November 2023 - ffmpeg-cvslog

avfilter/framesync: fix OOM case
by Paul B Mahol 30 Nov '23

30 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Mon Nov 27 13:04:16 2023 +0100| [a30adf9f96254f4870066c98a6dbf13fc74515a3] | committer: Paul B Mahol avfilter/framesync: fix OOM case Fixes OOM when caller keeps adding frames into filtergraph that reached EOF by other means, for example EOF is signalled by other filter in filtergraph or by buffersink. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a30adf9f96254f487… --- libavfilter/framesync.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/libavfilter/framesync.c b/libavfilter/framesync.c index 6cb4b21fed..23681b5183 100644 --- a/libavfilter/framesync.c +++ b/libavfilter/framesync.c @@ -354,8 +354,11 @@ static int consume_from_fifos(FFFrameSync *fs) int ff_framesync_activate(FFFrameSync *fs) { + AVFilterContext *ctx = fs->parent; int ret; + FF_FILTER_FORWARD_STATUS_BACK_ALL(ctx->outputs[0], ctx); + ret = framesync_advance(fs); if (ret < 0) return ret;

1 0

avfilter/arls_template: use defines for all constants
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Tue Nov 28 16:04:01 2023 +0100| [47e214245bc2fd373ac285fcd7c33ac128506948] | committer: Paul B Mahol avfilter/arls_template: use defines for all constants > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=47e214245bc2fd373… --- libavfilter/arls_template.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/libavfilter/arls_template.c b/libavfilter/arls_template.c index e5b91d0a85..d8b19d89a5 100644 --- a/libavfilter/arls_template.c +++ b/libavfilter/arls_template.c @@ -16,6 +16,8 @@ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#undef ZERO +#undef HALF #undef ONE #undef ftype #undef SAMPLE_FORMAT @@ -23,10 +25,14 @@ #define SAMPLE_FORMAT float #define ftype float #define ONE 1.f +#define HALF 0.5f +#define ZERO 0.f #else #define SAMPLE_FORMAT double #define ftype double #define ONE 1.0 +#define HALF 0.5 +#define ZERO 0.0 #endif #define fn3(a,b) a##_##b @@ -92,21 +98,21 @@ static ftype fn(process_sample)(AudioRLSContext *s, ftype input, ftype desired, for (int i = 0, pos = offset; i < order; i++, pos++) { const int ikernel_size = i * kernel_size; - u[i] = 0.f; + u[i] = ZERO; for (int k = 0, pos = offset; k < order; k++, pos++) u[i] += p[ikernel_size + k] * delay[pos]; g += u[i] * delay[pos]; } - g = 1.f / g; + g = ONE / g; for (int i = 0; i < order; i++) { const int ikernel_size = i * kernel_size; gains[i] = u[i] * g; coeffs[i] = coeffs[order + i] = coeffs[i] + gains[i] * e; - tmp[i] = 0.f; + tmp[i] = ZERO; for (int k = 0, pos = offset; k < order; k++, pos++) tmp[i] += p[ikernel_size + k] * delay[pos]; } @@ -122,7 +128,7 @@ static ftype fn(process_sample)(AudioRLSContext *s, ftype input, ftype desired, const int ikernel_size = i * kernel_size; for (int k = 0; k < order; k++) - p[ikernel_size + k] = (p[ikernel_size + k] - (dp[ikernel_size + k] + dp[kernel_size * k + i]) * 0.5f) * lambda; + p[ikernel_size + k] = (p[ikernel_size + k] - (dp[ikernel_size + k] + dp[kernel_size * k + i]) * HALF) * lambda; } switch (s->output_mode) {

