AudioConv.h

//----------------------------------------------------------------------------------------------------------------------
// Simple audio conversion and resampling library
// Copyright (c) 2024 Samuel Gomes
//
// This includes heavily modified code from the following projects:
// https://github.com/cpuimage/resampler
// https://github.com/deftio/companders
// https://github.com/dystopiancode/pcm-g711
// https://web.archive.org/web/20120415044546/http://hazelware.luggle.com/tutorials/mulawcompression.html
//----------------------------------------------------------------------------------------------------------------------

#pragma once

#include <cstdint>
#include <cstdlib>

static const auto AUDIOCONV_S8_TO_F32_MULTIPLER = 1.0f / 128.0f;
static const auto AUDIOCONV_S16_TO_F32_MULTIPLER = 1.0f / 32768.0f;
static const auto AUDIOCONV_S32_TO_F32_MULTIPLER = 1.0f / 2147483648.0f;
static const auto AUDIOCONV_F32_TO_S8_MULTIPLIER = 127.0f;
static const auto AUDIOCONV_F32_TO_S16_MULTIPLIER = 32767.0f;
static const auto AUDIOCONV_F32_TO_S32_MULTIPLIER = 2147483647.0f;

/// @brief Converts unsigned 8-bit audio samples to signed 8-bit inplace.
/// @param source The input unsigned 8-bit sample frame buffer.
/// @param samples The number of samples in the sample frame buffer, where samples = frames * channels.
void AudioConv_ConvertU8ToS8(uintptr_t source, uint32_t samples)
{
    if (!source or !samples)
        return;

    auto buffer = reinterpret_cast<uint8_t *>(source);

    for (size_t i = 0; i < samples; i++)
        buffer[i] ^= 0x80; // xor_eq
}

/// @brief Converts unsigned 16-bit audio samples to signed 16-bit inplace.
/// @param source The input unsigned 16-bit sample frame buffer.
/// @param samples The number of samples in the sample frame buffer, where samples = frames * channels.
void AudioConv_ConvertU16ToS16(uintptr_t source, uint32_t samples)
{
    if (!source or !samples)
        return;

    auto buffer = reinterpret_cast<uint16_t *>(source);

    for (size_t i = 0; i < samples; i++)
        buffer[i] ^= 0x8000; // xor_eq
}

/// @brief Converts unsigned 8-bit audio samples to floating point.
/// @param src The input unsigned 8-bit sample frame buffer.
/// @param samples The number of samples in the buffer, where samples = frames * channels.
/// @param dst The output floating point sample frame buffer. The buffer size must be at least samples * sizeof(float) bytes.
void AudioConv_ConvertU8ToF32(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const uint8_t *>(src);
    auto dstBuffer = reinterpret_cast<float *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = float(int8_t(srcBuffer[i] ^ 0x80)) * AUDIOCONV_S8_TO_F32_MULTIPLER;
}

/// @brief Converts signed 8-bit audio samples to floating point.
/// @param src The input signed 8-bit sample frame buffer.
/// @param samples The number of samples in the buffer, where samples = frames * channels.
/// @param dst The output floating point sample frame buffer. The buffer size must be at least samples * sizeof(float) bytes.
void AudioConv_ConvertS8ToF32(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const int8_t *>(src);
    auto dstBuffer = reinterpret_cast<float *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = (float)srcBuffer[i] * AUDIOCONV_S8_TO_F32_MULTIPLER;
}

/// @brief Converts unsigned 8-bit audio samples to signed 16-bit.
/// @param src The input unsigned 8-bit sample frame buffer.
/// @param samples The number of samples in the buffer, where samples = frames * channels.
/// @param dst The output signed 16-bit sample frame buffer. The buffer size must be at least samples * sizeof(int16_t) bytes.
void AudioConv_ConvertU8ToS16(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const uint8_t *>(src);
    auto dstBuffer = reinterpret_cast<int16_t *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = int8_t(srcBuffer[i] ^ 0x80) << 8;
}

