mirror of
https://github.com/wader/fq.git
synced 2024-11-23 18:56:52 +03:00
651 lines
22 KiB
Go
651 lines
22 KiB
Go
package flac
|
|
|
|
import (
|
|
"encoding/binary"
|
|
"math/bits"
|
|
|
|
"github.com/wader/fq/format"
|
|
"github.com/wader/fq/internal/mathex"
|
|
"github.com/wader/fq/pkg/checksum"
|
|
"github.com/wader/fq/pkg/decode"
|
|
"github.com/wader/fq/pkg/interp"
|
|
"github.com/wader/fq/pkg/scalar"
|
|
)
|
|
|
|
func init() {
|
|
interp.RegisterFormat(decode.Format{
|
|
Name: format.FLAC_FRAME,
|
|
Description: "FLAC frame",
|
|
DecodeFn: frameDecode,
|
|
DecodeInArg: format.FlacFrameIn{
|
|
BitsPerSample: 16, // TODO: maybe should not have a default value?
|
|
},
|
|
})
|
|
}
|
|
|
|
const (
|
|
SampleRateStreaminfo = 0b0000
|
|
)
|
|
|
|
const (
|
|
SampleSizeStreaminfo = 0b000
|
|
)
|
|
|
|
const (
|
|
BlockingStrategyFixed = 0
|
|
BlockingStrategyVariable = 1
|
|
)
|
|
|
|
var BlockingStrategyNames = scalar.UToSymStr{
|
|
BlockingStrategyFixed: "fixed",
|
|
BlockingStrategyVariable: "variable",
|
|
}
|
|
|
|
const (
|
|
BlockSizeEndOfHeader8 = 0b0110
|
|
BlockSizeEndOfHeader16 = 0b0111
|
|
)
|
|
|
|
const (
|
|
SampeleRateEndOfHeader8 = 0b1100
|
|
SampeleRateEndOfHeader16 = 0b1101
|
|
SampeleRateEndOfHeader160 = 0b1110
|
|
)
|
|
|
|
const (
|
|
SubframeConstant = "constant"
|
|
SubframeVerbatim = "verbatim"
|
|
SubframeFixed = "fixed"
|
|
SubframeLPC = "lpc"
|
|
SubframeReserved = "reserved"
|
|
)
|
|
|
|
const (
|
|
ChannelLeftSide = 0b1000
|
|
ChannelSideRight = 0b1001
|
|
ChannelMidSide = 0b1010
|
|
)
|
|
|
|
const (
|
|
ResidualCodingMethodRice = 0b00
|
|
ResidualCodingMethodRice2 = 0b01
|
|
)
|
|
|
|
var ResidualCodingMethodMap = scalar.UToScalar{
|
|
ResidualCodingMethodRice: scalar.S{Sym: uint64(4), Description: "rice"},
|
|
ResidualCodingMethodRice2: scalar.S{Sym: uint64(5), Description: "rice2"},
|
|
}
|
|
|
|
// TODO: generic enough?
