eden-emu/opus
15
0
Fork 0
mirror of https://github.com/xiph/opus.git synced 2025-05-22 11:28:31 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

LPC generation from the cepstral coefficients

Browse source
This commit is contained in:
Jean-Marc Valin 2018-10-16 20:23:20 -04:00
parent fb1d4fdec2
commit 4e331f377f
4 changed files with 26 additions and 15 deletions
Download patch file Download diff file Expand all files Collapse all files

29
dnn/denoise.c
View file

@ -325,12 +325,29 @@ static float frame_analysis(DenoiseState *st, signed char *iexc, short *pred, sh
RNN_COPY(st->analysis_mem, in, FRAME_SIZE); RNN_COPY(st->analysis_mem, in, FRAME_SIZE);
RNN_COPY(x0, x, WINDOW_SIZE); RNN_COPY(x0, x, WINDOW_SIZE);
apply_window(x); apply_window(x);
forward_transform(X, x);
#if TRAINING
for (i=lowpass;i<FREQ_SIZE;i++)
X[i].r = X[i].i = 0;
#endif
compute_band_energy(Ex, X);
{ {
float e; float e;
float g_1; float g_1;
_celt_autocorr(x, ac, NULL, 0, LPC_ORDER, WINDOW_SIZE); if (0) {
_celt_autocorr(x, ac, NULL, 0, LPC_ORDER, WINDOW_SIZE);
} else {
float Xr[FREQ_SIZE];
kiss_fft_cpx X_auto[FREQ_SIZE];
float x_auto[FRAME_SIZE];
interp_band_gain(Xr, Ex);
RNN_CLEAR(X_auto, FREQ_SIZE);
for (i=0;i<160;i++) X_auto[i].r = Xr[i];
inverse_transform(x_auto, X_auto);
for (i=0;i<LPC_ORDER+1;i++) ac[i] = x_auto[i];
}
/* -40 dB noise floor. */ /* -40 dB noise floor. */
ac[0] += ac[0]*1e-4 + 320/12; ac[0] += ac[0]*1e-4 + 320/12/38.;
/* Lag windowing. */ /* Lag windowing. */
for (i=1;i<LPC_ORDER+1;i++) ac[i] *= (1 - 6e-5*i*i); for (i=1;i<LPC_ORDER+1;i++) ac[i] *= (1 - 6e-5*i*i);
e = _celt_lpc(lpc, rc, ac, LPC_ORDER); e = _celt_lpc(lpc, rc, ac, LPC_ORDER);
@ -341,7 +358,7 @@ static float frame_analysis(DenoiseState *st, signed char *iexc, short *pred, sh
printf("\n"); printf("\n");
#endif #endif
#if 0 #if 0
printf("%f 1 ", e); printf("1 ");
for(i=0;i<LPC_ORDER;i++) printf("%f ", lpc[i]); for(i=0;i<LPC_ORDER;i++) printf("%f ", lpc[i]);
printf("\n"); printf("\n");
#endif #endif
@ -363,12 +380,6 @@ static float frame_analysis(DenoiseState *st, signed char *iexc, short *pred, sh
#endif #endif
} }
} }
forward_transform(X, x);
#if TRAINING
for (i=lowpass;i<FREQ_SIZE;i++)
X[i].r = X[i].i = 0;
#endif
compute_band_energy(Ex, X);
return g; return g;
} }

4
dnn/lpcnet.py
View file

@ -11,8 +11,8 @@ import numpy as np
import h5py import h5py
import sys import sys
rnn_units1=512 rnn_units1=384
rnn_units2=32 rnn_units2=16
pcm_bits = 8 pcm_bits = 8
embed_size = 128 embed_size = 128
pcm_levels = 2**pcm_bits pcm_levels = 2**pcm_bits

2
dnn/test_wavenet_audio.py
View file

@ -42,7 +42,7 @@ periods = (50*features[:,:,36:37]+100).astype('int16')
model.load_weights('wavenet5p0_30.h5') model.load_weights('wavenet5_384_10_G16np_50.h5')
order = 16 order = 16

6
dnn/train_wavenet_audio.py
View file

@ -120,8 +120,8 @@ periods = (50*features[:,:,36:37]+100).astype('int16')
in_data = np.concatenate([in_data, pred], axis=-1) in_data = np.concatenate([in_data, pred], axis=-1)
# dump models to disk as we go # dump models to disk as we go
checkpoint = ModelCheckpoint('lpcnet5_512_10_G32np_{epoch:02d}.h5') checkpoint = ModelCheckpoint('lpcnet9_384_10_G16_{epoch:02d}.h5')
#model.load_weights('wavenet4f2_30.h5') #model.load_weights('wavenet4f2_30.h5')
model.compile(optimizer=Adam(0.0005, amsgrad=True, decay=5e-5), loss='sparse_categorical_crossentropy', metrics=['sparse_categorical_accuracy']) model.compile(optimizer=Adam(0.001, amsgrad=True, decay=5e-5), loss='sparse_categorical_crossentropy', metrics=['sparse_categorical_accuracy'])
model.fit([in_data, in_exc, features, periods], out_data, batch_size=batch_size, epochs=60, validation_split=0.2, callbacks=[checkpoint, lpcnet.Sparsify(1000, 20000, 200, 0.1)]) model.fit([in_data, in_exc, features, periods], out_data, batch_size=batch_size, epochs=120, validation_split=0.0, callbacks=[checkpoint, lpcnet.Sparsify(2000, 40000, 400, 0.1)])