/// @brief Converts signed 8-bit audio samples to signed 16-bit.
/// @param src The input signed 8-bit sample frame buffer.
/// @param samples The number of samples in the buffer, where samples = frames * channels.
/// @param dst The output signed 16-bit sample frame buffer. The buffer size must be at least samples * sizeof(int16_t) bytes.
void AudioConv_ConvertS8ToS16(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const int8_t *>(src);
    auto dstBuffer = reinterpret_cast<int16_t *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = srcBuffer[i] << 8;
}

/// @brief Converts signed 16-bit audio samples to floating point.
/// @param src The input signed 16-bit sample frame buffer.
/// @param samples The number of samples in the buffer, where samples = frames * channels.
/// @param dst The output floating point sample frame buffer. The buffer size must be at least samples * sizeof(float) bytes.
void AudioConv_ConvertS16ToF32(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const int16_t *>(src);
    auto dstBuffer = reinterpret_cast<float *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = (float)srcBuffer[i] * AUDIOCONV_S16_TO_F32_MULTIPLER;
}

/// @brief Converts signed 32-bit audio samples to floating point.
/// @param src The input signed 32-bit sample frame buffer.
/// @param samples The number of samples in the buffer, where samples = frames * channels.
/// @param dst The output floating point sample frame buffer. The buffer size must be at least samples * sizeof(float) bytes.
void AudioConv_ConvertS32ToF32(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const int32_t *>(src);
    auto dstBuffer = reinterpret_cast<float *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = (float)srcBuffer[i] * AUDIOCONV_S32_TO_F32_MULTIPLER;
}

/// @brief Decodes an 8-bit signed integer using the A-Law.
/// @param number The number that will be decoded.
/// @return The decoded number.
static inline int16_t __AudioConv_DecodeALawSample(int8_t aLawByte)
{
    const static int16_t ALawDecompressTable[256] = {
        -5504, -5248, -6016, -5760, -4480, -4224, -4992, -4736, -7552,
        -7296, -8064, -7808, -6528, -6272, -7040, -6784, -2752, -2624,
        -3008, -2880, -2240, -2112, -2496, -2368, -3776, -3648, -4032,
        -3904, -3264, -3136, -3520, -3392, -22016, -20992, -24064, -23040,
        -17920, -16896, -19968, -18944, -30208, -29184, -32256, -31232, -26112,
        -25088, -28160, -27136, -11008, -10496, -12032, -11520, -8960, -8448,
        -9984, -9472, -15104, -14592, -16128, -15616, -13056, -12544, -14080,
        -13568, -344, -328, -376, -360, -280, -264, -312, -296,
        -472, -456, -504, -488, -408, -392, -440, -424, -88,
        -72, -120, -104, -24, -8, -56, -40, -216, -200,
        -248, -232, -152, -136, -184, -168, -1376, -1312, -1504,
        -1440, -1120, -1056, -1248, -1184, -1888, -1824, -2016, -1952,
        -1632, -1568, -1760, -1696, -688, -656, -752, -720, -560,
        -528, -624, -592, -944, -912, -1008, -976, -816, -784,
        -880, -848, 5504, 5248, 6016, 5760, 4480, 4224, 4992,
        4736, 7552, 7296, 8064, 7808, 6528, 6272, 7040, 6784,
        2752, 2624, 3008, 2880, 2240, 2112, 2496, 2368, 3776,
        3648, 4032, 3904, 3264, 3136, 3520, 3392, 22016, 20992,
        24064, 23040, 17920, 16896, 19968, 18944, 30208, 29184, 32256,
        31232, 26112, 25088, 28160, 27136, 11008, 10496, 12032, 11520,
        8960, 8448, 9984, 9472, 15104, 14592, 16128, 15616, 13056,
        12544, 14080, 13568, 344, 328, 376, 360, 280, 264,
        312, 296, 472, 456, 504, 488, 408, 392, 440,
        424, 88, 72, 120, 104, 24, 8, 56, 40,
        216, 200, 248, 232, 152, 136, 184, 168, 1376,
        1312, 1504, 1440, 1120, 1056, 1248, 1184, 1888, 1824,
        2016, 1952, 1632, 1568, 1760, 1696, 688, 656, 752,
        720, 560, 528, 624, 592, 944, 912, 1008, 976,
        816, 784, 880, 848};

