MP3 file format analysis

MP3 encoding

Signal Description:

(1) MP3 encoding input signal: PCM (Pulse Code Modulation) sound signal, some audio files in .wav format are PCM signals.

( 2 ) MP3 encoded output signal: transmission in MP3 format

WAV format file capacity = (sampling frequency X quantization number of bits X channel) X time / 8 (byte = 8 bits). When the 2 bytes of 14H~15H have a value of 1 , it indicates the PCM encoding format of the data bit , which can be used as the input of the MP3 encoder .

3. Analysis of the SHINE program
SHINE is a MP coding program written in C language, consisting of 11 source files in total. Add the source file to the newly created VC purchase mode console application to run, but it must use command line mode when running.

 

1. File data structure
A config_t structure type is defined in types.h , and a global variable configuration is initialized with it, which is equivalent to an ” object ” in an object-oriented language, and is used to encode data and parameters throughout the process. coding Save and manage

Define a wave_t structure type to store PCM pulse format file information, and use wave_t to define the wave variable in config_t , which stores the MP3 encoded source information as the input to the MP3 encoder.

A type of structure mpeg_t is defined which is used to store MP3 encoded information, and the mpeg variable is defined in config_t with mpeg_t , and the information stored in this variable is output as MP3 encoded parameter information.

typedestruct {

time_tstart_time;

char*infile;

round_wave;

char* output file;

mpeg_tmpeg;

} config_t;

The above structure is mainly used to store the ” header ” information , and the byte stream entity information after encoding the output is stored in the bs structure ( defined in the bitstream.h file ), and the bs structure is defined as

static structure

MP3 file format analysis

1. Summary of MP3 and MPEG files

An MP3 file is made up of frames, and a frame is the smallest unit of an MP3 file. The full name of MP3 must be MPEG1 Layer-3 audio file.

MPEG (MovingPictureExperts Group), MPGE audio layer refers to the sound part of the MPGE file, which is divided into three layers based on the quality and complexity of the encoding, namely Layer-1, Layer2 and Layer3, corresponding to MP1, MP2 and MP3 format files.

2. Structure of MP3 files
MP3 files are divided into 3 parts : TAG_V2(ID3V2 ) , Frame, TAG_V1(ID3V1) .

( 1 ) Frame format

The frame header is 4 bytes and its structure is as follows

typedef FrameHeader

{

unsigned intsync: 11; // synchronization information

unsigned intversion: 2; // version

unsigned intlayer: 2; // layer

unsigned intprotection: 1; // CRC check

unsigned intbitrate: 4; // Bit rate

unsigned intfrequency: 2; // sample rate

unsigned intpadding: 1; // adjust frame length

unsigned intprivate: 1; // reserved word

unsigned intmode: 2; // channel mode

unsigned int mode extension: 2; // extended mode

unsigned intcopyright: 1; // Copyright

unsigned original: 1; // original logo

unsigned inemphasis: 2; // emphasis mode

}

HEADER, *LPHEADER;

Each frame takes 26 ms to play, regardless of the length of the frame. The length of MAIN_DATA is

Length(MAIN_DATA) = ((version == MPEG1)?144:72) * bitrate/frequency + padding;

( 2 ) ID3V1 format

ID3V1 is stored at the end of the MP3 file, a total of 128 Bytes, all information is stored sequentially and the insufficient part is filled with ‘\0’, which can be opened and viewed with UltraEdit.

typedef tagID3V1 structure

{

char header[3];

char Title[30];

artist char[30];

album char[30];

char Year[4];

char Comment[28];

coal reserve;

character track;;

charGenus;

}

ID3V1,*pID3V1;

( 3 ) ID3V2 format

ID3V2 is stored in the header of the MP3 file and consists of a tag header and several tag frames.

The tag header is 10 bytes,

char header[3];

char see;

character review;

char Flag;

character size [4];

Each tag frame consists of a 10-byte frame header and at least one byte of variable-length content. The frame header is defined as follows:

char frame ID[4];

character size [4];

char Flags[2];

How does an mp3 file work?

What is the structure of an mp3?