|
|
func utf8Uint(d *decode.D) uint64 {
|
|
n := d.U8()
|
|
// leading ones, bit negate and count zeroes
|
|
c := bits.LeadingZeros8(^uint8(n))
|
|
// 0b0xxxxxxx 1 byte
|
|
// 0b110xxxxx 2 byte
|
|
// 0b1110xxxx 3 byte
|
|
// 0b11110xxx 4 byte
|
|
switch c {
|
|
case 0:
|
|
// nop
|
|
case 2, 3, 4:
|
|
n = n & ((1 << (8 - c - 1)) - 1)
|
|
for i := 1; i < c; i++ {
|
|
n = n<<6 | d.U8()&0x3f
|
|
}
|
|
default:
|
|
d.Errorf("invalid UTF8Uint")
|
|
}
|
|
return n
|
|
}
|
|
|
|
// in argument is an optional FlacFrameIn struct with stream info
|
|
func frameDecode(d *decode.D, in any) any {
|
|
frameStart := d.Pos()
|
|
blockSize := 0
|
|
channelAssignment := uint64(0)
|
|
channels := 0
|
|
sampleSize := 0
|
|
sideChannelIndex := -1
|
|
|
|
ffi, ok := in.(format.FlacFrameIn)
|
|
if ok {
|
|
sampleSize = ffi.BitsPerSample
|
|
}
|
|
|
|
d.FieldStruct("header", func(d *decode.D) {
|
|
// <14> 11111111111110
|
|
d.FieldU14("sync", d.AssertU(0b11111111111110), scalar.ActualBin)
|
|
|
|
// <1> Reserved
|
|
// 0 : mandatory value
|
|
// 1 : reserved for future use
|
|
d.FieldU1("reserved0", d.AssertU(0))
|
|
|
|
// <1> Blocking strategy:
|
|
// 0 : fixed-blocksize stream; frame header encodes the frame number
|
|
// 1 : variable-blocksize stream; frame header encodes the sample number
|
|
blockingStrategy := d.FieldU1("blocking_strategy", BlockingStrategyNames)
|
|
|
|
// <4> Block size in inter-channel samples:
|
|
// 0000 : reserved
|
|
// 0001 : 192 samples
|
|
// 0010-0101 : 576 * (2^(n-2)) samples, i.e. 576/1152/2304/4608
|
|
// 0110 : get 8 bit (blocksize-1) from end of header
|
|
// 0111 : get 16 bit (blocksize-1) from end of header
|
|
// 1000-1111 : 256 * (2^(n-8)) samples, i.e. 256/512/1024/2048/4096/8192/16384/32768
|
|
var blockSizeMap = scalar.UToScalar{
|
|
0b0000: {Description: "reserved"},
|
|
0b0001: {Sym: uint64(192)},
|
|
0b0010: {Sym: uint64(576)},
|
|
0b0011: {Sym: uint64(1152)},
|
|
0b0100: {Sym: uint64(2304)},
|
|
0b0101: {Sym: uint64(4608)},
|
|
0b0110: {Description: "end of header (8 bit)"},
|
|
0b0111: {Description: "end of header (16 bit)"},
|
|
0b1000: {Sym: uint64(256)},
|
|
0b1001: {Sym: uint64(512)},
|
|
0b1010: {Sym: uint64(1024)},
|
|
0b1011: {Sym: uint64(2048)},
|
|
0b1100: {Sym: uint64(4096)},
|
|
0b1101: {Sym: uint64(8192)},
|
|
0b1110: {Sym: uint64(16384)},
|
|
0b1111: {Sym: uint64(32768)},
|
|
}
|
|
blockSizeS := d.FieldScalarU4("block_size", blockSizeMap, scalar.ActualBin)
|
|
if blockSizeS.Sym != nil {
|
|
blockSize = int(blockSizeS.SymU())
|
|
}
|
|
|
|
// <4> Sample rate:
|
|
// 0000 : get from STREAMINFO metadata block
|
|
// 0001 : 88.2kHz
|
|
// 0010 : 176.4kHz
|
|
// 0011 : 192kHz
|
|
// 0100 : 8kHz
|
|
// 0101 : 16kHz
|
|
// 0110 : 22.05kHz
|
|
// 0111 : 24kHz
|
|
// 1000 : 32kHz
|
|
// 1001 : 44.1kHz
|
|
// 1010 : 48kHz
|
|
// 1011 : 96kHz
|
|
// 1100 : get 8 bit sample rate (in kHz) from end of header
|
|
// 1101 : get 16 bit sample rate (in Hz) from end of header
|
|
// 1110 : get 16 bit sample rate (in tens of Hz) from end of header
|
|
// 1111 : invalid, to prevent sync-fooling string of 1s
|
|
var sampleRateMap = scalar.UToScalar{
|
|
0b0000: {Description: "from streaminfo"},
|
|
0b0001: {Sym: uint64(88200)},
|
|
0b0010: {Sym: uint64(176400)},
|
|
0b0011: {Sym: uint64(192000)},
|
|