    return ALawDecompressTable[uint8_t(aLawByte)];
}

/// @brief Converts A-Law encoded audio samples to signed 16-bit samples.
/// @param src Pointer to the A-Law encoded audio samples buffer.
/// @param samples Number of samples in the buffer, where samples = frames * channels.
/// @param dst Pointer to the buffer where the signed 16-bit samples will be stored. The buffer size must be at least samples * sizeof(int16_t) bytes.
void AudioConv_ConvertALawToS16(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const int8_t *>(src);
    auto dstBuffer = reinterpret_cast<int16_t *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = __AudioConv_DecodeALawSample(srcBuffer[i]);
}

/// @brief Converts A-Law encoded audio samples to floating point samples.
/// @param src Pointer to the A-Law encoded audio samples buffer.
/// @param frames Number of samples in the buffer.
/// @param dst Pointer to the buffer where the floating point samples will be stored. The buffer size must be at least samples * sizeof(float) bytes.
void AudioConv_ConvertALawToF32(uintptr_t src, uint32_t frames, uintptr_t dst)
{
    if (!src or !dst or !frames)
        return;

    auto srcBuffer = reinterpret_cast<const int8_t *>(src);
    auto dstBuffer = reinterpret_cast<float *>(dst);

    for (size_t i = 0; i < frames; i++)
        dstBuffer[i] = (float)__AudioConv_DecodeALawSample(srcBuffer[i]) * AUDIOCONV_S16_TO_F32_MULTIPLER;
}

/// @brief Decodes an 8-bit signed integer using the mu-Law.
/// @param number The number that will be decoded.
/// @return The decoded number.
static inline int16_t __AudioConv_DecodeMuLawSample(int8_t uLawByte)
{
    static const int16_t MuLawDecompressTable[256] = {
        -32124, -31100, -30076, -29052, -28028, -27004, -25980, -24956,
        -23932, -22908, -21884, -20860, -19836, -18812, -17788, -16764,
        -15996, -15484, -14972, -14460, -13948, -13436, -12924, -12412,
        -11900, -11388, -10876, -10364, -9852, -9340, -8828, -8316,
        -7932, -7676, -7420, -7164, -6908, -6652, -6396, -6140,
        -5884, -5628, -5372, -5116, -4860, -4604, -4348, -4092,
        -3900, -3772, -3644, -3516, -3388, -3260, -3132, -3004,
        -2876, -2748, -2620, -2492, -2364, -2236, -2108, -1980,
        -1884, -1820, -1756, -1692, -1628, -1564, -1500, -1436,
        -1372, -1308, -1244, -1180, -1116, -1052, -988, -924,
        -876, -844, -812, -780, -748, -716, -684, -652,
        -620, -588, -556, -524, -492, -460, -428, -396,
        -372, -356, -340, -324, -308, -292, -276, -260,
        -244, -228, -212, -196, -180, -164, -148, -132,
        -120, -112, -104, -96, -88, -80, -72, -64,
        -56, -48, -40, -32, -24, -16, -8, -1,
        32124, 31100, 30076, 29052, 28028, 27004, 25980, 24956,
        23932, 22908, 21884, 20860, 19836, 18812, 17788, 16764,
        15996, 15484, 14972, 14460, 13948, 13436, 12924, 12412,
        11900, 11388, 10876, 10364, 9852, 9340, 8828, 8316,
        7932, 7676, 7420, 7164, 6908, 6652, 6396, 6140,
        5884, 5628, 5372, 5116, 4860, 4604, 4348, 4092,
        3900, 3772, 3644, 3516, 3388, 3260, 3132, 3004,
        2876, 2748, 2620, 2492, 2364, 2236, 2108, 1980,
        1884, 1820, 1756, 1692, 1628, 1564, 1500, 1436,
        1372, 1308, 1244, 1180, 1116, 1052, 988, 924,
        876, 844, 812, 780, 748, 716, 684, 652,
        620, 588, 556, 524, 492, 460, 428, 396,
        372, 356, 340, 324, 308, 292, 276, 260,
        244, 228, 212, 196, 180, 164, 148, 132,
        120, 112, 104, 96, 88, 80, 72, 64,
        56, 48, 40, 32, 24, 16, 8, 0};

