JP2001324996A - Method and device for reproducing mp3 music data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reproduce MP3 music data with spatial sound expanse and high level sound quality. SOLUTION: Based on a file B side, audio components (noise components) higher than 16 kHz are newly generated and added to the file B (hereafter, this file is called a file C). As a result, the file C can be put back to CD quality sound after a sound source is encoded and decoded to an MP3 file. In other words, the audio components above 16 kHz, that are lost during MP3 encoding and decoding, are put back to the sound quality of an MP3 CD sound source by incorporating a device which generates noise components into hardware such as an MP3 player.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to an MPEG-1 Audio La.
The present invention relates to a music data reproducing method and apparatus for reproducing music data (MP3 music data) irreversibly compressed according to the yer3 (hereinafter referred to as MP3) standard in a range close to the original sound.

[0002]

2. Description of the Related Art Conventionally, a WAV data format is a typical data standard for handling sound (music) on a computer. In the WAV data format, data can be converted to sound quality comparable to a CD (compact disk).

However, in the WAV data format, 192 KB (kilobyte) is required for recording sound (music) for one second. This capacity (192 KB) is equivalent to the amount of text data of 96,000 characters. Generally, when recording music for about 3 minutes, a capacity of about 50 MB (megabyte) is required. This was an enormous data size, which was difficult to handle on a computer.

In recent years, a compression technique called MP3 has been proposed, and it has become possible to compress the WAV data to 1/0 to 1/12.

[0005] MP3 is formally known as Motion Picture Exp.
ect Group-1 (MPEG-1) Audio Layer 3, which irreversibly compresses original music data.

Heretofore, as a means for publishing audio (music) files on the Internet, "real audio", "WAV files", and the like have conventionally been applied, but these two have advantages and disadvantages. "Real audio" can deliver audio (music) in real time, but it is far from the sound quality of a CD. On the other hand, "WAV file" has the sound quality of a CD but the file size described above. Was getting so big.

On the other hand, the reproduction sound quality of this MP3 is
As described above, the CD level can be maintained and the data size is extremely small, about 1/10 to 1/12.
Computers (including the Internet) are easy to handle.

Here, MP3 uses unnecessary frequencies (ultra high frequency range and ultra low frequency range) in order to reduce the data size.
Has been cut. As described above, the data can be reduced by using the human auditory characteristics to reduce the data size to the last minute. In such a frequency cut, basically, it is considered that a frequency in this band is hardly felt (audible) by a human, and is not a problem.

[0009]

SUMMARY OF THE INVENTION However, the applicant does not sense that the human hearing senses sound only in the frequency band that is actually felt as sound as described above, but rather the spatial spread of sound. It was found that the above cut super-treble is an important factor in order to feel the sound.

[0010] Therefore, the MP3 which has the sound quality equivalent to that of a CD
When music data is reproduced, the sound itself can be reproduced almost faithfully, but the concept of sound spread is not enough, and the sound is poor in spatial spread.

In view of the above facts, the present invention provides a method and apparatus for reproducing MP3 music data, which allows the sound to be spread spatially and reproduced at a higher level of sound quality even when reproducing MP3 music data. That is the purpose.

[0012]

SUMMARY OF THE INVENTION An MP3 music data reproducing method according to the present invention is an MPEG-1 Audio Layer 3 (hereinafter referred to as M-1).
P3) music data (MP
3 music data) in a range close to the original sound, in which a fundamental tone is extracted from the frequency component of the MP3 music data, and a frequency component to be a harmonic is generated from the fundamental tone and generated. Of the frequency components of the overtones, leave only the frequency components in the ultra-high frequency range and cut the others,
The remaining ultra-high frequency components are amplified based on the gain for the fundamental tone, a temporal change characteristic of the timbre is extracted from the MP3 music data, and based on the temporal change characteristic, the amplified ultra-high frequency component is amplified. The frequency component of the range and the original M
Synthesizes and reproduces the frequency component of the P3 music data,
It is characterized by: Further, in the present invention, the frequency component to be cut is 16 KHz or less.

An MP3 music data reproducing apparatus according to the present invention provides music data (MP3 music data) irreversibly compressed according to the MPEG-1 Audio Layer 3 (hereinafter referred to as MP3) standard.
A music data reproducing apparatus for reproducing in a range close to the original sound, an input means to which encoded MP3 music data is input, and a base tone from a frequency component of the MP3 music data input to the input means. A fundamental frequency extracting means for extracting, a harmonic frequency component generating means for generating a frequency component to be a harmonic from the fundamental frequency extracted by the fundamental frequency extracting means, and a frequency component of the harmonics generated by the harmonic frequency component generating means to MP3. A high frequency range component that exceeds a threshold value that is cut during compression, and a noise cut unit that cuts a frequency component equal to or less than the threshold value that becomes noise, and a frequency component that has been cut by the noise cut unit, Amplifying means for amplifying based on a gain for the fundamental tone,
A temporal change characteristic extracting means for extracting a temporal change characteristic of a timbre from the MP3 music data; and a temporal change characteristic of the timbre extracted by the temporal characteristic change extracting means.
Synthesizing means for synthesizing the supersonic frequency component amplified by the amplifying means and the frequency component of the original MP3 music data while synchronizing; and a synthetic MP synthesized by the synthesizing means.
And reproduction means for reproducing the three music data.

