ladybird/Userland/Libraries/LibAudio/QOATypes.h
kleines Filmröllchen 0b421a3764 LibAudio: Implement the Quite Okay Audio format
Brought to you by the inventor of QOI, QOA is a lossy audio codec that
is, as the name says, quite okay in compressing audio data reasonably
well without frequency transformation, mostly introducing some white
noise in the background. This implementation of QOA is fully compatible
with the qoa.h reference implementation as of 2023-02-25. Note that
there may be changes to the QOA format before a specification is
finalized, and there is currently no information on when that will
happen and which changes will be made.

This implementation of QOA can handle varying sample rate and varying
channel count files. The reference implementation does not produce these
files and cannot handle them, so their implementation is untested.

The QOA loader is capable of seeking in constant-bitrate streams.

QOA links:
https://phoboslab.org/log/2023/02/qoa-time-domain-audio-compression
https://github.com/phoboslab/qoa
2023-03-10 04:07:14 -07:00

143 lines
4.8 KiB
C++

/*
* Copyright (c) 2023, kleines Filmröllchen <filmroellchen@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#pragma once
#include <AK/Array.h>
#include <AK/Forward.h>
#include <AK/Math.h>
#include <AK/Types.h>
#include <math.h>
namespace Audio::QOA {
// 'qoaf'
static constexpr u32 const magic = 0x716f6166;
static constexpr size_t const header_size = sizeof(u64);
struct FrameHeader {
u8 num_channels;
u32 sample_rate; // 24 bits
u16 sample_count;
// TODO: might be removed and/or replaced
u16 frame_size;
static ErrorOr<FrameHeader> read_from_stream(Stream& stream);
};
static constexpr size_t const frame_header_size = sizeof(u64);
// Least mean squares (LMS) predictor FIR filter size.
static constexpr size_t const lms_history = 4;
static constexpr size_t const lms_state_size = 2 * lms_history * sizeof(u16);
// Only used for internal purposes; intermediate LMS states can be beyond 16 bits.
struct LMSState {
i32 history[lms_history] { 0, 0, 0, 0 };
i32 weights[lms_history] { 0, 0, 0, 0 };
LMSState() = default;
LMSState(u64 history_packed, u64 weights_packed);
i32 predict() const;
void update(i32 sample, i32 residual);
};
using PackedSlice = u64;
// A QOA slice in a more directly readable format, unpacked from the stored 64-bit format.
struct UnpackedSlice {
size_t scale_factor_index; // 4 bits packed
Array<u8, 20> residuals; // 3 bits packed
};
// Samples within a 64-bit slice.
static constexpr size_t const slice_samples = 20;
static constexpr size_t const max_slices_per_frame = 256;
static constexpr size_t const max_frame_samples = slice_samples * max_slices_per_frame;
// Defined as clamping limits by the spec.
static constexpr i32 const sample_minimum = -32768;
static constexpr i32 const sample_maximum = 32767;
// Quantization and scale factor tables computed from formulas given in qoa.h
constexpr Array<int, 17> generate_scale_factor_table()
{
Array<int, 17> scalefactor_table;
for (size_t s = 0; s < 17; ++s)
scalefactor_table[s] = static_cast<int>(AK::round<double>(AK::pow<double>(static_cast<double>(s + 1), 2.75)));
return scalefactor_table;
}
// FIXME: Get rid of the literal table once Clang understands our constexpr pow() and round() implementations.
#if defined(AK_COMPILER_CLANG)
static constexpr Array<int, 17> scale_factor_table = {
1, 7, 21, 45, 84, 138, 211, 304, 421, 562, 731, 928, 1157, 1419, 1715, 2048
};
#else
static constexpr Array<int, 17> scale_factor_table = generate_scale_factor_table();
#endif
constexpr Array<int, 17> generate_reciprocal_table()
{
Array<int, 17> reciprocal_table;
for (size_t s = 0; s < 17; ++s) {
reciprocal_table[s] = ((1 << 16) + scale_factor_table[s] - 1) / scale_factor_table[s];
}
return reciprocal_table;
}
constexpr Array<Array<int, 8>, 16> generate_dequantization_table()
{
Array<Array<int, 8>, 16> dequantization_table;
constexpr Array<double, 8> float_dequantization_table = { 0.75, -0.75, 2.5, -2.5, 4.5, -4.5, 7, -7 };
for (size_t scale = 0; scale < 16; ++scale) {
for (size_t quantization = 0; quantization < 8; ++quantization)
dequantization_table[scale][quantization] = static_cast<int>(AK::round<double>(
static_cast<double>(scale_factor_table[scale]) * float_dequantization_table[quantization]));
}
return dequantization_table;
}
#if defined(AK_COMPILER_CLANG)
static constexpr Array<int, 17> reciprocal_table = {
65536, 9363, 3121, 1457, 781, 475, 311, 216, 156, 117, 90, 71, 57, 47, 39, 32
};
static constexpr Array<Array<int, 8>, 16> dequantization_table = {
Array<int, 8> { 1, -1, 3, -3, 5, -5, 7, -7 },
{ 5, -5, 18, -18, 32, -32, 49, -49 },
{ 16, -16, 53, -53, 95, -95, 147, -147 },
{ 34, -34, 113, -113, 203, -203, 315, -315 },
{ 63, -63, 210, -210, 378, -378, 588, -588 },
{ 104, -104, 345, -345, 621, -621, 966, -966 },
{ 158, -158, 528, -528, 950, -950, 1477, -1477 },
{ 228, -228, 760, -760, 1368, -1368, 2128, -2128 },
{ 316, -316, 1053, -1053, 1895, -1895, 2947, -2947 },
{ 422, -422, 1405, -1405, 2529, -2529, 3934, -3934 },
{ 548, -548, 1828, -1828, 3290, -3290, 5117, -5117 },
{ 696, -696, 2320, -2320, 4176, -4176, 6496, -6496 },
{ 868, -868, 2893, -2893, 5207, -5207, 8099, -8099 },
{ 1064, -1064, 3548, -3548, 6386, -6386, 9933, -9933 },
{ 1286, -1286, 4288, -4288, 7718, -7718, 12005, -12005 },
{ 1536, -1536, 5120, -5120, 9216, -9216, 14336, -14336 },
};
#else
static constexpr Array<int, 17> reciprocal_table = generate_reciprocal_table();
static constexpr Array<Array<int, 8>, 16> dequantization_table = generate_dequantization_table();
#endif
static constexpr Array<int, 17> quantization_table = {
7, 7, 7, 5, 5, 3, 3, 1, // -8 ..-1
0, // 0
0, 2, 2, 4, 4, 6, 6, 6 // 1 .. 8
};
}