrelated to human voice.

As described above, the frequency data processing means 2
3 compresses the audio data input to the data compression means 2 to a data amount of about 1/5. The data amount of the compressed data generated by the frequency data processing means 23 is about one-fifth of the audio data. The amounts are the same. Therefore, if only the low-frequency block data is processed by the frequency data processing means 23 in the configuration as described above, the resolution of the data is increased, and the deterioration of the data caused by the data compression can be reduced. <Second Embodiment> A second embodiment will be described with reference to FIG. The second embodiment is an embodiment in which the band limiting unit 81 limits the derivation of the block data generated from the band division filter 21. In this case, the audio data input to the band splitting means 21 is first input to the first two-division filter 211, where it is split into the high-frequency component of the above-described band and the frequency component combining the low-band and the mid-band. You. In this case, since the recording band limiting key included in the operation unit 9 has been operated, the band limiting unit 81 generates a band limiting signal. The band limiting unit 81 supplies the band limited signal to the first two-division filter 211 shown in FIG. 5, and stops the output of the high-frequency block data. As a result, no high-frequency block data is output. However, the low / mid band block data is input to the second two-division filter 212.

The band limiting means 81 also stops the output of the block data of the middle band in the second two-part filter 212, and outputs only the block data of the low band from the second two-part filter 212. Will be. In addition. The processing performed below is the same as in the first embodiment, and redundant description will be omitted. <Third Embodiment> A third embodiment will be described with reference to FIG. The third embodiment is an example in which the band limiting unit 81 limits the band of the spectral data to be selected to the frequency selecting unit 231 in the frequency processing unit 23. In this case, the spell spectrum data generated by each of the three DCT units 221, 222, and 223 is input to the frequency processing unit 23. Normally, the input spell spectrum data is selected by the frequency selection unit 231 only for the frequency components of a specific level or higher as described above, and is transmitted to the encoding unit 232.

However, in this embodiment, since the recording band limiting key included in the operation unit 9 is operated, the band limiting unit 81 operates to generate the band limiting signal, and the frequency selecting unit for inputting the band declaration signal. 231 processes only low-frequency block data (frequency components in the range of DC to 5 KHz). As a result, only the low-frequency processed data is sent to the encoding means 232. Therefore, the encoding process of the encoding unit 232 is performed only on the low-frequency processed data. The processing performed below is common to the above-described embodiments.

[0018]

As described above, according to the present invention, there is an extremely large effect that a clear recording can be performed when only a person's voice such as a conference or a lecture is recorded by adding only a small configuration to the conventional technology. .

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a first embodiment of the present invention.

FIG. 2 is a functional block diagram of a second embodiment of the present invention.

FIG. 3 is a functional block diagram of a third embodiment of the present invention.

FIG. 4 is a diagram illustrating a conventional technique.

FIG. 5 is a detailed view of a band dividing means 21 according to the related art.

FIG. 6 is a detailed diagram of a conventional frequency data processing unit 23.

[Explanation of symbols]

 2 Data Compression Unit 21 Band Division Filter 221 First DCT Unit 222 Second DCT Unit 223 Third DCT Unit 23 Frequency Data Processing Unit 231 Frequency Data Selection Unit 232 Encoding Unit 8 System Controller 81 Band Limiting Unit

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Masato Fuuma 2-18-18 Keihanhondori, Moriguchi-shi, Osaka Sanyo Electric Co., Ltd. (72) Inventor Miyuki Okamoto 2-18 Keihanhondori, Moriguchi-shi, Osaka (56) References JP-A-62-265810 (JP, A) JP-A-63-3528 (JP, A) JP-A-4-227338 (JP, A) JP-A-4-230799 (JP, A) JP-A-5-183523 (JP, A)

Claims (3)

(57) [Claims] 1. A band dividing means for dividing audio data into a plurality of frequency bands and generating block data corresponding to each band, and inputting the block data as time axis data to spectrum data as frequency axis data. Audio data compression apparatus comprising: a plurality of frequency converters for converting the spectrum data; and a frequency data processing means for receiving the spectrum data from the plurality of frequency converters and performing a predetermined compression process to generate a fixed amount of compressed data. 5. The audio data compression unit according to claim 1, further comprising a band limiting unit that drives only the frequency converter that generates spectrum data included in an audio signal band among the plurality of the frequency converters. 2. Band division means for dividing audio data into a plurality of frequency bands to generate block data corresponding to each band, and inputting the block data as time axis data to spectrum data as frequency axis data. Audio data compression apparatus comprising: a plurality of frequency converters for converting the spectrum data; and a frequency data processing means for receiving the spectrum data from the plurality of frequency converters and performing a predetermined compression process to generate a fixed amount of compressed data. 5. The audio data compression unit according to claim 1, further comprising a band limitation unit that causes the band division unit to generate only block data included in an audio signal band. 3. Band division means for dividing audio data into a plurality of frequency bands to generate block data corresponding to each band, and inputting said block data as time axis data to spectrum data as frequency axis data. Audio data compression apparatus comprising: a plurality of frequency converters for converting the spectrum data; and a frequency data processing means for receiving the spectrum data from the plurality of frequency converters and performing a predetermined compression process to generate a fixed amount of compressed data. 5. The audio data compression means according to claim 1, further comprising a band limiting means for instructing said frequency data processing means to perform only the compression processing on the spectrum data included in the audio signal band.