July 2009 - FFmpeg-soc - lists.ffmpeg.org

[soc]: r4871 - amr/amrnbdec.c
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 17:36:07 2009 New Revision: 4871 Log: Comments to clarify that the ignored bits are just padding Modified: amr/amrnbdec.c Modified: amr/amrnbdec.c ============================================================================== --- amr/amrnbdec.c Fri Jul 31 17:10:43 2009 (r4870) +++ amr/amrnbdec.c Fri Jul 31 17:36:07 2009 (r4871) @@ -136,12 +136,12 @@ static enum Mode decode_bitstream(AMRCon // initialize get_bits init_get_bits(&p->gb, buf, buf_size * 8); - skip_bits(&p->gb, 1); - // set the mode - mode = get_bits(&p->gb, 4); - // set the bad frame indicator based on the quality bit - p->bad_frame_indicator = !get_bits1(&p->gb); - skip_bits(&p->gb, 2); + + // Decode the first octet. + skip_bits(&p->gb, 1); // padding bit + mode = get_bits(&p->gb, 4); // frame type + p->bad_frame_indicator = !get_bits1(&p->gb); // quality bit + skip_bits(&p->gb, 2); // two padding bits if (mode <= MODE_DTX) { uint16_t *data = (uint16_t *)&p->frame;

1 0

[soc]: r4870 - amr/amrnbdec.c
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 17:10:43 2009 New Revision: 4870 Log: Declare that the AMR decoder will output floats Modified: amr/amrnbdec.c Modified: amr/amrnbdec.c ============================================================================== --- amr/amrnbdec.c Fri Jul 31 16:47:33 2009 (r4869) +++ amr/amrnbdec.c Fri Jul 31 17:10:43 2009 (r4870) @@ -1222,5 +1222,5 @@ AVCodec amrnb_decoder = { .init = amrnb_decode_init, .decode = amrnb_decode_frame, .long_name = NULL_IF_CONFIG_SMALL("Adaptive Multi-Rate NarrowBand"), + .sample_fmts = (enum SampleFormat[]){SAMPLE_FMT_FLT,SAMPLE_FMT_NONE}, }; -

1 0

[soc]: r4869 - amr/amrnbdec.c
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 16:47:33 2009 New Revision: 4869 Log: Include the header for ff_celp_lspf2lpc instead of just declaring it Modified: amr/amrnbdec.c Modified: amr/amrnbdec.c ============================================================================== --- amr/amrnbdec.c Fri Jul 31 16:44:49 2009 (r4868) +++ amr/amrnbdec.c Fri Jul 31 16:47:33 2009 (r4869) @@ -37,9 +37,9 @@ #include "celp_filters.h" #include "acelp_filters.h" #include "acelp_vectors.h" -#include "amrnbdata.h" +#include "lsp.h" -void ff_celp_lspf2lpc(const double *lspf, float *lpc); +#include "amrnbdata.h" typedef struct AMRContext {