Further, according to the present invention, the threshold value is 16 KHz.

According to the above invention, the MP3 music data is
Due to lossy compression, when decompressed, it differs from the original frequency component, and in particular, the frequency components in the high frequency range (for example, 1
(6 kHz or more) is completely cut during compression, so this band is lost during decompression. This treble range is a band in which humans cannot hear, but is necessary when considering spatial expansion.

Therefore, the fundamental tone is extracted from the frequency components of the MP3 music data. In other words, most of the frequency components of the music data are complex tones in which fundamental and overtones are synthesized,
Among the components of the sound, a fundamental tone having the lowest frequency and a strong component is extracted.

As described above, since the overtone is synthesized with the fundamental tone, it is a composite tone. If the fundamental tone can be extracted, the overtone is synthesized, so that the fundamental frequency characteristic is basically the same as the original frequency characteristic. Should be able to be generated. However, since a single tone is not actually a single tone, a complex tone obtained by synthesizing the fundamental tone and the overtone in a normal frequency band becomes noise.

On the other hand, in the frequency band of 16 KHz or more, since the synthesized composite sound functions as a sound that creates a spatial spread, the other frequencies are cut off while leaving only the high frequency range frequency components. After that, the high frequency range frequency component is amplified based on the gain for the fundamental tone.

On the other hand, by extracting the temporal change characteristic of the timbre from the MP3 music data, based on the temporal change characteristic, the frequency component of the amplified supersonic range and the frequency component of the original MP3 music data are extracted. Can be synthesized while synchronizing. When this synthesized MP3 music data is reproduced, the sound pressure is present in the high frequency band required for the spread of sound, so that the sound quality can be made closer to that of a CD or the like.

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION (MP3 Music Data Generation (Compression)) The basic principle of MP3 compression is to remove useless parts of sound so that it fits in a specified bit rate. When the original music is music data stored on a CD, the bit rate is about 1400 kilobits, and by removing unnecessary parts of 9/10 from this, 128 kilobits of MP3 music data can be created.

The removal of the unnecessary portion is performed through the process shown in FIG.

The original music data (eg, music CD)
It is branched, and one is input to the sub-band decomposition unit 10,
The other is input to the Fourier transform (frequency analysis) unit 12.

The subband decomposition section 10 first resamples the original music data (for example, music CD) at a sampling frequency of 1/32. In this way, it is possible to narrow down the frequency range and perform sampling again to check the properties of the data for each frequency component included in the original music data.

On the other hand, from the result of the Fourier transform performed by the Fourier transform unit 12, the sound is discarded by other sounds and sent to a psychological auditory evaluation unit 14 for determining what is unnecessary.

The results of the subband decomposition section 10 and the psychological auditory evaluation section 14 are converted into an MDCT (Modified Discrete Cosine) section 16.
The subband-decomposed data is further converted into data in frequency units. At this stage, the digitized numerical value is converted into a notation in frequency and sent to the quantization unit 18. In the quantization unit 18, the MDCT unit 1
6 and the frequency components of the psychological auditory evaluation unit 14 are digitized, and division processing is appropriately performed for each frequency domain. That is, the quantization unit 18 determines which bits are assigned to which frequency from the result obtained by the psychological auditory evaluation.

The processing result of the quantization unit 18 is losslessly compressed in the Huffman encoding unit 20, the data is compressed, and MP3 music data is created. (Reproduction of MP3 music data) Here, the reproduction of the MP3 music data is generally performed by reproducing the data obtained by decompressing the losslessly compressed data using dedicated software or the like. The encoded MP3 music data is processed in the following procedure.

As shown in FIG. 2, in step 100, the encoded MP3 music data is fetched, and then in step 102, a fundamental tone is extracted from the frequency components of the MP3 music data, and harmonics are generated from the fundamental tone.

In the next step 104, of the frequency components of the synthesized sound (fundamental tone + overtone), only the frequency band of 16 KHz or more is left, and the rest is cut as noise. 1
In a frequency band of 6 KHz or more, the synthesized composite sound functions as a sound that creates a spatial spread.

Thereafter, in step 106, the frequency components in the high frequency range are amplified based on the gain for the fundamental tone, and then the process proceeds to step 108.