0b0100: {Sym: uint64(8000)},
|
|
0b0101: {Sym: uint64(16000)},
|
|
0b0110: {Sym: uint64(22050)},
|
|
0b0111: {Sym: uint64(24000)},
|
|
0b1000: {Sym: uint64(32000)},
|
|
0b1001: {Sym: uint64(44100)},
|
|
0b1010: {Sym: uint64(48000)},
|
|
0b1011: {Sym: uint64(96000)},
|
|
0b1100: {Description: "end of header (8 bit*1000)"},
|
|
0b1101: {Description: "end of header (16 bit)"},
|
|
0b1110: {Description: "end of header (16 bit*10)"},
|
|
0b1111: {Description: "invalid"},
|
|
}
|
|
sampleRateS := d.FieldScalarU4("sample_rate", sampleRateMap, scalar.ActualBin)
|
|
|
|
// <4> Channel assignment
|
|
// 0000-0111 : (number of independent channels)-1. Where defined, the channel order follows SMPTE/ITU-R recommendations. The assignments are as follows:
|
|
// 1 channel: mono
|
|
// 2 channels: left, right
|
|
// 3 channels: left, right, center
|
|
// 4 channels: front left, front right, back left, back right
|
|
// 5 channels: front left, front right, front center, back/surround left, back/surround right
|
|
// 6 channels: front left, front right, front center, LFE, back/surround left, back/surround right
|
|
// 7 channels: front left, front right, front center, LFE, back center, side left, side right
|
|
// 8 channels: front left, front right, front center, LFE, back left, back right, side left, side right
|
|
// 1000 : left/side stereo: channel 0 is the left channel, channel 1 is the side(difference) channel
|
|
// 1001 : right/side stereo: channel 0 is the side(difference) channel, channel 1 is the right channel
|
|
// 1010 : mid/side stereo: channel 0 is the mid(average) channel, channel 1 is the side(difference) channel
|
|
// 1011-1111 : reserved
|
|
// TODO: extract to tables and cleanup
|
|
var channelAssignmentMap = scalar.UToScalar{
|
|
0: {Sym: uint64(1), Description: "mono"},
|
|
1: {Sym: uint64(2), Description: "lr"},
|
|
2: {Sym: uint64(3), Description: "lrc"},
|
|
3: {Sym: uint64(4), Description: "fl,fr,bl,br"},
|
|
4: {Sym: uint64(5), Description: "fl,fr,fc,back/surround left,back/surround right"},
|
|
5: {Sym: uint64(6), Description: "fl,fr,fc,lfe,back/surround left,back/surround right"},
|
|
6: {Sym: uint64(7), Description: "fl,fr,fc,lfe,back center,sl,sr"},
|
|
7: {Sym: uint64(8), Description: "fl,fr,fc,lfe,back left,br,sl,sr"},
|
|
0b1000: {Sym: uint64(2), Description: "left/side stereo"},
|
|
0b1001: {Sym: uint64(2), Description: "right/side stereo"},
|
|
0b1010: {Sym: uint64(2), Description: "mid/side stereo"},
|
|
0b1011: {Sym: nil, Description: "reserved"},
|
|
0b1100: {Sym: nil, Description: "reserved"},
|
|
0b1101: {Sym: nil, Description: "reserved"},
|
|
0b1111: {Sym: nil, Description: "reserved"},
|
|
}
|
|
channelAssignmentS := d.FieldScalarU4("channel_assignment", channelAssignmentMap)
|
|
if channelAssignmentS.Sym == nil {
|
|
d.Fatalf("unknown number of channels")
|
|
}
|
|
channelAssignment = channelAssignmentS.ActualU()
|
|
channels = int(channelAssignmentS.SymU())
|
|
switch channelAssignmentS.ActualU() {
|
|
case ChannelLeftSide:
|
|
sideChannelIndex = 1
|
|
case ChannelSideRight:
|
|
sideChannelIndex = 0
|
|
case ChannelMidSide:
|
|
sideChannelIndex = 1
|
|
}
|
|
if sideChannelIndex != -1 {
|
|
d.FieldValueU("side_channel_index", uint64(sideChannelIndex))
|
|
}
|
|
|
|
// <3> Sample size in bits:
|
|
// 000 : get from STREAMINFO metadata block
|
|
// 001 : 8 bits per sample
|
|
// 010 : 12 bits per sample
|
|
// 011 : reserved
|
|
// 100 : 16 bits per sample
|
|
// 101 : 20 bits per sample
|
|
// 110 : 24 bits per sample
|
|
// 111 : reserved
|
|
var sampleSizeMap = scalar.UToScalar{
|
|
0b000: {Description: "from streaminfo"},
|
|
0b001: {Sym: uint64(8)},
|
|
0b010: {Sym: uint64(12)},
|
|
0b011: {Description: "reserved"},
|
|
0b100: {Sym: uint64(16)},
|
|
0b101: {Sym: uint64(20)},
|
|
0b110: {Sym: uint64(24)},
|
|
0b111: {Sym: uint64(32)},
|
|
}
|
|
sampleSizeS := d.FieldScalarU3("sample_size", sampleSizeMap, scalar.ActualBin)
|
|
switch sampleSizeS.ActualU() {
|
|
case SampleSizeStreaminfo:
|
|
sampleSize = ffi.BitsPerSample
|
|
default:
|
|
if sampleSizeS.Sym != nil {
|
|
sampleSize = int(sampleSizeS.SymU())
|
|
}
|
|
}
|
|
|
|
// <1> Reserved:
|
|
// 0 : mandatory value
|
|
// 1 : reserved for future use
|
|
d.FieldU1("reserved1", d.AssertU(0))
|
|
|
|
d.FieldStruct("end_of_header", func(d *decode.D) {
|
|
// if(variable blocksize)
|
|
// <8-56>:"UTF-8" coded sample number (decoded number is 36 bits) [4]
|
|
// else
|
|
// <8-48>:"UTF-8" coded frame number (decoded number is 31 bits) [4]
|
|
// 0 : fixed-blocksize stream; frame header encodes the frame number
|
|
// 1 : variable-blocksize stream; frame header encodes the sample number
|
|
switch blockingStrategy {
|
|
case BlockingStrategyVariable:
|
|
d.FieldUFn("sample_number", utf8Uint)
|
|
case BlockingStrategyFixed:
|
|
d.FieldUFn("frame_number", utf8Uint)
|
|
}
|
|
|
|
// if(blocksize bits == 011x)
|
|
// 8/16 bit (blocksize-1)
|
|
// 0110 : get 8 bit (blocksize-1) from end of header
|
|
// 0111 : get 16 bit (blocksize-1) from end of header
|
|
switch blockSizeS.ActualU() {
|
|
case BlockSizeEndOfHeader8:
|
|
blockSize = int(d.FieldU8("block_size", scalar.ActualUAdd(1)))
|
|
case BlockSizeEndOfHeader16:
|
|
blockSize = int(d.FieldU16("block_size", scalar.ActualUAdd(1)))
|
|
}
|
|
|
|
// if(sample rate bits == 11xx)
|
|
// 8/16 bit sample rate
|
|
// 1100 : get 8 bit sample rate (in kHz) from end of header
|
|
// 1101 : get 16 bit sample rate (in Hz) from end of header
|
|
// 1110 : get 16 bit sample rate (in tens of Hz) from end of header
|
|
switch sampleRateS.ActualU() {
|
|
case SampeleRateEndOfHeader8:
|
|
d.FieldUFn("sample_rate", func(d *decode.D) uint64 { return d.U8() * 1000 })
|
|
case SampeleRateEndOfHeader16:
|
|
d.FieldU16("sample_rate")
|
|
case SampeleRateEndOfHeader160:
|
|
d.FieldUFn("sample_rate", func(d *decode.D) uint64 { return d.U16() * 10 })
|
|
}
|
|
})
|
|
|
|
headerCRC := &checksum.CRC{Bits: 8, Table: checksum.ATM8Table}
|
|
d.CopyBits(headerCRC, d.BitBufRange(frameStart, d.Pos()-frameStart))
|
|
d.FieldU8("crc", d.ValidateUBytes(headerCRC.Sum(nil)), scalar.ActualHex)
|
|
})
|
|
|
|
var channelSamples [][]int64
|
|
d.FieldArray("subframes", func(d *decode.D) {
|
|
for channelIndex := 0; channelIndex < channels; channelIndex++ {
|
|
d.FieldStruct("subframe", func(d *decode.D) {
|
|
// <1> Zero bit padding, to prevent sync-fooling string of 1s
|
|
d.FieldU1("zero_bit", d.AssertU(0))
|
|
|
|
// <6> Subframe type:
|
|
// 000000 : SUBFRAME_CONSTANT
|
|
// 000001 : SUBFRAME_VERBATIM
|
|
// 00001x : reserved
|
|
// 0001xx : reserved
|
|