1 0

avfilter: add Affine Projection adaptive audio filter
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Sun Apr 30 17:06:00 2023 +0200| [f66536cc5816a4a3e9af67de26d41ea581505e30] | committer: Paul B Mahol avfilter: add Affine Projection adaptive audio filter > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=f66536cc5816a4a3e… --- Changelog | 1 + doc/filters.texi | 57 ++++++++ libavfilter/Makefile | 1 + libavfilter/aap_template.c | 227 +++++++++++++++++++++++++++++++ libavfilter/af_aap.c | 332 +++++++++++++++++++++++++++++++++++++++++++++ libavfilter/allfilters.c | 1 + libavfilter/version.h | 2 +- 7 files changed, 620 insertions(+), 1 deletion(-) diff --git a/Changelog b/Changelog index 98ea87a359..f00bc27ca4 100644 --- a/Changelog +++ b/Changelog @@ -6,6 +6,7 @@ version <next>: - EVC decoding using external library libxevd - EVC encoding using external library libxeve - QOA decoder and demuxer +- aap filter version 6.1: - libaribcaption decoder diff --git a/doc/filters.texi b/doc/filters.texi index 566bb94f97..27437c4a36 100644 --- a/doc/filters.texi +++ b/doc/filters.texi @@ -418,6 +418,63 @@ build. Below is a description of the currently available audio filters. +@section aap +Apply Affine Projection algorithm to the first audio stream using the second audio stream. + +This adaptive filter is used to estimate unknown audio based on multiple input audio samples. +Affine projection algorithm can make trade-offs between computation complexity with convergence speed. + +A description of the accepted options follows. + +@table @option +@item order +Set the filter order. + +@item projection +Set the projection order. + +@item mu +Set the filter mu. + +@item delta +Set the coefficient to initialize internal covariance matrix. + +@item out_mode +Set the filter output samples. It accepts the following values: +@table @option +@item i +Pass the 1st input. + +@item d +Pass the 2nd input. + +@item o +Pass difference between desired, 2nd input and error signal estimate. + +@item n +Pass difference between input, 1st input and error signal estimate. + +@item e +Pass error signal estimated samples. + +Default value is @var{o}. +@end table + +@item precision +Set which precision to use when processing samples. + +@table @option +@item auto +Auto pick internal sample format depending on other filters. + +@item float +Always use single-floating point precision sample format. + +@item double +Always use double-floating point precision sample format. +@end table +@end table + @section acompressor A compressor is mainly used to reduce the dynamic range of a signal. diff --git a/libavfilter/Makefile b/libavfilter/Makefile index de51c2a403..63725f91b4 100644 --- a/libavfilter/Makefile +++ b/libavfilter/Makefile @@ -35,6 +35,7 @@ OBJS-$(CONFIG_DNN) += dnn_filter_common.o include $(SRC_PATH)/libavfilter/dnn/Makefile # audio filters +OBJS-$(CONFIG_AAP_FILTER) += af_aap.o OBJS-$(CONFIG_ABENCH_FILTER) += f_bench.o OBJS-$(CONFIG_ACOMPRESSOR_FILTER) += af_sidechaincompress.o OBJS-$(CONFIG_ACONTRAST_FILTER) += af_acontrast.o diff --git a/libavfilter/aap_template.c b/libavfilter/aap_template.c new file mode 100644 index 0000000000..ea9c815a89 --- /dev/null +++ b/libavfilter/aap_template.c @@ -0,0 +1,227 @@ +/* + * 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 + */ + +#undef ZERO +#undef ONE +#undef ftype +#undef SAMPLE_FORMAT +#if DEPTH == 32 +#define SAMPLE_FORMAT float +#define ftype float +#define ONE 1.f +#define ZERO 0.f +#else +#define SAMPLE_FORMAT double +#define ftype double +#define ONE 1.0 +#define ZERO 0.0 +#endif + +#define fn3(a,b) a##_##b +#define fn2(a,b) fn3(a,b) +#define fn(a) fn2(a, SAMPLE_FORMAT) + +#if DEPTH == 64 +static double scalarproduct_double(const double *v1, const double *v2, int len) +{ + double p = 0.0; + + for (int i = 0; i < len; i++) + p += v1[i] * v2[i]; + + return p; +} +#endif + +static ftype fn(fir_sample)(AudioAPContext *s, ftype sample, ftype *delay, + ftype *coeffs, ftype *tmp, int *offset) +{ + const int order = s->order; + ftype output; + + delay[*offset] = sample; + + memcpy(tmp, coeffs + order - *offset, order * sizeof(ftype)); +#if DEPTH == 32 + output = s->fdsp->scalarproduct_float(delay, tmp, s->kernel_size); +#else + output = scalarproduct_double(delay, tmp, s->kernel_size); +#endif + + if (--(*offset) < 0) + *offset = order - 1; + + return output; +} + +static int fn(lup_decompose)(ftype **MA, const int N, const ftype tol, int *P) +{ + for (int i = 0; i <= N; i++) + P[i] = i; + + for (int i = 0; i < N; i++) { + ftype maxA = ZERO; + int imax = i; + + for (int k = i; k < N; k++) { + ftype absA = fabs(MA[k][i]); + if (absA > maxA) { + maxA = absA; + imax = k; + } + } + + if (maxA < tol) + return 0; + + if (imax != i) { + FFSWAP(int, P[i], P[imax]); + FFSWAP(ftype *, MA[i], MA[imax]); + P[N]++; + } + + for (int j = i + 1; j < N; j++) { + MA[j][i] /= MA[i][i]; + + for (int k = i + 1; k < N; k++) + MA[j][k] -= MA[j][i] * MA[i][k]; + } + } + + return 1; +} + +static void fn(lup_invert)(ftype *const *MA, const int *P, const int N, ftype **IA) +{ + for (int j = 0; j < N; j++) { + for (int i = 0; i < N; i++) { + IA[i][j] = P[i] == j ? ONE : ZERO; + + for (int k = 0; k < i; k++) + IA[i][j] -= MA[i][k] * IA[k][j]; + } + + for (int i = N - 1; i >= 0; i--) { + for (int k = i + 1; k < N; k++) + IA[i][j] -= MA[i][k] * IA[k][j]; + + IA[i][j] /= MA[i][i]; + } + } +} + +static ftype fn(process_sample)(AudioAPContext *s, ftype input, ftype desired, int ch) +{ + ftype *dcoeffs = (ftype *)s->dcoeffs->extended_data[ch]; + ftype *coeffs = (ftype *)s->coeffs->extended_data[ch]; + ftype *delay = (ftype *)s->delay->extended_data[ch]; + ftype **itmpmp = (ftype **)&s->itmpmp[s->projection * ch]; + ftype **tmpmp = (ftype **)&s->tmpmp[s->projection * ch]; + ftype *tmpm = (ftype *)s->tmpm->extended_data[ch]; + ftype *tmp = (ftype *)s->tmp->extended_data[ch]; + ftype *e = (ftype *)s->e->extended_data[ch]; + ftype *x = (ftype *)s->x->extended_data[ch]; + ftype *w = (ftype *)s->w->extended_data[ch]; + int *p = (int *)s->p->extended_data[ch]; + int *offset = (int *)s->offset->extended_data[ch]; + const int projection = s->projection; + const ftype delta = s->delta; + const int order = s->order; + const int length = projection + order; + const ftype mu = s->mu; + const ftype tol = 0.00001f; + ftype output; + + x[offset[2] + length] = x[offset[2]] = input; + delay[offset[0] + order] = input; + + output = fn(fir_sample)(s, input, delay, coeffs, tmp, offset); + e[offset[1]] = e[offset[1] + projection] = desired - output; + + for (int i = 0; i < projection; i++) { + const int iprojection = i * projection; + + for (int j = i; j < projection; j++) { + ftype sum = ZERO; + for (int k = 0; k < order; k++) + sum += x[offset[2] + i + k] * x[offset[2] + j + k]; + tmpm[iprojection + j] = sum; + if (i != j) + tmpm[j * projection + i] = sum; + } + + tmpm[iprojection + i] += delta; + } + + fn(lup_decompose)(tmpmp, projection, tol, p); + fn(lup_invert)(tmpmp, p, projection, itmpmp); + + for (int i = 0; i < projection; i++) { + ftype sum = ZERO; + for (int j = 0; j < projection; j++) + sum += itmpmp[i][j] * e[j + offset[1]]; + w[i] = sum; + } + + for (int i = 0; i < order; i++) { + ftype sum = ZERO; + for (int j = 0; j < projection; j++) + sum += x[offset[2] + i + j] * w[j]; + dcoeffs[i] = sum; + } + + for (int i = 0; i < order; i++) + coeffs[i] = coeffs[i + order] = coeffs[i] + mu * dcoeffs[i]; + + if (--offset[1] < 0) + offset[1] = projection - 1; + + if (--offset[2] < 0) + offset[2] = length - 1; + + switch (s->output_mode) { + case IN_MODE: output = input; break; + case DESIRED_MODE: output = desired; break; + case OUT_MODE: output = desired - output; break; + case NOISE_MODE: output = input - output; break; + case ERROR_MODE: break; + } + return output; +} + +static int fn(filter_channels)(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs) +{ + AudioAPContext *s = ctx->priv; + AVFrame *out = arg; + const int start = (out->ch_layout.nb_channels * jobnr) / nb_jobs; + const int end = (out->ch_layout.nb_channels * (jobnr+1)) / nb_jobs; + + for (int c = start; c < end; c++) { + const ftype *input = (const ftype *)s->frame[0]->extended_data[c]; + const ftype *desired = (const ftype *)s->frame[1]->extended_data[c]; + ftype *output = (ftype *)out->extended_data[c]; + + for (int n = 0; n < out->nb_samples; n++) { + output[n] = fn(process_sample)(s, input[n], desired[n], c); + if (ctx->is_disabled) + output[n] = input[n]; + } + } + + return 0; +} diff --git a/libavfilter/af_aap.c b/libavfilter/af_aap.c new file mode 100644 index 0000000000..96c8d27af4 --- /dev/null +++ b/libavfilter/af_aap.c @@ -0,0 +1,332 @@ +/* + * Copyright (c) 2023 Paul B Mahol + * + * 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 + */ + +#include "libavutil/channel_layout.h" +#include "libavutil/common.h" +#include "libavutil/float_dsp.h" +#include "libavutil/opt.h" + +#include "audio.h" +#include "avfilter.h" +#include "formats.h" +#include "filters.h" +#include "internal.h" + +enum OutModes { + IN_MODE, + DESIRED_MODE, + OUT_MODE, + NOISE_MODE, + ERROR_MODE, + NB_OMODES +}; + +typedef struct AudioAPContext { + const AVClass *class; + + int order; + int projection; + float mu; + float delta; + int output_mode; + int precision; + + int kernel_size; + AVFrame *offset; + AVFrame *delay; + AVFrame *coeffs; + AVFrame *e; + AVFrame *p; + AVFrame *x; + AVFrame *w; + AVFrame *dcoeffs; + AVFrame *tmp; + AVFrame *tmpm; + AVFrame *itmpm; + + void **tmpmp; + void **itmpmp; + + AVFrame *frame[2]; + + int (*filter_channels)(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs); + + AVFloatDSPContext *fdsp; +} AudioAPContext; + +#define OFFSET(x) offsetof(AudioAPContext, x) +#define A AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM +#define AT AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_RUNTIME_PARAM + +static const AVOption aap_options[] = { + { "order", "set the filter order", OFFSET(order), AV_OPT_TYPE_INT, {.i64=16}, 1, INT16_MAX, A }, + { "projection", "set the filter projection", OFFSET(projection), AV_OPT_TYPE_INT, {.i64=2}, 1, 256, A }, + { "mu", "set the filter mu", OFFSET(mu), AV_OPT_TYPE_FLOAT, {.dbl=0.0001},0,1, AT }, + { "delta", "set the filter delta", OFFSET(delta), AV_OPT_TYPE_FLOAT, {.dbl=0.001},0, 1, AT }, + { "out_mode", "set output mode", OFFSET(output_mode), AV_OPT_TYPE_INT, {.i64=OUT_MODE}, 0, NB_OMODES-1, AT, "mode" }, + { "i", "input", 0, AV_OPT_TYPE_CONST, {.i64=IN_MODE}, 0, 0, AT, "mode" }, + { "d", "desired", 0, AV_OPT_TYPE_CONST, {.i64=DESIRED_MODE}, 0, 0, AT, "mode" }, + { "o", "output", 0, AV_OPT_TYPE_CONST, {.i64=OUT_MODE}, 0, 0, AT, "mode" }, + { "n", "noise", 0, AV_OPT_TYPE_CONST, {.i64=NOISE_MODE}, 0, 0, AT, "mode" }, + { "e", "error", 0, AV_OPT_TYPE_CONST, {.i64=ERROR_MODE}, 0, 0, AT, "mode" }, + { "precision", "set processing precision", OFFSET(precision), AV_OPT_TYPE_INT, {.i64=0}, 0, 2, A, "precision" }, + { "auto", "set auto processing precision", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, A, "precision" }, + { "float", "set single-floating point processing precision", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, A, "precision" }, + { "double","set double-floating point processing precision", 0, AV_OPT_TYPE_CONST, {.i64=2}, 0, 0, A, "precision" }, + { NULL } +}; + +AVFILTER_DEFINE_CLASS(aap); + +static int query_formats(AVFilterContext *ctx) +{ + AudioAPContext *s = ctx->priv; + static const enum AVSampleFormat sample_fmts[3][3] = { + { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_DBLP, AV_SAMPLE_FMT_NONE }, + { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE }, + { AV_SAMPLE_FMT_DBLP, AV_SAMPLE_FMT_NONE }, + }; + int ret; + + if ((ret = ff_set_common_all_channel_counts(ctx)) < 0) + return ret; + + if ((ret = ff_set_common_formats_from_list(ctx, sample_fmts[s->precision])) < 0) + return ret; + + return ff_set_common_all_samplerates(ctx); +} + +static int activate(AVFilterContext *ctx) +{ + AudioAPContext *s = ctx->priv; + int i, ret, status; + int nb_samples; + int64_t pts; + + FF_FILTER_FORWARD_STATUS_BACK_ALL(ctx->outputs[0], ctx); + + nb_samples = FFMIN(ff_inlink_queued_samples(ctx->inputs[0]), + ff_inlink_queued_samples(ctx->inputs[1])); + for (i = 0; i < ctx->nb_inputs && nb_samples > 0; i++) { + if (s->frame[i]) + continue; + + if (ff_inlink_check_available_samples(ctx->inputs[i], nb_samples) > 0) { + ret = ff_inlink_consume_samples(ctx->inputs[i], nb_samples, nb_samples, &s->frame[i]); + if (ret < 0) + return ret; + } + } + + if (s->frame[0] && s->frame[1]) { + AVFrame *out; + + out = ff_get_audio_buffer(ctx->outputs[0], s->frame[0]->nb_samples); + if (!out) { + av_frame_free(&s->frame[0]); + av_frame_free(&s->frame[1]); + return AVERROR(ENOMEM); + } + + ff_filter_execute(ctx, s->filter_channels, out, NULL, + FFMIN(ctx->outputs[0]->ch_layout.nb_channels, ff_filter_get_nb_threads(ctx))); + + out->pts = s->frame[0]->pts; + out->duration = s->frame[0]->duration; + + av_frame_free(&s->frame[0]); + av_frame_free(&s->frame[1]); + + ret = ff_filter_frame(ctx->outputs[0], out); + if (ret < 0) + return ret; + } + + if (!nb_samples) { + for (i = 0; i < 2; i++) { + if (ff_inlink_acknowledge_status(ctx->inputs[i], &status, &pts)) { + ff_outlink_set_status(ctx->outputs[0], status, pts); + return 0; + } + } + } + + if (ff_outlink_frame_wanted(ctx->outputs[0])) { + for (i = 0; i < 2; i++) { + if (s->frame[i] || ff_inlink_queued_samples(ctx->inputs[i]) > 0) + continue; + ff_inlink_request_frame(ctx->inputs[i]); + return 0; + } + } + return 0; +} + +#define DEPTH 32 +#include "aap_template.c" + +#undef DEPTH +#define DEPTH 64 +#include "aap_template.c" + +static int config_output(AVFilterLink *outlink) +{ + const int channels = outlink->ch_layout.nb_channels; + AVFilterContext *ctx = outlink->src; + AudioAPContext *s = ctx->priv; + + s->kernel_size = FFALIGN(s->order, 16); + + if (!s->offset) + s->offset = ff_get_audio_buffer(outlink, 3); + if (!s->delay) + s->delay = ff_get_audio_buffer(outlink, 2 * s->kernel_size); + if (!s->dcoeffs) + s->dcoeffs = ff_get_audio_buffer(outlink, s->kernel_size); + if (!s->coeffs) + s->coeffs = ff_get_audio_buffer(outlink, 2 * s->kernel_size); + if (!s->e) + s->e = ff_get_audio_buffer(outlink, 2 * s->projection); + if (!s->p) + s->p = ff_get_audio_buffer(outlink, s->projection + 1); + if (!s->x) + s->x = ff_get_audio_buffer(outlink, 2 * (s->projection + s->order)); + if (!s->w) + s->w = ff_get_audio_buffer(outlink, s->projection); + if (!s->tmp) + s->tmp = ff_get_audio_buffer(outlink, s->kernel_size); + if (!s->tmpm) + s->tmpm = ff_get_audio_buffer(outlink, s->projection * s->projection); + if (!s->itmpm) + s->itmpm = ff_get_audio_buffer(outlink, s->projection * s->projection); + + if (!s->tmpmp) + s->tmpmp = av_calloc(s->projection * channels, sizeof(*s->tmpmp)); + if (!s->itmpmp) + s->itmpmp = av_calloc(s->projection * channels, sizeof(*s->itmpmp)); + + if (!s->offset || !s->delay || !s->dcoeffs || !s->coeffs || !s->tmpmp || !s->itmpmp || + !s->e || !s->p || !s->x || !s->w || !s->tmp || !s->tmpm || !s->itmpm) + return AVERROR(ENOMEM); + + switch (outlink->format) { + case AV_SAMPLE_FMT_DBLP: + for (int ch = 0; ch < channels; ch++) { + double *itmpm = (double *)s->itmpm->extended_data[ch]; + double *tmpm = (double *)s->tmpm->extended_data[ch]; + double **itmpmp = (double **)&s->itmpmp[s->projection * ch]; + double **tmpmp = (double **)&s->tmpmp[s->projection * ch]; + + for (int i = 0; i < s->projection; i++) { + itmpmp[i] = &itmpm[i * s->projection]; + tmpmp[i] = &tmpm[i * s->projection]; + } + } + + s->filter_channels = filter_channels_double; + break; + case AV_SAMPLE_FMT_FLTP: + for (int ch = 0; ch < channels; ch++) { + float *itmpm = (float *)s->itmpm->extended_data[ch]; + float *tmpm = (float *)s->tmpm->extended_data[ch]; + float **itmpmp = (float **)&s->itmpmp[s->projection * ch]; + float **tmpmp = (float **)&s->tmpmp[s->projection * ch]; + + for (int i = 0; i < s->projection; i++) { + itmpmp[i] = &itmpm[i * s->projection]; + tmpmp[i] = &tmpm[i * s->projection]; + } + } + + s->filter_channels = filter_channels_float; + break; + } + + return 0; +} + +static av_cold int init(AVFilterContext *ctx) +{ + AudioAPContext *s = ctx->priv; + + s->fdsp = avpriv_float_dsp_alloc(0); + if (!s->fdsp) + return AVERROR(ENOMEM); + + return 0; +} + +static av_cold void uninit(AVFilterContext *ctx) +{ + AudioAPContext *s = ctx->priv; + + av_freep(&s->fdsp); + + av_frame_free(&s->offset); + av_frame_free(&s->delay); + av_frame_free(&s->dcoeffs); + av_frame_free(&s->coeffs); + av_frame_free(&s->e); + av_frame_free(&s->p); + av_frame_free(&s->w); + av_frame_free(&s->x); + av_frame_free(&s->tmp); + av_frame_free(&s->tmpm); + av_frame_free(&s->itmpm); + + av_freep(&s->tmpmp); + av_freep(&s->itmpmp); +} + +static const AVFilterPad inputs[] = { + { + .name = "input", + .type = AVMEDIA_TYPE_AUDIO, + }, + { + .name = "desired", + .type = AVMEDIA_TYPE_AUDIO, + }, +}; + +static const AVFilterPad outputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_AUDIO, + .config_props = config_output, + }, +}; + +const AVFilter ff_af_aap = { + .name = "aap", + .description = NULL_IF_CONFIG_SMALL("Apply Affine Projection algorithm to first audio stream."), + .priv_size = sizeof(AudioAPContext), + .priv_class = &aap_class, + .init = init, + .uninit = uninit, + .activate = activate, + FILTER_INPUTS(inputs), + FILTER_OUTPUTS(outputs), + FILTER_QUERY_FUNC(query_formats), + .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_INTERNAL | + AVFILTER_FLAG_SLICE_THREADS, + .process_command = ff_filter_process_command, +}; diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c index aa49703c6e..ed7c32be94 100644 --- a/libavfilter/allfilters.c +++ b/libavfilter/allfilters.c @@ -21,6 +21,7 @@ #include "avfilter.h" +extern const AVFilter ff_af_aap; extern const AVFilter ff_af_abench; extern const AVFilter ff_af_acompressor; extern const AVFilter ff_af_acontrast; diff --git a/libavfilter/version.h b/libavfilter/version.h index 537df129cd..7642b670d1 100644 --- a/libavfilter/version.h +++ b/libavfilter/version.h @@ -31,7 +31,7 @@ #include "version_major.h" -#define LIBAVFILTER_VERSION_MINOR 13 +#define LIBAVFILTER_VERSION_MINOR 14 #define LIBAVFILTER_VERSION_MICRO 100