    return MuLawDecompressTable[uint8_t(uLawByte)];
}

/// @brief Converts mu-Law encoded audio samples to signed 16-bit samples.
/// @param src Pointer to the mu-Law encoded audio samples buffer.
/// @param samples Number of samples in the buffer, where samples = frames * channels.
/// @param dst Pointer to the buffer where the signed 16-bit samples will be stored. The buffer size must be at least samples * sizeof(int16_t) bytes.
void AudioConv_ConvertMuLawToS16(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const int8_t *>(src);
    auto dstBuffer = reinterpret_cast<int16_t *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = __AudioConv_DecodeMuLawSample(srcBuffer[i]);
}

/// @brief Converts mu-Law encoded audio samples to floating point samples.
/// @param src Pointer to the mu-Law encoded audio samples buffer.
/// @param samples Number of samples in the buffer.
/// @param dst Pointer to the buffer where the floating point samples will be stored. The buffer size must be at least samples * sizeof(float) bytes.
void AudioConv_ConvertMuLawToF32(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src or !dst or !samples)
        return;

    auto srcBuffer = reinterpret_cast<const int8_t *>(src);
    auto dstBuffer = reinterpret_cast<float *>(dst);

    for (size_t i = 0; i < samples; i++)
        dstBuffer[i] = (float)__AudioConv_DecodeMuLawSample(srcBuffer[i]) * AUDIOCONV_S16_TO_F32_MULTIPLER;
}

/// @brief Converts 4-bit ADPCM compressed audio samples to 8-bit signed samples.
/// @param src Pointer to the ADPCM compressed audio samples buffer.
/// @param srcLen The number of bytes in the input buffer.
/// @param compTab Pointer to the compression table used to decode the ADPCM codes.
/// @param dst Pointer to the buffer where the 8-bit signed samples will be stored. The buffer size must be at least srcLen * 2 bytes.
void AudioConv_ConvertADPCM4ToS8(uintptr_t src, uint32_t srcLen, const char *compTab, uintptr_t dst)
{
    auto srcBuffer = reinterpret_cast<const uint8_t *>(src);
    auto dstBuffer = reinterpret_cast<int8_t *>(dst);

    int8_t delta = 0;

    for (size_t i = 0; i < srcLen; i++)
    {
        delta += compTab[*srcBuffer & 0x0F];
        *(dstBuffer++) = delta;
        delta += compTab[(*srcBuffer >> 4) & 0x0F];
        *(dstBuffer++) = delta;
        srcBuffer++;
    }
}