// 001xxx : if(xxx <= 4) SUBFRAME_FIXED, xxx=order ; else reserved
|
|
// 01xxxx : reserved
|
|
// 1xxxxx : SUBFRAME_LPC, xxxxx=order-1
|
|
lpcOrder := -1
|
|
var subframeTypeRangeMap = scalar.URangeToScalar{
|
|
{Range: [2]uint64{0b000000, 0b000000}, S: scalar.S{Sym: SubframeConstant}},
|
|
{Range: [2]uint64{0b000001, 0b000001}, S: scalar.S{Sym: SubframeVerbatim}},
|
|
{Range: [2]uint64{0b000010, 0b000011}, S: scalar.S{Sym: SubframeReserved}},
|
|
{Range: [2]uint64{0b000100, 0b000111}, S: scalar.S{Sym: SubframeReserved}},
|
|
{Range: [2]uint64{0b001000, 0b001100}, S: scalar.S{Sym: SubframeFixed}},
|
|
{Range: [2]uint64{0b001101, 0b001111}, S: scalar.S{Sym: SubframeReserved}},
|
|
{Range: [2]uint64{0b010000, 0b011111}, S: scalar.S{Sym: SubframeReserved}},
|
|
{Range: [2]uint64{0b100000, 0b111111}, S: scalar.S{Sym: SubframeLPC}},
|
|
}
|
|
subframeTypeS := d.FieldScalarU6("subframe_type", subframeTypeRangeMap, scalar.ActualBin)
|
|
switch subframeTypeS.SymStr() {
|
|
case SubframeFixed:
|
|
lpcOrder = int(subframeTypeS.ActualU() & 0b111)
|
|
case SubframeLPC:
|
|
lpcOrder = int((subframeTypeS.ActualU() & 0b11111) + 1)
|
|
}
|
|
if lpcOrder != -1 {
|
|
d.FieldValueU("lpc_order", uint64(lpcOrder))
|
|
}
|
|
|
|
// 'Wasted bits-per-sample' flag:
|
|
// 0 : no wasted bits-per-sample in source subblock, k=0
|
|
// 1 : k wasted bits-per-sample in source subblock, k-1 follows, unary coded; e.g. k=3 => 001 follows, k=7 => 0000001 follows.
|
|
wastedBitsFlag := d.FieldU1("wasted_bits_flag")
|
|
var wastedBitsK int
|
|
if wastedBitsFlag != 0 {
|
|
wastedBitsK = int(d.FieldUnary("wasted_bits_k", 0, scalar.ActualUAdd(1)))
|
|
}
|
|
|
|
subframeSampleSize := sampleSize - wastedBitsK
|
|
if subframeSampleSize < 1 {
|
|
d.Fatalf("subframeSampleSize %d < 1", subframeSampleSize)
|
|
}
|
|
// if channel is side, add en extra sample bit
|
|
// https://github.com/xiph/flac/blob/37e675b777d4e0de53ac9ff69e2aea10d92e729c/src/libFLAC/stream_decoder.c#L2040
|
|
if channelIndex == sideChannelIndex {
|
|
subframeSampleSize++
|
|
}
|
|
d.FieldValueU("subframe_sample_size", uint64(subframeSampleSize))
|
|
|
|
decodeWarmupSamples := func(samples []int64, n int, sampleSize int) {
|
|
if len(samples) < n {
|
|
d.Fatalf("decodeWarmupSamples outside block size")
|
|
}
|
|
|
|
d.FieldArray("warmup_samples", func(d *decode.D) {
|
|
for i := 0; i < n; i++ {
|
|
samples[i] = d.FieldS("value", sampleSize)
|
|
}
|
|
})
|
|
}
|
|
|
|
decodeResiduals := func(samples []int64) {
|
|
samplesLen := len(samples)
|
|
n := 0
|
|
|
|
// <2> Residual coding method:
|
|
// 00 : partitioned Rice coding with 4-bit Rice parameter; RESIDUAL_CODING_METHOD_PARTITIONED_RICE follows
|
|
// 01 : partitioned Rice coding with 5-bit Rice parameter; RESIDUAL_CODING_METHOD_PARTITIONED_RICE2 follows
|
|
// 10-11 : reserved
|
|
var riceEscape int
|
|
var riceBits int
|
|
residualCodingMethod := d.FieldU2("residual_coding_method", scalar.UToScalar{
|
|
0b00: scalar.S{Sym: uint64(4), Description: "rice"},
|
|
0b01: scalar.S{Sym: uint64(5), Description: "rice2"},
|
|
})
|
|
switch residualCodingMethod {
|
|
case ResidualCodingMethodRice:
|
|
riceEscape = 0b1111
|
|
riceBits = 4
|
|
case ResidualCodingMethodRice2:
|
|
riceEscape = 0b11111
|
|
riceBits = 5
|
|
}
|
|
|
|
// <4> Partition order.