On the other hand, in step 110, the envelope is extracted simultaneously with the processing of steps 102 to 106, and in step 108, the amplified data is passed through an envelope follower circuit.
Synchronize with the original MP3 music data, and then in step 112, the amplified supersonic frequency component and the original MP3
The frequency components of the three music data are synthesized while synchronizing them (synthesized MP3 music data). The synthesized MP3 music data is reproduced by a dedicated software or the like (step 1).
14)

EXAMPLES In order to compare the original MP3 music data and the synthesized MP3 music data with the music CD data, the system shown in Table 1 below was used.

[0031]

[Table 1] In the above system, first, the CD sound source (Sample Rate
After importing 44.1KHz) to LOGIC AUDIO via Protools hardware (hereinafter this file is called file A), PE
Convert to MP3 file with AK 2.1 + Shockwave Export XTRA.

This file A is reproduced by Sound Jam MP and copied to PCM-7040 via Protools hardware.
Import to LOGIC AUDIO again via Protools hardware (this file will be referred to as file B hereinafter). The above two types of file A and file B were reproduced at the same time, and their frequency characteristics were measured with Spectra Foo.

As the largest change in the conversion to the file B, it was recognized that the sound of 16 kHz or more does not exist on the file B side. In order to make the frequency characteristics of these two types of files almost the same, DUY DSPider, MacDSP FilterBank, Waves Re
Using the naissance EQ, a new audio component (noise component) of 16 kHz or more was created and added to the file B (hereinafter, this file is referred to as file C). As a result, file C was able to return to CD quality sound even after the sound source was encoded and decoded into an MP3 file.

That is, by incorporating a device for generating a noise component created in the above experiment into hardware such as an MP3 player, the audio component of 16 KHz or more which is lost during MP3 encoding and decoding, the MP3 CD You can restore the sound quality of the sound source.

[0035]

As described above, the method and apparatus for reproducing MP3 music data according to the present invention can provide a sound with a wider space and reproduce at a higher level of sound quality even when reproducing MP3 music data. It has an excellent effect of being able to.

[Brief description of the drawings]

FIG. 1 is a control block diagram for generating MP3 music data.

FIG. 2 is a flowchart showing a procedure for adding overtone noise to a frequency band of 16 kHz or more of MP3 music data.

FIG. 3 is a frequency characteristic diagram of the same music data;
(A) CD original (file A), (B) MP3
Shown before adding noise, and (C) after adding MP3 noise.

[Explanation of symbols]

 Reference Signs List 10 Subband decomposition unit 12 Fourier transform unit 14 Psychological hearing evaluation unit 16 MDCT unit 18 Quantization unit 20 Huffman coding unit

Claims (4)

[Claims] 1. An MPEG-1 Audio Layer 3 (hereinafter, MP3)
A music data reproduction method for reproducing music data (MP3 music data) irreversibly compressed according to a standard in a range close to the original sound, comprising extracting a fundamental tone from a frequency component of the MP3 music data, Generate a frequency component that will be a harmonic, and, of the frequency components of the generated harmonic, leave only the frequency component in the ultra-high frequency range and cut off the others.The remaining frequency component in the ultra-high frequency range is used as a gain for the fundamental frequency. Amplifying based on the above, extracting a temporal change characteristic of the timbre from the MP3 music data, and based on the temporal change characteristic, the amplified frequency component of the supersonic range, the frequency component of the original MP3 music data, A method for reproducing MP3 music data, comprising: synthesizing and reproducing. 2. The MP3 music data reproducing method according to claim 1, wherein the frequency component to be cut is 16 KHz or less. 3. An MPEG-1 Audio Layer 3 (hereinafter, MP3)
A music data reproduction device for reproducing music data (MP3 music data) irreversibly compressed according to a standard in a range close to the original sound, and input means for inputting encoded MP3 music data; A fundamental tone extracting means for extracting a fundamental tone from a frequency component of the MP3 music data input to the means; a harmonic frequency component generating means for producing a frequency component to be a harmonic from the fundamental tone extracted by the fundamental tone extracting means; Noise cut means for cutting out frequency components below the threshold value which become noise while leaving high frequency range frequency components exceeding a threshold value which is cut during compression to MP3 among frequency components of harmonics generated by the generation means; Amplifying means for amplifying the frequency component cut by the noise cut means based on a gain for the fundamental tone; A temporal change characteristic extracting means for extracting a temporal change characteristic of the timbre from the MP3 music data; and a super-amplifier amplified by the amplifying means based on the temporal change characteristic of the timbre extracted by the temporal characteristic change extractor. MP3 music data reproduction comprising: synthesizing means for synthesizing frequency components of a musical range and frequency components of original MP3 music data while synchronizing; and reproducing means for reproducing the synthesized MP3 music data synthesized by the synthesizing means. apparatus. 4. The MP3 music data reproducing apparatus according to claim 3, wherein said threshold value is 16 KHz.