1 0

[soc]: r4868 - amr/amr-ffmpeg.diff
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 16:44:49 2009 New Revision: 4868 Log: Absorb qcelp_lsp.c into lsp.c because it's also useful for AMR Modified: amr/amr-ffmpeg.diff Modified: amr/amr-ffmpeg.diff ============================================================================== --- amr/amr-ffmpeg.diff Fri Jul 31 16:36:17 2009 (r4867) +++ amr/amr-ffmpeg.diff Fri Jul 31 16:44:49 2009 (r4868) @@ -6,10 +6,19 @@ Index: libavcodec/Makefile OBJS-$(CONFIG_AC3_ENCODER) += ac3enc.o ac3tab.o ac3.o OBJS-$(CONFIG_ALAC_DECODER) += alac.o OBJS-$(CONFIG_ALAC_ENCODER) += alacenc.o lpc.o -+OBJS-$(CONFIG_AMRNB_DECODER) += amrnbdec.o qcelp_lsp.o celp_filters.o celp_math.o acelp_filters.o acelp_vectors.o ++OBJS-$(CONFIG_AMRNB_DECODER) += amrnbdec.o celp_filters.o celp_math.o acelp_filters.o acelp_vectors.o lsp.o OBJS-$(CONFIG_AMV_DECODER) += sp5xdec.o mjpegdec.o mjpeg.o OBJS-$(CONFIG_APE_DECODER) += apedec.o OBJS-$(CONFIG_ASV1_DECODER) += asv1.o mpeg12data.o +@@ -180,7 +181,7 @@ + OBJS-$(CONFIG_PPM_DECODER) += pnmenc.o pnm.o + OBJS-$(CONFIG_PPM_ENCODER) += pnmenc.o + OBJS-$(CONFIG_PTX_DECODER) += ptx.o +-OBJS-$(CONFIG_QCELP_DECODER) += qcelpdec.o qcelp_lsp.o celp_math.o celp_filters.o acelp_vectors.o ++OBJS-$(CONFIG_QCELP_DECODER) += qcelpdec.o celp_math.o celp_filters.o acelp_vectors.o lsp.o + OBJS-$(CONFIG_QDM2_DECODER) += qdm2.o mpegaudiodec.o mpegaudiodecheader.o mpegaudio.o mpegaudiodata.o + OBJS-$(CONFIG_QDRAW_DECODER) += qdrw.o + OBJS-$(CONFIG_QPEG_DECODER) += qpeg.o Index: libavcodec/allcodecs.c =================================================================== --- libavcodec/allcodecs.c (revision 19401) @@ -214,16 +223,195 @@ Index: libavcodec/acelp_vectors.h + const int length); + #endif /* AVCODEC_ACELP_VECTORS_H */ +Index: libavcodec/qcelpdec.c +=================================================================== +--- libavcodec/qcelpdec.c (revision 19401) ++++ libavcodec/qcelpdec.c (working copy) +@@ -38,6 +38,7 @@ + #include "celp_math.h" + #include "celp_filters.h" + #include "acelp_vectors.h" ++#include "lsp.h" + + #undef NDEBUG + #include <assert.h> +@@ -76,13 +77,6 @@ + } QCELPContext; + + /** +- * Reconstructs LPC coefficients from the line spectral pair frequencies. +- * +- * TIA/EIA/IS-733 2.4.3.3.5 +- */ +-void ff_celp_lspf2lpc(const double *lspf, float *lpc); +- +-/** + * Initialize the speech codec according to the specification. + * + * TIA/EIA/IS-733 2.4.9 Index: libavcodec/qcelp_lsp.c =================================================================== --- libavcodec/qcelp_lsp.c (revision 19401) +++ libavcodec/qcelp_lsp.c (working copy) -@@ -27,8 +27,6 @@ - * @remark Development mentored by Benjamin Larson - */ - +@@ -1,81 +0,0 @@ +-/* +- * QCELP decoder +- * Copyright (c) 2007 Reynaldo H. Verdejo Pinochet +- * +- * This file is part of FFmpeg. +- * +- * FFmpeg is free software; you can redistribute it and/or +- * modify it under the terms of the GNU Lesser General Public +- * License as published by the Free Software Foundation; either +- * version 2.1 of the License, or (at your option) any later version. +- * +- * FFmpeg is distributed in the hope that it will be useful, +- * but WITHOUT ANY WARRANTY; without even the implied warranty of +- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +- * Lesser General Public License for more details. +- * +- * You should have received a copy of the GNU Lesser General Public +- * License along with FFmpeg; if not, write to the Free Software +- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA +- */ +- +-/** +- * @file libavcodec/qcelp_lsp.c +- * QCELP decoder +- * @author