1 0

lavc/hevcdsp_qpel_neon: using movi.16b instead of movi.2d
by xufuji456 28 Nov '23

28 Nov '23

ffmpeg | branch: master | xufuji456 <839789740(a)qq.com> | Wed Nov 15 11:53:12 2023 +0800| [cc86343b960793a822d6c51b58a1a7e3319cb217] | committer: Martin Storsjö lavc/hevcdsp_qpel_neon: using movi.16b instead of movi.2d Building iOS platform with arm64, the compiler has a warning: "instruction movi.2d with immediate #0 may not function correctly on this CPU, converting to movi.16b" Signed-off-by: xufuji456 <839789740(a)qq.com> Signed-off-by: Martin Storsjö <martin(a)martin.st> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=cc86343b960793a82… --- libavcodec/aarch64/hevcdsp_epel_neon.S | 136 ++++++++++++++++----------------- libavcodec/aarch64/hevcdsp_qpel_neon.S | 52 ++++++------- libavcodec/aarch64/mpegaudiodsp_neon.S | 8 +- libswscale/aarch64/hscale.S | 72 ++++++++--------- 4 files changed, 134 insertions(+), 134 deletions(-) diff --git a/libavcodec/aarch64/hevcdsp_epel_neon.S b/libavcodec/aarch64/hevcdsp_epel_neon.S index a2a051210f..c077c204cc 100644 --- a/libavcodec/aarch64/hevcdsp_epel_neon.S +++ b/libavcodec/aarch64/hevcdsp_epel_neon.S @@ -694,7 +694,7 @@ function ff_hevc_put_hevc_epel_h4_8_neon_i8mm, export=1 trn1 v4.2s, v4.2s, v5.2s trn1 v6.2s, v6.2s, v7.2s trn1 v4.2d, v4.2d, v6.2d - movi v16.2d, #0 + movi v16.16b, #0 usdot v16.4s, v4.16b, v30.16b xtn v16.4h, v16.4s st1 {v16.4h}, [x0], x10 @@ -714,8 +714,8 @@ function ff_hevc_put_hevc_epel_h6_8_neon_i8mm, export=1 trn2 v17.2s, v4.2s, v5.2s trn1 v6.2s, v6.2s, v7.2s trn1 v16.2d, v16.2d, v6.2d - movi v18.2d, #0 - movi v19.2d, #0 + movi v18.16b, #0 + movi v19.16b, #0 usdot v18.4s, v16.16b, v30.16b usdot v19.2s, v17.8b, v30.8b xtn v18.4h, v18.4s @@ -736,8 +736,8 @@ function ff_hevc_put_hevc_epel_h8_8_neon_i8mm, export=1 ext v7.16b, v4.16b, v4.16b, #3 zip1 v20.4s, v4.4s, v6.4s zip1 v21.4s, v5.4s, v7.4s - movi v16.2d, #0 - movi v17.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 usdot v16.4s, v20.16b, v30.16b usdot v17.4s, v21.16b, v30.16b xtn v16.4h, v16.4s @@ -761,9 +761,9 @@ function ff_hevc_put_hevc_epel_h12_8_neon_i8mm, export=1 trn1 v4.4s, v20.4s, v21.4s trn2 v5.4s, v20.4s, v21.4s trn1 v6.4s, v22.4s, v23.4s - movi v16.2d, #0 - movi v17.2d, #0 - movi v18.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 + movi v18.16b, #0 usdot v16.4s, v4.16b, v30.16b usdot v17.4s, v5.16b, v30.16b usdot v18.4s, v6.16b, v30.16b @@ -788,10 +788,10 @@ function ff_hevc_put_hevc_epel_h16_8_neon_i8mm, export=1 zip2 v22.4s, v0.4s, v6.4s zip1 v21.4s, v5.4s, v7.4s zip2 v23.4s, v5.4s, v7.4s - movi v16.2d, #0 - movi v17.2d, #0 - movi v18.2d, #0 - movi v19.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 + movi v18.16b, #0 + movi v19.16b, #0 usdot v16.4s, v20.16b, v30.16b usdot v17.4s, v21.16b, v30.16b usdot v18.4s, v22.16b, v30.16b @@ -815,14 +815,14 @@ function ff_hevc_put_hevc_epel_h24_8_neon_i8mm, export=1 ext v26.16b, v1.16b, v1.16b, #1 ext v27.16b, v1.16b, v1.16b, #2 ext v28.16b, v1.16b, v1.16b, #3 - movi v16.2d, #0 - movi v17.2d, #0 - movi v18.2d, #0 - movi v19.2d, #0 - movi v20.2d, #0 - movi v21.2d, #0 - movi v22.2d, #0 - movi v23.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 + movi v18.16b, #0 + movi v19.16b, #0 + movi v20.16b, #0 + movi v21.16b, #0 + movi v22.16b, #0 + movi v23.16b, #0 usdot v16.4s, v0.16b, v30.16b usdot v17.4s, v5.16b, v30.16b usdot v18.4s, v6.16b, v30.16b @@ -861,14 +861,14 @@ function ff_hevc_put_hevc_epel_h32_8_neon_i8mm, export=1 ext v26.16b, v1.16b, v2.16b, #1 ext v27.16b, v1.16b, v2.16b, #2 ext v28.16b, v1.16b, v2.16b, #3 - movi v16.2d, #0 - movi v17.2d, #0 - movi v18.2d, #0 - movi v19.2d, #0 - movi v20.2d, #0 - movi v21.2d, #0 - movi v22.2d, #0 - movi v23.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 + movi v18.16b, #0 + movi v19.16b, #0 + movi v20.16b, #0 + movi v21.16b, #0 + movi v22.16b, #0 + movi v23.16b, #0 usdot v16.4s, v0.16b, v30.16b usdot v17.4s, v5.16b, v30.16b usdot v18.4s, v6.16b, v30.16b @@ -900,18 +900,18 @@ function ff_hevc_put_hevc_epel_h48_8_neon_i8mm, export=1 ext v16.16b, v1.16b, v2.16b, #1 ext v17.16b, v1.16b, v2.16b, #2 ext v18.16b, v1.16b, v2.16b, #3 - movi v20.2d, #0 - movi v21.2d, #0 - movi v22.2d, #0 - movi v23.2d, #0 + movi v20.16b, #0 + movi v21.16b, #0 + movi v22.16b, #0 + movi v23.16b, #0 usdot v20.4s, v0.16b, v30.16b usdot v21.4s, v4.16b, v30.16b usdot v22.4s, v5.16b, v30.16b usdot v23.4s, v6.16b, v30.16b - movi v24.2d, #0 - movi v25.2d, #0 - movi v26.2d, #0 - movi v27.2d, #0 + movi v24.16b, #0 + movi v25.16b, #0 + movi v26.16b, #0 + movi v27.16b, #0 usdot v24.4s, v1.16b, v30.16b usdot v25.4s, v16.16b, v30.16b usdot v26.4s, v17.16b, v30.16b @@ -928,10 +928,10 @@ function ff_hevc_put_hevc_epel_h48_8_neon_i8mm, export=1 ext v4.16b, v2.16b, v3.16b, #1 ext v5.16b, v2.16b, v3.16b, #2 ext v6.16b, v2.16b, v3.16b, #3 - movi v20.2d, #0 - movi v21.2d, #0 - movi v22.2d, #0 - movi v23.2d, #0 + movi v20.16b, #0 + movi v21.16b, #0 + movi v22.16b, #0 + movi v23.16b, #0 usdot v20.4s, v2.16b, v30.16b usdot v21.4s, v4.16b, v30.16b usdot v22.4s, v5.16b, v30.16b @@ -957,18 +957,18 @@ function ff_hevc_put_hevc_epel_h64_8_neon_i8mm, export=1 ext v16.16b, v1.16b, v2.16b, #1 ext v17.16b, v1.16b, v2.16b, #2 ext v18.16b, v1.16b, v2.16b, #3 - movi v20.2d, #0 - movi v21.2d, #0 - movi v22.2d, #0 - movi v23.2d, #0 + movi v20.16b, #0 + movi v21.16b, #0 + movi v22.16b, #0 + movi v23.16b, #0 usdot v20.4s, v0.16b, v30.16b usdot v21.4s, v4.16b, v30.16b usdot v22.4s, v5.16b, v30.16b usdot v23.4s, v6.16b, v30.16b - movi v24.2d, #0 - movi v25.2d, #0 - movi v26.2d, #0 - movi v27.2d, #0 + movi v24.16b, #0 + movi v25.16b, #0 + movi v26.16b, #0 + movi v27.16b, #0 usdot v24.4s, v1.16b, v30.16b usdot v25.4s, v16.16b, v30.16b usdot v26.4s, v17.16b, v30.16b @@ -989,18 +989,18 @@ function ff_hevc_put_hevc_epel_h64_8_neon_i8mm, export=1 ext v16.16b, v3.16b, v7.16b, #1 ext v17.16b, v3.16b, v7.16b, #2 ext v18.16b, v3.16b, v7.16b, #3 - movi v20.2d, #0 - movi v21.2d, #0 - movi v22.2d, #0 - movi v23.2d, #0 + movi v20.16b, #0 + movi v21.16b, #0 + movi v22.16b, #0 + movi v23.16b, #0 usdot v20.4s, v2.16b, v30.16b usdot v21.4s, v4.16b, v30.16b usdot v22.4s, v5.16b, v30.16b usdot v23.4s, v6.16b, v30.16b - movi v24.2d, #0 - movi v25.2d, #0 - movi v26.2d, #0 - movi v27.2d, #0 + movi v24.16b, #0 + movi v25.16b, #0 + movi v26.16b, #0 + movi v27.16b, #0 usdot v24.4s, v3.16b, v30.16b usdot v25.4s, v16.16b, v30.16b usdot v26.4s, v17.16b, v30.16b @@ -1593,7 +1593,7 @@ function ff_hevc_put_hevc_epel_uni_w_h4_8_neon_i8mm, export=1 trn1 v0.2s, v0.2s, v2.2s trn1 v1.2s, v1.2s, v3.2s zip1 v0.4s, v0.4s, v1.4s - movi v16.2d, #0 + movi v16.16b, #0 usdot v16.4s, v0.16b, v28.16b mul v16.4s, v16.4s, v30.4s sqrshl v16.4s, v16.4s, v31.4s @@ -1620,8 +1620,8 @@ function ff_hevc_put_hevc_epel_uni_w_h6_8_neon_i8mm, export=1 trn2 v6.2s, v0.2s, v1.2s trn1 v5.2s, v2.2s, v3.2s zip1 v4.2d, v4.2d, v5.2d - movi v16.2d, #0 - movi v17.