WAV audio file format is a Microsoft and IBM audio file format standard for storing audio on PC.

It's a variant of the Resource Interchange File Format (RIFF). This file format is generic meta-format for storing data in tagged chunks. It is clone of older Audio Interchange File Format (AIFF) used on Macintosh computers. The only difference is how write/read multi-byte integers, in little-endian format or in big-endian format.

RIFF

RIFF files consist of a simple header followed by "chunks".
Header:

• 4 bytes - the ASCII identifier "RIFF"
• 4 bytes - an unsigned 32-bit integer with the size of overall file (except 8 bytes)
• 4 bytes - an ASCII identifier for this particular file type (such as "WAVE")
  

Chunk:

• 4 bytes - an ASCII identifier for this chunk ("fmt " or "data")
• 4 bytes - an unsigned 32-bit integer with the size of this chunk (expect 8 bytes)
• Variable-sized field, the size given in the previous field
• A pad byte, if the chunk's size is not even

WAVE

The WAVE format is a subset of RIFF specification. A WAVE file is often just a RIFF file with "WAVE" chunk which consists of two sub-chunk:

• "fmt " chunk - specifying the data format
• "data" chunk - containing the actual sample data.
  

 

This chunk contains information about how the wave data is stored and should be played back including the type of compression used, number of channels, sample rate, etc.

 

• 4 bytes - ChunkID. ChunkID is always "fmt "
• 4 bytes - ChunkDataSize. This is the size the rest of the "fmt " sub-chunk
• 2 bytes - CompressionCode. It specifies the type of compression used on the wave data. A value of 1 means there is no compression (Pulse Code Modulation (PCM) code)
• 2 bytes - NumberOfChannels. The number of channels specifies how many separate audio signals that are encoded in the wave "data" chunk. A value of 1 means a mono signal, a value of 2 means a stereo signal, etc.
• 4 bytes - SampleRate. The number of sample slices per second (44100, 48000, 96000, ...)
• 4 bytes - AverageBytesPerSecond. This value indicates how many bytes of wave data must be streamed to a D/A converter per second in order to play the wave file. This value can be easily calculated with the formula: AverageBytesPerSecond = BlockAlign * SampleRate
• 2 bytes - BlockAlign. The number of bytes per sample slice including all channels.  This value can be calculated by the formula: BlockAlign = SignificantBitsPerSample * NumberOfChannels/8
• 2 bytes - SignificantBitsPerSample. This value specifies the number of bits used to define each sample. (Usually 8, 16, 24, 32 bits)
• 2 bytes - ExtraFormatBytes. It doesn't exist if the CompressionCode is 1 (uncompressed PCM file)

 

 

"data" sub-chunk

The "data" sub-chunk contains the size of the data and the actual sound data.

• 4 bytes - ChunkID. ChunkID is always "data"
• 4 bytes - ChunkDataSize. This is the number of bytes in the data. If CompressionCode=1 (PCM), this value can be calculated by the formula: ChunkDataSize = NumberOfSamples * BlockAlign
• Data. The actual sound data. If NumberOfChannels =1, and  SignificantBitsPerSample=16 you have:

1. sample0_for_channel1 (2 bytes)
2. sample1_for_channel1 (2 bytes), etc.    

    In case when NumberOfChannels =2, you have:

1. sample0_for_channel1 (2 bytes)
2. sample0_for_channel2 (2 bytes)
3. sample1_for_channel1 (2 bytes)
4. sample1_for_channel2 (2 bytes), etc.

 

Wav File Format - References:    

 

 

• "WAVE PCM soundfile format", CCRMA
• "Multiple Channel Audio Data and WAVE Files", Microsoft
• "AES standard for network and file transfer of audio - Audio-file transfer and exchange - Radio traffic audio delivery extension to the broadcast - WAVE-file format", AES 2002
• Wikipedia
• "Wave Files", SonicSpot
• Borg
• "The WAVE File Format", Lightlink