Reynaldo H. Verdejo Pinochet +- * @remark FFmpeg merging spearheaded by Kenan Gillet +- * @remark Development mentored by Benjamin Larson +- */ +- -#include "libavutil/mathematics.h" - - /** - * Computes the Pa / (1 + z(-1)) or Qa / (1 - z(-1)) coefficients - * needed for LSP to LPC conversion. +-/** +- * Computes the Pa / (1 + z(-1)) or Qa / (1 - z(-1)) coefficients +- * needed for LSP to LPC conversion. +- * We only need to calculate the 6 first elements of the polynomial. +- * +- * @param lspf line spectral pair frequencies +- * @param f [out] polynomial input/output as a vector +- * +- * TIA/EIA/IS-733 2.4.3.3.5-1/2 +- */ +-static void lsp2polyf(const double *lspf, double *f, int lp_half_order) +-{ +- int i, j; +- +- f[0] = 1.0; +- f[1] = -2 * lspf[0]; +- lspf -= 2; +- for(i=2; i<=lp_half_order; i++) +- { +- double val = -2 * lspf[2*i]; +- f[i] = val * f[i-1] + 2*f[i-2]; +- for(j=i-1; j>1; j--) +- f[j] += f[j-1] * val + f[j-2]; +- f[1] += val; +- } +-} +- +-/** +- * Reconstructs LPC coefficients from the line spectral pair frequencies. +- * +- * @param lspf line spectral pair frequencies +- * @param lpc linear predictive coding coefficients +- */ +-void ff_celp_lspf2lpc(const double *lspf, float *lpc) +-{ +- double pa[6], qa[6]; +- int i; +- +- lsp2polyf(lspf, pa, 5); +- lsp2polyf(lspf + 1, qa, 5); +- +- for (i=4; i>=0; i--) +- { +- double paf = pa[i+1] + pa[i]; +- double qaf = qa[i+1] - qa[i]; +- +- lpc[i ] = 0.5*(paf+qaf); +- lpc[9-i] = 0.5*(paf-qaf); +- } +-} +Index: libavcodec/lsp.c +=================================================================== +--- libavcodec/lsp.c (revision 19401) ++++ libavcodec/lsp.c (working copy) +@@ -1,6 +1,7 @@ + /* + * LSP routines for ACELP-based codecs + * ++ * Copyright (c) 2007 Reynaldo H. Verdejo Pinochet (QCELP decoder) + * Copyright (c) 2008 Vladimir Voroshilov + * + * This file is part of FFmpeg. +@@ -118,3 +119,48 @@ + /* LSP values for second subframe (3.2.5 of G.729)*/ + ff_acelp_lsp2lpc(lp_2nd, lsp_2nd, lp_order >> 1); + } ++ ++/** ++ * Computes the Pa / (1 + z(-1)) or Qa / (1 - z(-1)) coefficients ++ * needed for LSP to LPC conversion. ++ * We only need to calculate the 6 first elements of the polynomial. ++ * ++ * @param lspf line spectral pair frequencies ++ * @param f [out] polynomial input/output as a vector ++ * ++ * TIA/EIA/IS-733 2.4.3.3.5-1/2 ++ */ ++static void lsp2polyf(const double *lspf, double *f, int lp_half_order) ++{ ++ int i, j; ++ ++ f[0] = 1.0; ++ f[1] = -2 * lspf[0]; ++ lspf -= 2; ++ for(i=2; i<=lp_half_order; i++) ++ { ++ double val = -2 * lspf[2*i]; ++ f[i] = val * f[i-1] + 2*f[i-2]; ++ for(j=i-1; j>1; j--) ++ f[j] += f[j-1] * val + f[j-2]; ++ f[1] += val; ++ } ++} ++ ++void ff_celp_lspf2lpc(const double *lspf, float *lpc) ++{ ++ double pa[6], qa[6]; ++ int i; ++ ++ lsp2polyf(lspf, pa, 5); ++ lsp2polyf(lspf + 1, qa, 5); ++ ++ for (i=4; i>=0; i--) ++ { ++ double paf = pa[i+1] + pa[i]; ++ double qaf = qa[i+1] - qa[i]; ++ ++ lpc[i ] = 0.5*(paf+qaf); ++ lpc[9-i] = 0.5*(paf-qaf); ++ } ++} +Index: libavcodec/lsp.h +=================================================================== +--- libavcodec/lsp.h (revision 19401) ++++ libavcodec/lsp.h (working copy) +@@ -67,4 +67,14 @@ + */ + void ff_acelp_lp_decode(int16_t* lp_1st, int16_t* lp_2nd, const int16_t* lsp_2nd, const int16_t* lsp_prev, int lp_order); + ++/** ++ * Reconstructs LPC coefficients from the line spectral pair frequencies. ++ * ++ * TIA/EIA/IS-733 2.4.3.3.5 ++ * ++ * @param lspf line spectral pair frequencies ++ * @param lpc linear predictive coding coefficients ++ */ ++void ff_celp_lspf2lpc(const double *lspf, float *lpc); ++ + #endif /* AVCODEC_LSP_H */