An MP3 song has three
versions: 96 Kbps (96 kilobits per second), 128 Kbps and 192 Kbps. Kbps (bit rate), which indicates the amount of music data per second,
the higher the Kbps value, the better the sound quality, and the larger the file, the MP3 standard stipulates that an MP3 file with a constant bit rate is called CBR, and most of the
MP3 files are CBR, and MP3 file with changing bit rate is called VBR, and the length of each FRAME can be changed. The following are
the differences between CBR and VBR:
1) CBR: The size of the FRAME with a fixed bitrate is fixed (the formula is as above), as long as the total length of the file and the length of the frame are known, mp3 can be calculated from the 26ms needed to play each frame. The total playback time can also be monitored by counting the number of frames to control operations such as fast forward, fast rewind, and slow playback. Note: Sometimes not all frames are the same length and some frames may be one or more bytes longer.
2) VBR: VBR is an algorithm released by XING company, so there will be “XING” keyword in the MP3 FRAME (many popular
Small software can also perform VBR compression (it is not known if they comply with this agreement), it is stored in the first valid FRAME in the MP3 file and identifies that the MP3 file is VBR. At the same time, the first FRAME stores the total number of FRAMES of the MP3 file, which makes it easy to get the total playing time, and at the same time, there are 100 bytes to store the FRAME INDEX of 100 times segments of the total playing time. . Suppose a 4 minute MP3 song 240S is divided into 100 segments, and the time difference between two adjacent INDEXes is 2.4S, so through this INDEX as long as some FRAMES are processed before and then we can quickly find the FRAME header we need to fast forward. Table 2 Explanation of the byte of
structure of the first frame of the VBR 1-4 file The same standard sound frame header as CBR 5-40 Save the VBR file logo “Xing” (58 69 6E 67), the specific location of this logo depends on the adopted standard MPEG and the sound depends on the channel mode. The leading and trailing bytes of the flag are not used. 36-39 MPEG-1 and non-mono (common) 21-24 MPEG-1 and mono 21-24 MPEG-2 and non-mono 13-16 MPEG-2 and mono 41-44 Flags, indicates whether the frame number, Se stores information about file length, directory table, and VBR scale, and if so, 01 02 04 08. 45-48 frame number (including first frame) 49-52 file length 53-152 file table directory, used for byte positioning according to time. 153-156 VBR scale for bit rate changes

How does an mp3 file work?

What is the structure of an mp3?

MP3 file is composed of frame (frame), frame is the smallest composition unit of MP3 file. MP3 full name should be MPEG1 Layer 3 audio files. MPEG
(Motion Picture Experts Group) translates into Chinese as Moving Picture Experts Group, and refers specifically to moving video and audio compression standards.
MPEG1 standard, also known as MPEG audio layer, which is divided into three layers based on compression quality and encoding complexity, namely,
Layer-1, Layer2 and Layer3, which correspond to the three sound files of MP1, MP2 and MP3 respectively, and use different
levels of audio files according to different purposes. The higher the MPEG audio encoding level, the more complex the encoder and the higher the compression ratio. The compression ratios of MP1 and MP2 are 4:1 and
6:1-8:1 respectively, while the compression ratio of MP3 is as high as 10:1-8:1. 12:1, meaning one minute of CD-quality music requires 10MB
of storage space without compression, but only about 1 MB after MP3 compression encoding. However, MP3 uses a lossy compression method for audio signals.

Low sound distortion, MP3 adopts “sensory coding technology”, that is, when encoding, the audio file is first analyzed for frequency spectrum, and then the noise level is filtered by a filter, and then each bit remaining is sparse and arranged by quantization, and finally form an MP3 file with a higher compression ratio, and make the file
compressed achieve a sound effect closer to the original sound source when played back.

2. The complete structure of the file
MP3 The MP3 file is roughly divided into three parts: TAG_V2 (ID3V2), Frame, TAG_V1 (ID3V1) ID3V2
contains
information like author, composer, album, etc. The length is not fixed, which expands the information volume of ID3V1.
frame
… Frame A series
of frames, the number is determined by the file size and the length of the frame
. The length of each FRAME may not be fixed, or it may be fixed. It is determined by the bit rate. Each FRAME is divided into two parts: frame header and data entity. The frame header records information such as bit rate, sample rate, and mp3 version, and each frame is independent of each other. ID3V1 contains information like author, composer, album, etc. ., and the length is 128BYTE. 3. MP3 Frame Format 1. Frame Header Format The frame header is 4 bytes long. For fixed bitrate MP3 files, the frame header format of all frames is the same. The data structure is as follows: typedef FrameHeader { unsigned int sync: 11; // unsigned synchronization information int version: 2; // version