S.O.N.G/opus-go
S.O.N.G/opus-go
3
0
Fork 0
Code Issues Pull requests Releases Activity

Add stubs for rest of Silk Decoder

Browse source 
Also include the implemented excitation functions now that we have
reached them.
This commit is contained in:
Sean DuBois 2022-09-05 23:50:07 -04:00
parent 5919d29d78
commit f20a9eedf1
4 changed files with 100 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
decoder.go
View file

@ -38,7 +38,7 @@ func (d *Decoder) Decode(in []byte) (bandwidth Bandwidth, isStereo bool, frames
}
for _, encodedFrame := range encodedFrames {
decoded, err := d.silkDecoder.Decode(encodedFrame, tocHeader.isStereo(), cfg.frameDuration().nanoseconds(), silk.Bandwidth(cfg.bandwidth()))
decoded, err := d.silkDecoder.Decode(encodedFrame, []byte{}, tocHeader.isStereo(), cfg.frameDuration().nanoseconds(), silk.Bandwidth(cfg.bandwidth()))
if err != nil {
return 0, false, nil, err
}

102
internal/silk/decoder.go
View file

@ -1067,6 +1067,63 @@ func (d *Decoder) limitLPCFilterPredictionGain(a32Q17 []int32) (aQ12 []float64)
return
}
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.6.1
func (d *Decoder) decodePitchLags(signalType frameSignalType) error {
if signalType == frameSignalTypeVoiced {
return errUnsupportedVoicedFrames
}
return nil
}
// This allows the encoder to trade off the prediction gain between
// packets against the recovery time after packet loss.
//
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.6.3
func (d *Decoder) decodeLTPScalingParamater(signalType frameSignalType) (float64, error) {
// An LTP scaling parameter appears after the LTP filter coefficients if
// and only if
//
// o This is a voiced frame (see Section 4.2.7.3), and
// o Either
// * This SILK frame corresponds to the first time interval of the
// current Opus frame for its type (LBRR or regular), or
//
// * This is an LBRR frame where the LBRR flags (see Section 4.2.4)
// indicate the previous LBRR frame in the same channel is not
// coded.
// Frames that do not code the scaling parameter
// use the default factor of 15565 (approximately 0.95).
if signalType != frameSignalTypeVoiced {
return 15565.0, nil
}
// TODO
return 0, errUnsupportedVoicedFrames
}
// SILK uses a separate 5-tap pitch filter for each subframe, selected
// from one of three codebooks.
//
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.6.2
func (d *Decoder) decodeLTPFilterCoefficients(signalType frameSignalType) error {
if signalType == frameSignalTypeVoiced {
return errUnsupportedVoicedFrames
}
// TODO
return nil
}
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.9.1
func (d *Decoder) ltpSynthesis() {
}
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.9.2
func (d *Decoder) lpcSynthesis() {
}
// Decode decodes many SILK subframes
// An overview of the decoder is given in Figure 14.
//
@ -1099,7 +1156,7 @@ func (d *Decoder) limitLPCFilterPredictionGain(a32Q17 []int32) (aQ12 []float64)
// 8: Resampled signal
//
// https://datatracker.ietf.org/doc/html/rfc6716#section-4.2.1
func (d *Decoder) Decode(in []byte, isStereo bool, nanoseconds int, bandwidth Bandwidth) (decoded []byte, err error) {
func (d *Decoder) Decode(in, out []byte, isStereo bool, nanoseconds int, bandwidth Bandwidth) ([]byte, error) {
if nanoseconds != nanoseconds20Ms {
return nil, errUnsupportedSilkFrameDuration
} else if isStereo {
@ -1113,7 +1170,7 @@ func (d *Decoder) Decode(in []byte, isStereo bool, nanoseconds int, bandwidth Ba
return nil, errUnsupportedSilkLowBitrateRedundancy
}
signalType, _ := d.determineFrameType(voiceActivityDetected)
signalType, quantizationOffsetType := d.determineFrameType(voiceActivityDetected)
// https://datatracker.ietf.org/doc/html/rfc6716#section-4.2.7.4
d.decodeSubframeQuantizations(signalType)
@ -1145,10 +1202,43 @@ func (d *Decoder) Decode(in []byte, isStereo bool, nanoseconds int, bandwidth Ba
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.5.8
d.limitLPCFilterPredictionGain(a32Q17)
if signalType == frameSignalTypeVoiced {
return nil, errUnsupportedVoicedFrames
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.6.1
if err := d.decodePitchLags(signalType); err != nil {
return nil, err
}
d.isPreviousFrameVoiced = signalType == frameSignalTypeVoiced
return
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.6.2
if err := d.decodeLTPFilterCoefficients(signalType); err != nil {
return nil, err
}
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.6.3
_, err = d.decodeLTPScalingParamater(signalType)
if err != nil {
return nil, err
}
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.7
lcgSeed := d.decodeLinearCongruentialGeneratorSeed()
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.8
shellblocks := d.decodeShellblocks(nanoseconds, bandwidth)
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.8.1
rateLevel := d.decodeRatelevel(signalType == frameSignalTypeVoiced)
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.8.2
pulsecounts, lsbcounts := d.decodePulseAndLSBCounts(shellblocks, rateLevel)
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.8.6
d.decodeExcitation(signalType, quantizationOffsetType, lcgSeed, pulsecounts, lsbcounts)
// https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.9.1
d.ltpSynthesis()
//https://www.rfc-editor.org/rfc/rfc6716.html#section-4.2.7.9.2
d.lpcSynthesis()
d.isPreviousFrameVoiced = signalType == frameSignalTypeVoiced
return out, nil
}

4
internal/silk/decoder_test.go
View file

@ -28,7 +28,7 @@ func createRangeDecoder(data []byte, bitsRead uint, rangeSize uint32, highAndCod
func TestDecode20MsOnly(t *testing.T) {
d := &Decoder{}
_, err := d.Decode(testSilkFrame(), false, 1, BandwidthWideband)
_, err := d.Decode(testSilkFrame(), []byte{}, false, 1, BandwidthWideband)
if !errors.Is(err, errUnsupportedSilkFrameDuration) {
t.Fatal(err)
}
@ -36,7 +36,7 @@ func TestDecode20MsOnly(t *testing.T) {
func TestDecodeStereoTODO(t *testing.T) {
d := &Decoder{}
_, err := d.Decode(testSilkFrame(), true, nanoseconds20Ms, BandwidthWideband)
_, err := d.Decode(testSilkFrame(), []byte{}, true, nanoseconds20Ms, BandwidthWideband)
if !errors.Is(err, errUnsupportedSilkStereo) {
t.Fatal(err)
}

2
internal/silk/errors.go
View file

@ -6,6 +6,6 @@ var (
errUnsupportedSilkFrameDuration = errors.New("only silk frames with a duration of 20ms supported")
errUnsupportedSilkStereo = errors.New("silk decoder does not support stereo")
errUnsupportedSilkLowBitrateRedundancy = errors.New("silk decoder does not low bit-rate redundancy")
errUnsupportedLSFInterpolation = errors.New("silk decoder does not support LSF Interpolation yet")
errUnsupportedLSFInterpolation = errors.New("silk decoder does not support LSF Interpolation")
errUnsupportedVoicedFrames = errors.New("silk decoder does not support voiced frames")
)