1 0

[soc]: r4867 - amr/amr-ffmpeg.diff
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 16:36:17 2009 New Revision: 4867 Log: qcelp_lsp.c is purely arithmetic and does not need any headers Modified: amr/amr-ffmpeg.diff Modified: amr/amr-ffmpeg.diff ============================================================================== --- amr/amr-ffmpeg.diff Fri Jul 31 16:11:38 2009 (r4866) +++ amr/amr-ffmpeg.diff Fri Jul 31 16:36:17 2009 (r4867) @@ -214,3 +214,16 @@ Index: libavcodec/acelp_vectors.h + const int length); + #endif /* AVCODEC_ACELP_VECTORS_H */ +Index: libavcodec/qcelp_lsp.c +=================================================================== +--- libavcodec/qcelp_lsp.c (revision 19401) ++++ libavcodec/qcelp_lsp.c (working copy) +@@ -27,8 +27,6 @@ + * @remark Development mentored by Benjamin Larson + */ + +-#include "libavutil/mathematics.h" +- + /** + * Computes the Pa / (1 + z(-1)) or Qa / (1 - z(-1)) coefficients + * needed for LSP to LPC conversion.

1 0

[soc]: r4866 - amr/amrnbdec.c
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 16:11:38 2009 New Revision: 4866 Log: Make use of ff_weighted_vector_sumf Modified: amr/amrnbdec.c Modified: amr/amrnbdec.c ============================================================================== --- amr/amrnbdec.c Fri Jul 31 15:55:41 2009 (r4865) +++ amr/amrnbdec.c Fri Jul 31 16:11:38 2009 (r4866) @@ -927,9 +927,8 @@ static int synthesis(AMRContext *p, floa p->pitch_vector[i] *= 0.25; // construct the excitation vector - for (i = 0; i < AMR_SUBFRAME_SIZE; i++) - excitation[i] = p->pitch_gain[4] * p->pitch_vector[i] + - fixed_gain * fixed_vector[i]; + ff_weighted_vector_sumf(excitation, p->pitch_vector, fixed_vector, + p->pitch_gain[4], fixed_gain, AMR_SUBFRAME_SIZE); // if an overflow has been detected, pitch vector contribution emphasis and // adaptive gain control are skipped

1 0

[soc]: r4865 - amr/amr-ffmpeg.diff
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 15:55:41 2009 New Revision: 4865 Log: Move the convolve filter so it's easier to compare to its fixed-point version Modified: amr/amr-ffmpeg.diff Modified: amr/amr-ffmpeg.diff ============================================================================== --- amr/amr-ffmpeg.diff Fri Jul 31 15:22:25 2009 (r4864) +++ amr/amr-ffmpeg.diff Fri Jul 31 15:55:41 2009 (r4865) @@ -53,8 +53,8 @@ Index: libavcodec/celp_filters.c =================================================================== --- libavcodec/celp_filters.c (revision 19401) +++ libavcodec/celp_filters.c (working copy) -@@ -84,6 +84,31 @@ - return 0; +@@ -50,6 +50,31 @@ + } } +void ff_celp_convolve_circf( @@ -82,9 +82,9 @@ Index: libavcodec/celp_filters.c + } +} + - void ff_celp_lp_synthesis_filterf( - float *out, - const float* filter_coeffs, + int ff_celp_lp_synthesis_filter( + int16_t *out, + const int16_t* filter_coeffs, @@ -120,6 +145,6 @@ { out[n] = in[n];