/// @brief Converts a dual mono audio buffer to a stereo interleaved audio buffer.
/// @tparam T Data type of the audio samples.
/// @param src Pointer to the dual mono audio buffer.
/// @param samples Number of samples in the buffer.
/// @param dst Pointer to the buffer where the stereo interleaved audio samples will be stored. The buffer size must be at least samples * sizeof(T) bytes.
template <typename T>
void AudioConv_ConvertDualMonoToStereo(uintptr_t src, uint32_t samples, uintptr_t dst)
{
    if (!src || !dst || samples < 4)
        return;

    auto srcBuffer = reinterpret_cast<const T *>(src);
    auto dstBuffer = reinterpret_cast<T *>(dst);

    uint32_t halfLength = samples >> 1;

    for (size_t i = 0, j = 0; i < halfLength; i++, j += 2)
    {
        dstBuffer[j] = srcBuffer[i];
        dstBuffer[j + 1] = srcBuffer[halfLength + i];
    }
}

// Specializations of AudioConv_ConvertDualMonoToStereoInterleaved() for different data types
#define AudioConv_ConvertDualMonoToStereoS8(_src_, _samples_, _dst_) AudioConv_ConvertDualMonoToStereo<int8_t>(_src_, _samples_, _dst_)
#define AudioConv_ConvertDualMonoToStereoS16(_src_, _samples_, _dst_) AudioConv_ConvertDualMonoToStereo<int16_t>(_src_, _samples_, _dst_)
#define AudioConv_ConvertDualMonoToStereoF32(_src_, _samples_, _dst_) AudioConv_ConvertDualMonoToStereo<float>(_src_, _samples_, _dst_)

/// @brief Resamples an audio buffer. Set output to NULL to get the output buffer size in samples frames.
/// @tparam T The sample data type.
/// @param src The input sample frame buffer.
/// @param dst The output sample frame buffer.
/// @param inSampleRate The input sample rate.
/// @param outSampleRate The output sample rate.
/// @param inputSize The number of samples frames in the input.
/// @param channels The number of channels for both input and output.
/// @return The number of samples frames written to the output.
template <typename T>
uint64_t AudioConv_Resample(uintptr_t src, uintptr_t dst, int inSampleRate, int outSampleRate, uint64_t inputSize, uint32_t channels)
{
    auto input = reinterpret_cast<const T *>(src);

    if (!input)
        return 0;

    auto outputSize = (uint64_t)(inputSize * (double)outSampleRate / (double)inSampleRate);
    outputSize -= outputSize % channels;

    auto output = reinterpret_cast<T *>(dst);

    if (!output)
        return outputSize;

    auto stepDist = ((double)inSampleRate / (double)outSampleRate);
    const uint64_t fixedFraction = (1LL << 32);
    const double normFixed = (1.0 / (1LL << 32));
    auto step = ((uint64_t)(stepDist * fixedFraction + 0.5));
    uint64_t curOffset = 0;

    for (uint32_t i = 0; i < outputSize; i += 1)
    {
        for (uint32_t c = 0; c < channels; c += 1)
        {
            *output++ = static_cast<T>(input[c] + (input[c + channels] - input[c]) * ((double)(curOffset >> 32) + ((curOffset & (fixedFraction - 1)) * normFixed)));
        }
        curOffset += step;
        input += (curOffset >> 32) * channels;
        curOffset &= (fixedFraction - 1);
    }

    return outputSize;
}

// Specializations of AudioConv_Resample() for different data types
#define AudioConv_ResampleS16(_src_, _dst_, _src_sample_rate_, _dst_sample_rate_, _src_size_, _channels_) AudioConv_Resample<int16_t>(_src_, _dst_, _src_sample_rate_, _dst_sample_rate_, _src_size_, _channels_)
#define AudioConv_ResampleS32(_src_, _dst_, _src_sample_rate_, _dst_sample_rate_, _src_size_, _channels_) AudioConv_Resample<int32_t>(_src_, _dst_, _src_sample_rate_, _dst_sample_rate_, _src_size_, _channels_)
#define AudioConv_ResampleF32(_src_, _dst_, _src_sample_rate_, _dst_sample_rate_, _src_size_, _channels_) AudioConv_Resample<float>(_src_, _dst_, _src_sample_rate_, _dst_sample_rate_, _src_size_, _channels_)