|
|
partitionOrder := int(d.FieldU4("partition_order"))
|
|
// There will be 2^order partitions.
|
|
ricePartitions := 1 << partitionOrder
|
|
d.FieldValueU("rice_partitions", uint64(ricePartitions))
|
|
|
|
d.FieldArray("partitions", func(d *decode.D) {
|
|
for i := 0; i < ricePartitions; i++ {
|
|
d.FieldStruct("partition", func(d *decode.D) {
|
|
// Encoding parameter:
|
|
// <4(+5)> Encoding parameter:
|
|
// 0000-1110 : Rice parameter.
|
|
// 1111 : Escape code, meaning the partition is in unencoded binary form using n bits per sample; n follows as a 5-bit number.
|
|
// Or:
|
|
// <5(+5)> Encoding parameter:
|
|
// 00000-11110 : Rice parameter.
|
|
// 11111 : Escape code, meaning the partition is in unencoded binary form using n bits per sample; n follows as a 5-bit number.
|
|
// Encoded residual. The number of samples (n) in the partition is determined as follows:
|
|
// if the partition order is zero, n = frame's blocksize - predictor order
|
|
// else if this is not the first partition of the subframe, n = (frame's blocksize / (2^partition order))
|
|
// else n = (frame's blocksize / (2^partition order)) - predictor order
|
|
var count int
|
|
if partitionOrder == 0 {
|
|
count = blockSize - lpcOrder
|
|
} else if i != 0 {
|
|
count = blockSize / ricePartitions
|
|
} else {
|
|
count = (blockSize / ricePartitions) - lpcOrder
|
|
}
|
|
|
|
d.FieldValueU("count", uint64(count))
|
|
|
|
riceParameter := int(d.FieldU("rice_parameter", riceBits))
|
|
|
|
if samplesLen < n+count {
|
|
d.Fatalf("decodeResiduals outside block size")
|
|
}
|
|
if count < 0 {
|
|
d.Fatalf("negative sample count %d", count)
|
|
}
|
|
|
|
if riceParameter == riceEscape {
|
|
escapeSampleSize := int(d.FieldU5("escape_sample_size"))
|
|
if escapeSampleSize == 0 {
|
|
// Zero sample size, we can just skip ahead count samples as they are already zero. From spec:
|
|
// Note that it is possible that the number of bits is 0, which means all residual samples in that partition have
|
|
// a value of 0, and no bits code for the partition itself.