1 0

[soc]: r4862 - amr/amrnbdec.c
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 15:07:57 2009 New Revision: 4862 Log: Cosmetics: whitespace around operators Modified: amr/amrnbdec.c Modified: amr/amrnbdec.c ============================================================================== --- amr/amrnbdec.c Fri Jul 31 14:50:49 2009 (r4861) +++ amr/amrnbdec.c Fri Jul 31 15:07:57 2009 (r4862) @@ -219,7 +219,7 @@ static void interpolate_lsf(float lsf_q[ static void adjust_lsf(float *m) { int i; - float tmp=0.0; + float tmp = 0.0; for (i = 0; i < LP_FILTER_ORDER; i++) tmp = m[i] = FFMAX(m[i], tmp + MIN_LSF_SPACING); } @@ -522,7 +522,7 @@ static void reconstruct_fixed_vector(int // reset the code memset(fixed_vector, 0, AMR_SUBFRAME_SIZE * sizeof(float)); - for (i=0; i<nr_pulses; i++) + for (i = 0; i < nr_pulses; i++) fixed_vector[pulse_position[i]] = ((sign >> i) & 1) ? 1.0 : -1.0; } @@ -679,7 +679,7 @@ static void pitch_sharpening(AMRContext // conduct pitch sharpening as appropriate (section 6.1.2) if (p->pitch_lag_int < AMR_SUBFRAME_SIZE) for (i = p->pitch_lag_int; i < AMR_SUBFRAME_SIZE; i++) - fixed_vector[i] += p->beta*fixed_vector[i-p->pitch_lag_int]; + fixed_vector[i] += p->beta * fixed_vector[i-p->pitch_lag_int]; // Save pitch sharpening factor for the next subframe // MODE_475 only updates on the 2nd and 4th subframes - this follows from @@ -936,8 +936,9 @@ static int synthesis(AMRContext *p, floa if (p->pitch_gain[4] > 0.5 && !overflow) { float excitation_temp[AMR_SUBFRAME_SIZE]; float pitch_factor = (p->cur_frame_mode == MODE_122 ? 0.25 : 0.5) - *FFMIN(p->pitch_gain[4], p->cur_frame_mode == MODE_122 ? 1.0 : SHARP_MAX) - *p->pitch_gain[4]; + * FFMIN(p->pitch_gain[4], + p->cur_frame_mode == MODE_122 ? 1.0 : SHARP_MAX) + * p->pitch_gain[4]; for (i = 0; i < AMR_SUBFRAME_SIZE; i++) // emphasize pitch vector contribution @@ -1095,7 +1096,7 @@ static void postfilter(AMRContext *p, fl for (i = 0; i < AMR_SUBFRAME_SIZE; i++) { p->postfilter_agc = AMR_AGC_ALPHA * p->postfilter_agc + - (1.0 - AMR_AGC_ALPHA) * gain_scale_factor; + (1.0 - AMR_AGC_ALPHA) * gain_scale_factor; buf_out[i] *= p->postfilter_agc; } } @@ -1195,7 +1196,7 @@ static int amrnb_decode_frame(AVCodecCon // scale from PCM values to [-1,1) here too. // FIXME: merge with qcelpdec.c buf_out[i] = av_clipf(buf_out[i] * 2.0 * AMR_SAMPLE_SCALE, - -1.0, 32767.0/32768.0); + -1.0, 32767.0 / 32768.0); /* Update averaged lsf vector (used for fixed gain smoothing). *