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 usdot v16.4s, v4.16b, v28.16b usdot v17.2s, v6.8b, v28.8b mul v16.4s, v16.4s, v30.4s @@ -1640,8 +1640,8 @@ function ff_hevc_put_hevc_epel_uni_w_h6_8_neon_i8mm, export=1 endfunc .macro EPEL_UNI_W_H_CALC s0, s1, d0, d1 - movi \d0\().2d, #0 - movi \d1\().2d, #0 + movi \d0\().16b, #0 + movi \d1\().16b, #0 usdot \d0\().4s, \s0\().16b, v28.16b usdot \d1\().4s, \s1\().16b, v28.16b mul \d0\().4s, \d0\().4s, v30.4s @@ -1687,7 +1687,7 @@ function ff_hevc_put_hevc_epel_uni_w_h12_8_neon_i8mm, export=1 zip2 v7.4s, v1.4s, v3.4s zip1 v6.4s, v6.4s, v7.4s EPEL_UNI_W_H_CALC v4, v5, v16, v17 - movi v18.2d, #0 + movi v18.16b, #0 usdot v18.4s, v6.16b, v28.16b mul v18.4s, v18.4s, v30.4s sqrshl v18.4s, v18.4s, v31.4s @@ -2575,7 +2575,7 @@ DISABLE_I8MM .endm .macro EPEL_UNI_W_V4_CALC d0, s0, s1, s2, s3 - movi \d0\().2d, #0 + movi \d0\().16b, #0 umlsl \d0\().8h, \s0\().8b, v0.8b umlal \d0\().8h, \s1\().8b, v1.8b umlal \d0\().8h, \s2\().8b, v2.8b @@ -2626,7 +2626,7 @@ function ff_hevc_put_hevc_epel_uni_w_v4_8_neon, export=1 endfunc .macro EPEL_UNI_W_V8_CALC d0, s0, s1, s2, s3, t0, t1 - movi \d0\().2d, #0 + movi \d0\().16b, #0 umlsl \d0\().8h, \s0\().8b, v0.8b umlal \d0\().8h, \s1\().8b, v1.8b umlal \d0\().8h, \s2\().8b, v2.8b @@ -2720,8 +2720,8 @@ function ff_hevc_put_hevc_epel_uni_w_v8_8_neon, export=1 endfunc .macro EPEL_UNI_W_V12_CALC d0, d1, s0, s1, s2, s3, t0, t1, t2, t3 - movi \d0\().2d, #0 - movi \d1\().2d, #0 + movi \d0\().16b, #0 + movi \d1\().16b, #0 umlsl \d0\().8h, \s0\().8b, v0.8b umlsl2 \d1\().8h, \s0\().16b, v0.16b umlal \d0\().8h, \s1\().8b, v1.8b @@ -2793,8 +2793,8 @@ function ff_hevc_put_hevc_epel_uni_w_v12_8_neon, export=1 endfunc .macro EPEL_UNI_W_V16_CALC d0, d1, s0, s1, s2, s3, t0, t1, t2, t3 - movi \d0\().2d, #0 - movi \d1\().2d, #0 + movi \d0\().16b, #0 + movi \d1\().16b, #0 umlsl \d0\().8h, \s0\().8b, v0.8b umlsl2 \d1\().8h, \s0\().16b, v0.16b umlal \d0\().8h, \s1\().8b, v1.8b diff --git a/libavcodec/aarch64/hevcdsp_qpel_neon.S b/libavcodec/aarch64/hevcdsp_qpel_neon.S index 8adfa38ccf..bcee627cba 100644 --- a/libavcodec/aarch64/hevcdsp_qpel_neon.S +++ b/libavcodec/aarch64/hevcdsp_qpel_neon.S @@ -2180,8 +2180,8 @@ function ff_hevc_put_hevc_qpel_uni_w_h4_8_neon_i8mm, export=1 ext v3.16b, v0.16b, v0.16b, #3 zip1 v0.2d, v0.2d, v1.2d zip1 v2.2d, v2.2d, v3.2d - movi v16.2d, #0 - movi v17.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 usdot v16.4s, v0.16b, v28.16b usdot v17.4s, v2.16b, v28.16b addp v16.4s, v16.4s, v17.4s @@ -2210,9 +2210,9 @@ function ff_hevc_put_hevc_qpel_uni_w_h6_8_neon_i8mm, export=1 zip1 v0.2d, v0.2d, v1.2d zip1 v2.2d, v2.2d, v3.2d zip1 v4.2d, v4.2d, v5.2d - movi v16.2d, #0 - movi v17.2d, #0 - movi v18.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 + movi v18.16b, #0 usdot v16.4s, v0.16b, v28.16b usdot v17.4s, v2.16b, v28.16b usdot v18.4s, v4.16b, v28.16b @@ -2236,10 +2236,10 @@ endfunc .macro QPEL_UNI_W_H_CALC s0, s1, s2, s3, d0, d1, d2, d3 - movi \d0\().2d, #0 - movi \d1\().2d, #0 - movi \d2\().2d, #0 - movi \d3\().2d, #0 + movi \d0\().16b, #0 + movi \d1\().16b, #0 + movi \d2\().16b, #0 + movi \d3\().16b, #0 usdot \d0\().4s, \s0\().16b, v28.16b usdot \d1\().4s, \s1\().16b, v28.16b usdot \d2\().4s, \s2\().16b, v28.16b @@ -2255,8 +2255,8 @@ endfunc .endm .macro QPEL_UNI_W_H_CALC_HALF s0, s1, d0, d1 - movi \d0\().2d, #0 - movi \d1\().2d, #0 + movi \d0\().16b, #0 + movi \d1\().16b, #0 usdot \d0\().4s, \s0\().16b, v28.16b usdot \d1\().4s, \s1\().16b, v28.16b addp \d0\().4s, \d0\().4s, \d1\().4s @@ -2606,8 +2606,8 @@ function ff_hevc_put_hevc_qpel_h4_8_neon_i8mm, export=1 ext v3.16b, v0.16b, v0.16b, #3 zip1 v0.2d, v0.2d, v1.2d zip1 v2.2d, v2.2d, v3.2d - movi v16.2d, #0 - movi v17.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 usdot v16.4s, v0.16b, v31.16b usdot v17.4s, v2.16b, v31.16b addp v16.4s, v16.4s, v17.4s @@ -2633,9 +2633,9 @@ function ff_hevc_put_hevc_qpel_h6_8_neon_i8mm, export=1 zip1 v0.2d, v0.2d, v1.2d zip1 v2.2d, v2.2d, v3.2d zip1 v4.2d, v4.2d, v5.2d - movi v16.2d, #0 - movi v17.2d, #0 - movi v18.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 + movi v18.16b, #0 usdot v16.4s, v0.16b, v31.16b usdot v17.4s, v2.16b, v31.16b usdot v18.4s, v4.16b, v31.16b @@ -2668,10 +2668,10 @@ function ff_hevc_put_hevc_qpel_h8_8_neon_i8mm, export=1 zip1 v2.2d, v2.2d, v3.2d zip1 v4.2d, v4.2d, v5.2d zip1 v6.2d, v6.2d, v7.2d - movi v16.2d, #0 - movi v17.2d, #0 - movi v18.2d, #0 - movi v19.2d, #0 + movi v16.16b, #0 + movi v17.16b, #0 + movi v18.16b, #0 + movi v19.16b, #0 usdot v16.4s, v0.16b, v31.16b usdot v17.4s, v2.16b, v31.16b usdot v18.4s, v4.16b, v31.16b @@ -2688,10 +2688,10 @@ function ff_hevc_put_hevc_qpel_h8_8_neon_i8mm, export=1 endfunc .macro QPEL_H_CALC s0, s1, s2, s3, d0, d1, d2, d3 - movi \d0\().2d, #0 - movi \d1\().2d, #0 - movi \d2\().2d, #0 - movi \d3\().2d, #0 + movi \d0\().16b, #0 + movi \d1\().16b, #0 + movi \d2\().16b, #0 + movi \d3\().16b, #0 usdot \d0\().4s, \s0\().16b, v31.16b usdot \d1\().4s, \s1\().16b, v31.16b usdot \d2\().4s, \s2\().16b, v31.16b @@ -2716,8 +2716,8 @@ function ff_hevc_put_hevc_qpel_h12_8_neon_i8mm, export=1 QPEL_H_CALC v16, v1, v2, v3, v20, v21, v22, v23 addp v20.4s, v20.4s, v22.4s addp v21.4s, v21.4s, v23.4s - movi v24.2d, #0 - movi v25.2d, #0 + movi v24.16b, #0 + movi v25.16b, #0 usdot v24.4s, v18.16b, v31.16b usdot v25.4s, v19.16b, v31.16b addp v24.4s, v24.4s, v25.4s diff --git a/libavcodec/aarch64/mpegaudiodsp_neon.S b/libavcodec/aarch64/mpegaudiodsp_neon.S index b6ef131228..9799c271d0 100644 --- a/libavcodec/aarch64/mpegaudiodsp_neon.S +++ b/libavcodec/aarch64/mpegaudiodsp_neon.S @@ -56,7 +56,7 @@ function ff_mpadsp_apply_window_\type\()_neon, export=1 .ifc \type, fixed ld1r {v16.2s}, [x2] // dither_state sxtl v16.2d, v16.2s - movi v29.2d, #0 + movi v29.16b, #0 movi v30.2d, #(1<<OUT_SHIFT)-1 trn1 v31.2d, v29.2d, v30.2d trn2 v30.2d, v30.2d, v29.2d @@ -74,9 +74,9 @@ function ff_mpadsp_apply_window_\type\()_neon, export=1 add x11, x6, #(32)<<2 // w + 32 add x12, x7, #(32)<<2 // w2 + 32 mov x15, #8 - movi v17.2d, #0 - movi v18.2d, #0 - movi v19.2d, #0 + movi v17.16b, #0 + movi v18.16b, #0 + movi v19.16b, #0 2: subs x15, x15, #1 ld1 {v0.4s}, [x8], x9 diff --git a/libswscale/aarch64/hscale.S b/libswscale/aarch64/hscale.S index b49443c964..b3873fc4b7 100644 --- a/libswscale/aarch64/hscale.S +++ b/libswscale/aarch64/hscale.S @@ -50,10 +50,10 @@ function ff_hscale8to15_X8_neon, export=1 add x12, x16, x7 // filter1 = filter0 + filterSize*2 add x13, x12, x7 // filter2 = filter1 + filterSize*2 add x4, x13, x7 // filter3 = filter2 + filterSize*2 - movi v0.2d, #0 // val sum part 1 (for dst[0]) - movi v1.2d, #0 // val sum part 2 (for dst[1]) - movi v2.2d, #0 // val sum part 3 (for dst[2]) - movi v3.2d, #0 // val sum part 4 (for dst[3]) + movi v0.16b, #0 // val sum part 1 (for dst[0]) + movi v1.16b, #0 // val sum part 2 (for dst[1]) + movi v2.16b, #0 // val sum part 3 (for dst[2]) + movi v3.16b, #0 // val sum part 4 (for dst[3]) add x17, x3, w8, uxtw // srcp + filterPos[0] add x8, x3, w0, uxtw // srcp + filterPos[1] add x0, x3, w11, uxtw // srcp + filterPos[2] @@ -108,10 +108,10 @@ function ff_hscale8to15_X4_neon, export=1 ldp w8, w9, [x5] // filterPos[idx + 0], [idx + 1] ldp w10, w11, [x5, #8] // filterPos[idx + 2], [idx + 3] - movi v16.2d, #0 // initialize accumulator for idx + 0 - movi v17.2d, #0 // initialize accumulator for idx + 1 - movi v18.2d, #0 // initialize accumulator for idx + 2 - movi v19.2d, #0 // initialize accumulator for idx + 3 + movi v16.16b, #0 // initialize accumulator for idx + 0 + movi v17.16b, #0 // initialize accumulator for idx + 1 + movi v18.16b, #0 // initialize accumulator for idx + 2 + movi v19.16b, #0 // initialize accumulator for idx + 3 mov x12, x4 // filter pointer for idx + 0 add x13, x4, x7 // filter pointer for idx + 1 @@ -253,8 +253,8 @@ function ff_hscale8to15_4_neon, export=1 ldp w12, w13, [x5, #16] // filterPos[idx + 4][0..3], [idx + 5][0..3], next iteration ldp w14, w15, [x5, #24] // filterPos[idx + 6][0..3], [idx + 7][0..3], next iteration - movi v0.2d, #0 // Clear madd accumulator for idx 0..3 - movi v5.2d, #0 // Clear madd accumulator for idx 4..7 + movi v0.16b, #0 // Clear madd accumulator for idx 0..3 + movi v5.16b, #0 // Clear madd accumulator for idx 4..7 ld4 {v1.8h, v2.8h, v3.8h, v4.8h}, [x4], #64 // load filter idx + 0..7 @@ -299,8 +299,8 @@ function ff_hscale8to15_4_neon, export=1 ld4 {v16.8b, v17.8b, v18.8b, v19.8b}, [sp] ld4 {v1.8h, v2.8h, v3.8h, v4.8h}, [x4], #64 // load filter idx + 0..7 - movi v0.2d, #0 // Clear madd accumulator for idx 0..3 - movi v5.2d, #0 // Clear madd accumulator for idx 4..7 + movi v0.16b, #0 // Clear madd accumulator for idx 0..3 + movi v5.16b, #0 // Clear madd accumulator for idx 4..7 uxtl v16.8h, v16.8b // unsigned extend long, covert src data to 16-bit uxtl v17.8h, v17.8b // unsigned extend long, covert src data to 16-bit @@ -499,10 +499,10 @@ function ff_hscale8to19_X8_neon, export=1 ldr w11, [x5], #4 // filterPos[idx + 2] add x4, x13, x7 // filter3 = filter2 + filterSize*2 ldr w9, [x5], #4 // filterPos[idx + 3] - movi v0.2d, #0 // val sum part 1 (for dst[0]) - movi v1.2d, #0 // val sum part 2 (for dst[1]) - movi v2.2d, #0 // val sum part 3 (for dst[2]) - movi v3.2d, #0 // val sum part 4 (for dst[3]) + movi v0.16b, #0 // val sum part 1 (for dst[0]) + movi v1.16b, #0 // val sum part 2 (for dst[1]) + movi v2.16b, #0 // val sum part 3 (for dst[2]) + movi v3.16b, #0 // val sum part 4 (for dst[3]) add x17, x3, w8, uxtw // srcp + filterPos[0] add x8, x3, w0, uxtw // srcp + filterPos[1] add x0, x3, w11, uxtw // srcp + filterPos[2] @@ -560,10 +560,10 @@ function ff_hscale8to19_X4_neon, export=1 ldp w8, w9, [x5] ldp w10, w11, [x5, #8] - movi v16.2d, #0 // initialize accumulator for idx + 0 - movi v17.2d, #0 // initialize accumulator for idx + 1 - movi v18.2d, #0 // initialize accumulator for idx + 2 - movi v19.2d, #0 // initialize accumulator for idx + 3 + movi v16.16b, #0 // initialize accumulator for idx + 0 + movi v17.16b, #0 // initialize accumulator for idx + 1 + movi v18.16b, #0 // initialize accumulator for idx + 2 + movi v19.16b, #0 // initialize accumulator for idx + 3 mov x12, x4 // filter + 0 add x13, x4, x7 // filter + 1 @@ -865,10 +865,10 @@ function ff_hscale16to15_X8_neon_asm, export=1 add x12, x16, x7 // filter1 = filter0 + filterSize*2 add x13, x12, x7 // filter2 = filter1 + filterSize*2 add x4, x13, x7 // filter3 = filter2 + filterSize*2 - movi v0.2d, #0 // val sum part 1 (for dst[0]) - movi v1.2d, #0 // val sum part 2 (for dst[1]) - movi v2.2d, #0 // val sum part 3 (for dst[2]) - movi v3.2d, #0 // val sum part 4 (for dst[3]) + movi v0.16b, #0 // val sum part 1 (for dst[0]) + movi v1.16b, #0 // val sum part 2 (for dst[1]) + movi v2.16b, #0 // val sum part 3 (for dst[2]) + movi v3.16b, #0 // val sum part 4 (for dst[3]) add x17, x3, w8, uxtw // srcp + filterPos[0] add x8, x3, w10, uxtw // srcp + filterPos[1] add x10, x3, w11, uxtw // srcp + filterPos[2] @@ -945,10 +945,10 @@ function ff_hscale16to15_X4_neon_asm, export=1 ldp w8, w9, [x5] ldp w10, w11, [x5, #8] - movi v16.2d, #0 // initialize accumulator for idx + 0 - movi v17.2d, #0 // initialize accumulator for idx + 1 - movi v18.2d, #0 // initialize accumulator for idx + 2 - movi v19.2d, #0 // initialize accumulator for idx + 3 + movi v16.16b, #0 // initialize accumulator for idx + 0 + movi v17.16b, #0 // initialize accumulator for idx + 1 + movi v18.16b, #0 // initialize accumulator for idx + 2 + movi v19.16b, #0 // initialize accumulator for idx + 3 mov x12, x4 // filter + 0 add x13, x4, x7 // filter + 1 @@ -1270,10 +1270,10 @@ function ff_hscale16to19_X8_neon_asm, export=1 add x13, x12, x7 // filter2 = filter1 + filterSize*2 lsl w10, w10, #1 add x4, x13, x7 // filter3 = filter2 + filterSize*2 - movi v0.2d, #0 // val sum part 1 (for dst[0]) - movi v1.2d, #0 // val sum part 2 (for dst[1]) - movi v2.2d, #0 // val sum part 3 (for dst[2]) - movi v3.2d, #0 // val sum part 4 (for dst[3]) + movi v0.16b, #0 // val sum part 1 (for dst[0]) + movi v1.16b, #0 // val sum part 2 (for dst[1]) + movi v2.16b, #0 // val sum part 3 (for dst[2]) + movi v3.16b, #0 // val sum part 4 (for dst[3]) add x17, x3, w8, uxtw // srcp + filterPos[0] add x8, x3, w10, uxtw // srcp + filterPos[1] add x10, x3, w11, uxtw // srcp + filterPos[2] @@ -1348,10 +1348,10 @@ function ff_hscale16to19_X4_neon_asm, export=1 ldp w8, w9, [x5] ldp w10, w11, [x5, #8] - movi v16.2d, #0 // initialize accumulator for idx + 0 - movi v17.2d, #0 // initialize accumulator for idx + 1 - movi v18.2d, #0 // initialize accumulator for idx + 2 - movi v19.2d, #0 // initialize accumulator for idx + 3 + movi v16.16b, #0 // initialize accumulator for idx + 0 + movi v17.16b, #0 // initialize accumulator for idx + 1 + movi v18.16b, #0 // initialize accumulator for idx + 2 + movi v19.16b, #0 // initialize accumulator for idx + 3 mov x12, x4 // filter + 0 add x13, x4, x7 // filter + 1