|
|
n += count
|
|
} else {
|
|
d.RangeFn(d.Pos(), int64(count*escapeSampleSize), func(d *decode.D) {
|
|
d.FieldRawLen("samples", int64(count*escapeSampleSize))
|
|
})
|
|
for j := 0; j < count; j++ {
|
|
samples[n] = d.S(escapeSampleSize)
|
|
n++
|
|
}
|
|
}
|
|
} else {
|
|
samplesStart := d.Pos()
|
|
for j := 0; j < count; j++ {
|
|
high := d.Unary(0)
|
|
low := d.U(riceParameter)
|
|
samples[n] = mathex.ZigZag[uint64, int64](high<<riceParameter | low)
|
|
n++
|
|
}
|
|
samplesStop := d.Pos()
|
|
d.RangeFn(samplesStart, samplesStop-samplesStart, func(d *decode.D) {
|
|
d.FieldRawLen("samples", d.BitsLeft())
|
|
})
|
|
}
|
|
})
|
|
}
|
|
})
|
|
}
|
|
|
|
// modifies input samples slice and returns it
|
|
decodeLPC := func(lpcOrder int, samples []int64, coeffs []int64, shift int64) {
|
|
for i := lpcOrder; i < len(samples); i++ {
|
|
r := int64(0)
|
|
for j := 0; j < len(coeffs); j++ {
|
|
c := coeffs[j]
|
|
s := samples[i-j-1]
|
|
r += c * s
|
|
}
|
|
samples[i] = samples[i] + (r >> shift)
|
|
}
|
|
}
|
|
|
|
samples := make([]int64, blockSize)
|
|
switch subframeTypeS.SymStr() {
|
|
case SubframeConstant:
|
|
// <n> Unencoded constant value of the subblock, n = frame's bits-per-sample.
|
|
v := d.FieldS("value", subframeSampleSize)
|
|
for i := 0; i < blockSize; i++ {
|
|
samples[i] = v
|
|
}
|
|
case SubframeVerbatim:
|
|
// <n*i> Unencoded subblock; n = frame's bits-per-sample, i = frame's blocksize.
|
|
// TODO: refactor into some kind of FieldBitBufLenFn?
|
|
d.RangeFn(d.Pos(), int64(blockSize*subframeSampleSize), func(d *decode.D) {
|
|
d.FieldRawLen("samples", d.BitsLeft())
|
|
})
|
|
|
|
for i := 0; i < blockSize; i++ {
|
|
samples[i] = d.S(subframeSampleSize)
|
|
}
|
|
case SubframeFixed:
|
|
// <n> Unencoded warm-up samples (n = frame's bits-per-sample * predictor order).
|
|
decodeWarmupSamples(samples, lpcOrder, subframeSampleSize)
|
|
// Encoded residual
|
|
decodeResiduals(samples[lpcOrder:])
|
|
// http://www.hpl.hp.com/techreports/1999/HPL-1999-144.pdf
|
|
fixedCoeffs := [][]int64{
|
|
{},
|
|
{1},
|
|
{2, -1},
|
|
{3, -3, 1},
|
|
{4, -6, 4, -1},
|
|
}
|
|
coeffs := fixedCoeffs[lpcOrder]
|
|
decodeLPC(lpcOrder, samples, coeffs, 0)
|
|
case SubframeLPC:
|
|
// <n> Unencoded warm-up samples (n = frame's bits-per-sample * lpc order).
|
|
decodeWarmupSamples(samples, lpcOrder, subframeSampleSize)
|
|
// <4> (Quantized linear predictor coefficients' precision in bits)-1 (1111 = invalid).
|
|
precision := int(d.FieldU4("precision", scalar.ActualUAdd(1)))
|
|
// <5> Quantized linear predictor coefficient shift needed in bits (NOTE: this number is signed two's-complement).
|
|
shift := d.FieldS5("shift")
|
|
if shift < 0 {
|
|
d.Fatalf("negative LPC shift %d", shift)
|
|
}
|
|
// <n> Unencoded predictor coefficients (n = qlp coeff precision * lpc order) (NOTE: the coefficients are signed two's-complement).
|
|
var coeffs []int64
|
|
d.FieldArray("coefficients", func(d *decode.D) {
|
|
for i := 0; i < lpcOrder; i++ {
|
|
coeffs = append(coeffs, d.FieldS("value", precision))
|
|
}
|
|
})
|
|
// Encoded residual
|
|
decodeResiduals(samples[lpcOrder:])
|
|
decodeLPC(lpcOrder, samples, coeffs, shift)
|
|
}
|
|
|
|
if wastedBitsK != 0 {
|
|
for i := 0; i < len(samples); i++ {
|
|
samples[i] <<= wastedBitsK
|
|
}
|
|
}
|
|
|
|
channelSamples = append(channelSamples, samples)
|
|
})
|
|
}
|
|
})
|
|
|
|
// <?> Zero-padding to byte alignment.