2 1

[soc]: r4864 - amr/amrnbdec.c
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 15:22:25 2009 New Revision: 4864 Log: Shrink the integer types in the context. This is a bit cumbersome because pitch lag can be 143 so must be a uint8_t, but its calculation currently uses negative numbers. Modified: amr/amrnbdec.c Modified: amr/amrnbdec.c ============================================================================== --- amr/amrnbdec.c Fri Jul 31 15:08:22 2009 (r4863) +++ amr/amrnbdec.c Fri Jul 31 15:22:25 2009 (r4864) @@ -46,7 +46,7 @@ typedef struct AMRContext { GetBitContext gb; AMRNBFrame frame; ///< decoded AMR parameters (lsf coefficients, codebook indexes, etc) - int bad_frame_indicator; ///< bad frame ? 1 : 0 + uint8_t bad_frame_indicator; ///< bad frame ? 1 : 0 enum Mode cur_frame_mode; ///< current frame mode enum RXFrameType cur_frame_type; ///< current frame type @@ -59,7 +59,7 @@ typedef struct AMRContext { float lpc[4][LP_FILTER_ORDER]; ///< lpc coefficient vectors for 4 subframes - int pitch_lag_int; ///< integer part of pitch lag from current subframe + uint8_t pitch_lag_int; ///< integer part of pitch lag from current subframe float excitation_buf[PITCH_LAG_MAX + LP_FILTER_ORDER + 1 + AMR_SUBFRAME_SIZE]; ///< excitation buffer float *excitation; ///< pointer to the current excitation vector in excitation_buf @@ -71,8 +71,8 @@ typedef struct AMRContext { float fixed_gain[5]; ///< quantified fixed gains for the current and previous four subframes float beta; ///< beta = previous pitch_gain, bounded by [0.0,SHARP_MAX] - int diff_count; ///< the number of subframes for which diff has been above 0.65 - int hang_count; ///< the number of subframes since a hangover period started + uint8_t diff_count; ///< the number of subframes for which diff has been above 0.65 + uint8_t hang_count; ///< the number of subframes since a hangover period started float prev_sparse_fixed_gain; ///< previous fixed gain; used by anti-sparseness processing to determine "onset" uint8_t prev_ir_filter_strength; ///< previous impulse response filter strength; 0 - strong, 1 - medium, 2 - none @@ -486,17 +486,18 @@ static void decode_pitch_vector(AMRConte const AMRNBSubframe *amr_subframe, const int subframe) { - int pitch_lag_frac; + int pitch_lag_int, pitch_lag_frac; // decode integer and fractional parts of pitch lag from parsed pitch // index - decode_pitch_lag(&p->pitch_lag_int, &pitch_lag_frac, amr_subframe->p_lag, + decode_pitch_lag(&pitch_lag_int, &pitch_lag_frac, amr_subframe->p_lag, p->pitch_lag_int, subframe, p->cur_frame_mode); // interpolate the past excitation at the pitch lag to obtain the pitch // vector - interp_pitch_vector(p->excitation, p->pitch_lag_int, pitch_lag_frac, + interp_pitch_vector(p->excitation, pitch_lag_int, pitch_lag_frac, p->cur_frame_mode); + p->pitch_lag_int = pitch_lag_int; // store previous lag in a uint8_t memcpy(p->pitch_vector, p->excitation, AMR_SUBFRAME_SIZE * sizeof(float)); } /// @}

1 0

[soc]: r4863 - amr/amrnbdec.c
by cmcq 31 Jul '09

31 Jul '09

Author: cmcq Date: Fri Jul 31 15:08:22 2009 New Revision: 4863 Log: Remove redundant variable Modified: amr/amrnbdec.c Modified: amr/amrnbdec.c ============================================================================== --- amr/amrnbdec.c Fri Jul 31 15:07:57 2009 (r4862) +++ amr/amrnbdec.c Fri Jul 31 15:08:22 2009 (r4863) @@ -486,12 +486,11 @@ static void decode_pitch_vector(AMRConte const AMRNBSubframe *amr_subframe, const int subframe) { - int prev_pitch_lag_int = p->pitch_lag_int; int pitch_lag_frac; // decode integer and fractional parts of pitch lag from parsed pitch // index decode_pitch_lag(&p->pitch_lag_int, &pitch_lag_frac, amr_subframe->p_lag, - prev_pitch_lag_int, subframe, p->cur_frame_mode); + p->pitch_lag_int, subframe, p->cur_frame_mode); // interpolate the past excitation at the pitch lag to obtain the pitch // vector

1 0