PcmToWav

PHP converts PCM to WAV

This extension can quickly convert PCM format sound wave files into WAV format audio files. Currently, it only provides solutions for company projects.

The extended reference is in helviojunior/WaveGenerator, special thanks to you!

Install

composer require jade/pcm-to-wav

use

PcmToWav e; $input_file = './file/test.pcm'; // 准备输入的文件 $output_file = './file/test.wav'; // 预计输出的文件 $data = PcmToWav e::init($pcm_file, $wav_file); // 调用转换 ">

use PcmToWav e P cmToWave ;

$input_file = ' ./file/test.pcm ' ; // 准备输入的文件
$output_file = ' ./file/test.wav ' ; // 预计输出的文件
$data = PcmToWav e::init( $pcm_file , $wav_file ) ; // 调用转换

Test Demo usage

进入扩展包目录
cd vendor/jade/pcm-to-wav
composer install
cd test
php Test.php

Principle introduction

What are PCM and WAV ?

PCM : PCM (Pulse Code Modulation----pulse code modulation recording). The so-called PCM recording is to convert analog signals such as sound into symbolic pulse trains and then record them. PCM signal is a digital signal composed of 1 , 0 and other symbols without any encoding or compression processing. Compared with analog signals, it is less susceptible to noise and distortion in the transmission system. The dynamic range is wide and the sound quality can be quite good.

WAV : WAV is a lossless audio file format. WAV complies with the PIFF (Resource Interchange File Format) specification. All WAV have a file header, which contains the encoding parameters of the audio stream. WAV has no hard and fast rules for encoding audio streams. In addition to PCM , almost all encodings that support ACM specification can encode WAV audio streams.

The relationship between PCM and WAV

Simply put, PCM is the original data of audio, and WAV is a container that encapsulates audio data, and its format is very simple. It just adds some header information related to the audio data at the beginning of the data.

First, let’s take a look at the format rules of WAV, as shown below

After understanding these rules, we can start coding

1. ChunkID occupies 4 bytes, fixed value "RIFF"

 $ChunkID = array(0x52, 0x49, 0x46, 0x46); // RIFF 16进制的0x52等于10进制中的82，82对应的ASCII码为R

2. ChunkSize occupies 4 bytes, and the value is 4 + (8 + SubChunk1Size) + (8 + SubChunk2Size). If the original data is PCM, it is simplified to 36 + SubChunk2Size

 $ChunkSize = array(0x0, 0x0, 0x0, 0x0);
$ChunkSize = self::getLittleEndianByteArray(36 + $dataSize);

3. Format occupies 4byte, fixed value "WAVE"

 $FileFormat = array(0x57, 0x41, 0x56, 0x45); // WAVE

4. Subchunk1ID occupies 4 bytes, fixed value "ftm" (note the space completion of 4 digits)

 $Subchunk1ID = array(0x66, 0x6D, 0x74, 0x20); // fmt

5. Subchunk1Size occupies 4byte, and when the data is PCM, the value is 16

 $Subchunk1Size = array(0x10, 0x0, 0x0, 0x0); // 16 little endian

6. AudioFormat occupies 2 bytes. When the data is PCM, the value is 1. Other values ​​indicate that the data has been compressed in some way.

 $AudioFormat = array(0x1, 0x0); // PCM = 1 little endian

7. NumChannels occupies 2 bytes, corresponding to channelConfig in AudioRecord, Mono = 1, Stereo = 2

 if ($numchannels == 2) {
    $fmt->NumChannels = array(0x2, 0x0); // 立体声为2
} else {
    $fmt->NumChannels = array(0x1, 0x0); // 单声道为1
}

8. SampleRate occupies 4 bytes, corresponding to sampleRateInHz in AudioRecord, which is the sampling frequency, such as 8000, 16000, 44100

 $SampleRate = self::getLittleEndianByteArray($samplerate);

9. ByteRate occupies 4byte, and the value is SampleRate * BlockAlign

 self::getLittleEndianByteArray($samplerate * $numchannels * ($bitspersample / 8));

10. BlockAlign occupies 2 bytes, and the value is NumChannels * BitsPerSample / 8

 self::getLittleEndianByteArray($numchannels * ($bitspersample / 8), true);

11. BitsPerSample occupies 2byte, corresponding to audioFormat in AudioRecord, 8bits = 8, 16bits = 16

 self::getLittleEndianByteArray($bitspersample, true);

12. Subchunk2ID occupies 4byte, fixed value "data", that is

 $Subchunk2ID = array(0x64, 0x61, 0x74, 0x61); // data

13. Subchunk2Size occupies 4byte, which describes the length of the audio data, which is the size of the pcm file.

 self::getLittleEndianByteArray(filesize($filename));

14. Data accounts for bytes of the pcm file size, representing the original PCM audio data.

getLittleEndianByteArray method description

getLittleEndianByteArray mainly processes the passed numbers and converts them into the data that needs to be used. The length of the array will be returned based on the number of bytes.

 private static function getLittleEndianByteArray($lValue, $short = false)
    {
        $b = array(0, 0, 0, 0);
        $running = $lValue / pow(16, 6);
        $b[3] = floor($running);
        $running -= $b[3];
        $running *= 256;
        $b[2] = floor($running);
        $running -= $b[2];
        $running *= 256;
        $b[1] = floor($running);
        $running -= $b[1];
        $running *= 256;
        $b[0] = round($running);

        if ($short) { // 为 `true` 时返回长度为2的数组
            $tmp = array_slice($b, 0, 2);
            $b = $tmp;
        }

        return $b;
    }

The first 44 bytes of information in the entire file have been basically explained. Let’s talk about the implementation of processing class files. The processing logic here first temporarily creates a file with only 44 bytes, and then appends the data of PCM file to the file. Finally, according to the WAV format rules, the actual header 44-byte information is actually calculated and the file modification pointer points to the beginning of the file, and then modified to the newly generated data.