|
|
d.FieldU("byte_align", d.ByteAlignBits(), d.AssertU(0))
|
|
// <16> CRC-16 (polynomial = x^16 + x^15 + x^2 + x^0, initialized with 0) of everything before the crc, back to and including the frame header sync code
|
|
footerCRC := &checksum.CRC{Bits: 16, Table: checksum.ANSI16Table}
|
|
d.CopyBits(footerCRC, d.BitBufRange(frameStart, d.Pos()-frameStart))
|
|
d.FieldRawLen("footer_crc", 16, d.ValidateBitBuf(footerCRC.Sum(nil)), scalar.RawHex)
|
|
|
|
streamSamples := len(channelSamples[0])
|
|
for j := 0; j < len(channelSamples); j++ {
|
|
if streamSamples > len(channelSamples[j]) {
|
|
d.Fatalf("different amount of samples in channels %d != %d", streamSamples, len(channelSamples[j]))
|
|
}
|
|
}
|
|
|
|
// Transform mid/side channels into left, right
|
|
// mid = (left + right)/2
|
|
// side = left - right
|
|
switch channelAssignment {
|
|
case ChannelLeftSide:
|
|
for i := 0; i < len(channelSamples[0]); i++ {
|
|
channelSamples[1][i] = channelSamples[0][i] - channelSamples[1][i]
|
|
}
|
|
case ChannelSideRight:
|
|
for i := 0; i < len(channelSamples[0]); i++ {
|
|
channelSamples[0][i] = channelSamples[1][i] + channelSamples[0][i]
|
|
}
|
|
case ChannelMidSide:
|
|
for i := 0; i < len(channelSamples[0]); i++ {
|
|
m := channelSamples[0][i]
|
|
s := channelSamples[1][i]
|
|
m = m<<1 | s&1
|
|
channelSamples[0][i] = (m + s) >> 1
|
|
channelSamples[1][i] = (m - s) >> 1
|
|
}
|
|
default:
|
|
// not stereo or no side channel
|
|
}
|
|
|
|
outSampleSize := sampleSize + (sampleSize % 8)
|
|
|
|
bytesPerSample := outSampleSize / 8
|
|
p := 0
|
|
le := binary.LittleEndian
|
|
|
|
interleavedSamplesBuf := ffi.SamplesBuf
|
|
interleavedSamplesBufLen := len(channelSamples) * streamSamples * bytesPerSample
|
|
// TODO: decode read buffer?
|
|
// TODO: reuse buffer if possible
|
|
if interleavedSamplesBuf == nil || len(interleavedSamplesBuf) < interleavedSamplesBufLen {
|
|
interleavedSamplesBuf = make([]byte, interleavedSamplesBufLen)
|
|
}
|
|
|
|
// TODO: speedup by using more cache friendly memory layout for samples
|
|
for i := 0; i < streamSamples; i++ {
|
|
for j := 0; j < len(channelSamples); j++ {
|
|
|
|
s := channelSamples[j][i]
|
|
switch outSampleSize {
|
|
case 8:
|
|
interleavedSamplesBuf[p] = byte(s)
|
|
case 16:
|
|
le.PutUint16(interleavedSamplesBuf[p:], uint16(s))
|
|
case 24:
|
|
interleavedSamplesBuf[p] = byte(s)
|
|
le.PutUint16(interleavedSamplesBuf[p+1:], uint16(s>>8))
|
|
case 32:
|
|
le.PutUint32(interleavedSamplesBuf[p:], uint32(s))
|
|
}
|
|
p += bytesPerSample
|
|
}
|
|
}
|
|
|
|
return format.FlacFrameOut{
|
|
SamplesBuf: interleavedSamplesBuf,
|
|
Samples: uint64(streamSamples),
|
|
Channels: channels,
|
|
BitsPerSample: outSampleSize,
|
|
}
|
|
}
|