1 0

avfilter/af_anlms: set output frame duration
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Tue Nov 28 13:08:12 2023 +0100| [67ce690bc6b64623bbe4ec79d4be7cad01cdfbbf] | committer: Paul B Mahol avfilter/af_anlms: set output frame duration > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=67ce690bc6b64623b… --- libavfilter/af_anlms.c | 1 + 1 file changed, 1 insertion(+) diff --git a/libavfilter/af_anlms.c b/libavfilter/af_anlms.c index 1516cad0a5..17ec8db31f 100644 --- a/libavfilter/af_anlms.c +++ b/libavfilter/af_anlms.c @@ -142,6 +142,7 @@ static int activate(AVFilterContext *ctx) FFMIN(ctx->outputs[0]->ch_layout.nb_channels, ff_filter_get_nb_threads(ctx))); out->pts = s->frame[0]->pts; + out->duration = s->frame[0]->duration; av_frame_free(&s->frame[0]); av_frame_free(&s->frame[1]);

1 0

avfilter/af_arls: set output frame duration
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Tue Nov 28 13:07:49 2023 +0100| [411c516453065323c32a01211497ed4e8fffd621] | committer: Paul B Mahol avfilter/af_arls: set output frame duration > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=411c516453065323c… --- libavfilter/af_arls.c | 1 + 1 file changed, 1 insertion(+) diff --git a/libavfilter/af_arls.c b/libavfilter/af_arls.c index 301e963d8a..4e77f9a561 100644 --- a/libavfilter/af_arls.c +++ b/libavfilter/af_arls.c @@ -139,6 +139,7 @@ static int activate(AVFilterContext *ctx) FFMIN(ctx->outputs[0]->ch_layout.nb_channels, ff_filter_get_nb_threads(ctx))); out->pts = s->frame[0]->pts; + out->duration = s->frame[0]->duration; av_frame_free(&s->frame[0]); av_frame_free(&s->frame[1]);

1 0

avfilter/af_amix: set output frame duration
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Tue Nov 28 13:05:49 2023 +0100| [bafbb0697e05e78aca5b3eccc9c64fca8f05249b] | committer: Paul B Mahol avfilter/af_amix: set output frame duration > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=bafbb0697e05e78ac… --- libavfilter/af_amix.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/libavfilter/af_amix.c b/libavfilter/af_amix.c index 120a97f48c..39353d6642 100644 --- a/libavfilter/af_amix.c +++ b/libavfilter/af_amix.c @@ -379,6 +379,9 @@ static int output_frame(AVFilterLink *outlink) av_frame_free(&in_buf); out_buf->pts = s->next_pts; + out_buf->duration = av_rescale_q(out_buf->nb_samples, av_make_q(1, outlink->sample_rate), + outlink->time_base); + if (s->next_pts != AV_NOPTS_VALUE) s->next_pts += nb_samples;

1 0

avfilter/af_amultiply: set output frame duration
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Tue Nov 28 12:55:16 2023 +0100| [358aced447fb1fe11da8bf888dab4e9cade7bcdf] | committer: Paul B Mahol avfilter/af_amultiply: set output frame duration > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=358aced447fb1fe11… --- libavfilter/af_amultiply.c | 1 + 1 file changed, 1 insertion(+) diff --git a/libavfilter/af_amultiply.c b/libavfilter/af_amultiply.c index 7871487cdd..4ae355c794 100644 --- a/libavfilter/af_amultiply.c +++ b/libavfilter/af_amultiply.c @@ -74,6 +74,7 @@ static int activate(AVFilterContext *ctx) return AVERROR(ENOMEM); out->pts = s->frames[0]->pts; + out->duration = s->frames[0]->duration; if (av_get_packed_sample_fmt(ctx->inputs[0]->format) == AV_SAMPLE_FMT_FLT) { for (i = 0; i < s->planes; i++) {

1 0

avfilter/af_amerge: use already provided outlink
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Tue Nov 28 12:50:10 2023 +0100| [8b9c400f1d63eae28b1142eb998012b5327d95f9] | committer: Paul B Mahol avfilter/af_amerge: use already provided outlink > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=8b9c400f1d63eae28… --- libavfilter/af_amerge.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/libavfilter/af_amerge.c b/libavfilter/af_amerge.c index 6ef05e213f..83318f99e4 100644 --- a/libavfilter/af_amerge.c +++ b/libavfilter/af_amerge.c @@ -156,7 +156,7 @@ static int config_output(AVFilterLink *outlink) char buf[128]; int i; - s->bps = av_get_bytes_per_sample(ctx->outputs[0]->format); + s->bps = av_get_bytes_per_sample(outlink->format); outlink->time_base = ctx->inputs[0]->time_base; av_bprint_init(&bp, 0, AV_BPRINT_SIZE_AUTOMATIC); @@ -166,7 +166,7 @@ static int config_output(AVFilterLink *outlink) av_bprintf(&bp, "%s", buf); } av_bprintf(&bp, " -> out:"); - av_channel_layout_describe(&ctx->outputs[0]->ch_layout, buf, sizeof(buf)); + av_channel_layout_describe(&outlink->ch_layout, buf, sizeof(buf)); av_bprintf(&bp, "%s", buf); av_log(ctx, AV_LOG_VERBOSE, "%s\n", bp.str); @@ -234,7 +234,7 @@ static int try_push_frame(AVFilterContext *ctx, int nb_samples) ins[i] = inbuf[i]->data[0]; } - outbuf = ff_get_audio_buffer(ctx->outputs[0], nb_samples); + outbuf = ff_get_audio_buffer(outlink, nb_samples); if (!outbuf) { free_frames(s->nb_inputs, inbuf); return AVERROR(ENOMEM); @@ -279,7 +279,7 @@ FF_ENABLE_DEPRECATION_WARNINGS } free_frames(s->nb_inputs, inbuf); - return ff_filter_frame(ctx->outputs[0], outbuf); + return ff_filter_frame(outlink, outbuf); } static int activate(AVFilterContext *ctx)

1 0

avfilter: no need to request more samples if internal frame is available
by Paul B Mahol 28 Nov '23

28 Nov '23

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Tue Nov 28 12:19:08 2023 +0100| [c979ccdfd77e0e8ebaa1604dbc2a11497f24cf97] | committer: Paul B Mahol avfilter: no need to request more samples if internal frame is available > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=c979ccdfd77e0e8eb… --- libavfilter/af_amultiply.c | 2 +- libavfilter/af_anlms.c | 2 +- libavfilter/af_arls.c | 2 +- libavfilter/af_asdr.c | 4 ++-- 4 files changed, 5 insertions(+), 5 deletions(-) diff --git a/libavfilter/af_amultiply.c b/libavfilter/af_amultiply.c index 0e2bdacd8f..7871487cdd 100644 --- a/libavfilter/af_amultiply.c +++ b/libavfilter/af_amultiply.c @@ -110,7 +110,7 @@ static int activate(AVFilterContext *ctx) if (ff_outlink_frame_wanted(ctx->outputs[0])) { for (i = 0; i < 2; i++) { - if (ff_inlink_queued_samples(ctx->inputs[i]) > 0) + if (s->frames[i] || ff_inlink_queued_samples(ctx->inputs[i]) > 0) continue; ff_inlink_request_frame(ctx->inputs[i]); return 0; diff --git a/libavfilter/af_anlms.c b/libavfilter/af_anlms.c index 9d3c44575b..1516cad0a5 100644 --- a/libavfilter/af_anlms.c +++ b/libavfilter/af_anlms.c @@ -162,7 +162,7 @@ static int activate(AVFilterContext *ctx) if (ff_outlink_frame_wanted(ctx->outputs[0])) { for (i = 0; i < 2; i++) { - if (ff_inlink_queued_samples(ctx->inputs[i]) > 0) + if (s->frame[i] || ff_inlink_queued_samples(ctx->inputs[i]) > 0) continue; ff_inlink_request_frame(ctx->inputs[i]); return 0; diff --git a/libavfilter/af_arls.c b/libavfilter/af_arls.c index 85e4f92425..301e963d8a 100644 --- a/libavfilter/af_arls.c +++ b/libavfilter/af_arls.c @@ -159,7 +159,7 @@ static int activate(AVFilterContext *ctx) if (ff_outlink_frame_wanted(ctx->outputs[0])) { for (i = 0; i < 2; i++) { - if (ff_inlink_queued_samples(ctx->inputs[i]) > 0) + if (s->frame[i] || ff_inlink_queued_samples(ctx->inputs[i]) > 0) continue; ff_inlink_request_frame(ctx->inputs[i]); return 0; diff --git a/libavfilter/af_asdr.c b/libavfilter/af_asdr.c index 3942f48053..bb5c30faba 100644 --- a/libavfilter/af_asdr.c +++ b/libavfilter/af_asdr.c @@ -186,11 +186,11 @@ static int activate(AVFilterContext *ctx) if (ff_outlink_frame_wanted(outlink)) { for (int i = 0; i < 2; i++) { - if (ff_inlink_queued_samples(ctx->inputs[i]) > 0) + if (s->cache[i] || ff_inlink_queued_samples(ctx->inputs[i]) > 0) continue; ff_inlink_request_frame(ctx->inputs[i]); + return 0; } - return 0; } return FFERROR_NOT_READY;

1 0