--- Is this correct?
--- libavfilter/Makefile | 1 + libavfilter/allfilters.c | 1 + libavfilter/avf_showfreqs.c | 333 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 335 insertions(+) create mode 100644 libavfilter/avf_showfreqs.c diff --git a/libavfilter/Makefile b/libavfilter/Makefile index 1865e49..049bbe8 100644 --- a/libavfilter/Makefile +++ b/libavfilter/Makefile @@ -256,6 +256,7 @@ OBJS-$(CONFIG_ADRAWGRAPH_FILTER) += f_drawgraph.o OBJS-$(CONFIG_AVECTORSCOPE_FILTER) += avf_avectorscope.o OBJS-$(CONFIG_CONCAT_FILTER) += avf_concat.o OBJS-$(CONFIG_SHOWCQT_FILTER) += avf_showcqt.o +OBJS-$(CONFIG_SHOWFREQS_FILTER) += avf_showfreqs.o OBJS-$(CONFIG_SHOWSPECTRUM_FILTER) += avf_showspectrum.o OBJS-$(CONFIG_SHOWVOLUME_FILTER) += avf_showvolume.o OBJS-$(CONFIG_SHOWWAVES_FILTER) += avf_showwaves.o diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c index f8e97bc..63a274a 100644 --- a/libavfilter/allfilters.c +++ b/libavfilter/allfilters.c @@ -271,6 +271,7 @@ void avfilter_register_all(void) REGISTER_FILTER(AVECTORSCOPE, avectorscope, avf); REGISTER_FILTER(CONCAT, concat, avf); REGISTER_FILTER(SHOWCQT, showcqt, avf); + REGISTER_FILTER(SHOWFREQS, showfreqs, avf); REGISTER_FILTER(SHOWSPECTRUM, showspectrum, avf); REGISTER_FILTER(SHOWVOLUME, showvolume, avf); REGISTER_FILTER(SHOWWAVES, showwaves, avf); diff --git a/libavfilter/avf_showfreqs.c b/libavfilter/avf_showfreqs.c new file mode 100644 index 0000000..54902b9 --- /dev/null +++ b/libavfilter/avf_showfreqs.c @@ -0,0 +1,333 @@ +/* + * Copyright (c) 2015 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 <math.h> + +#include "libavcodec/avfft.h" +#include "libavutil/avassert.h" +#include "libavutil/channel_layout.h" +#include "libavutil/intreadwrite.h" +#include "libavutil/opt.h" +#include "avfilter.h" +#include "internal.h" + +enum DisplayMode { LINE, BAR, DOT, NB_MODES }; +enum DisplayScale { LINEAR, SQRT, CBRT, LOG, NB_SCALES }; +enum WindowFunc { WFUNC_NONE, WFUNC_HANN, WFUNC_HAMMING, WFUNC_BLACKMAN, NB_WFUNC }; + +typedef struct ShowFreqsContext { + const AVClass *class; + int w, h; + RDFTContext *rdft; ///< Real Discrete Fourier Transform context + int rdft_bits; ///< number of bits (RDFT window size = 1<<rdft_bits) + FFTSample **rdft_data; ///< bins holder for each (displayed) channels + float *window_func_lut; ///< Window function LUT + int win_func; + int nb_channels; + int mode; + int scale; +} ShowFreqsContext; + +#define OFFSET(x) offsetof(ShowFreqsContext, x) +#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM + +static const AVOption showfreqs_options[] = { + { "size", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str = "1024x512"}, 0, 0, FLAGS }, + { "s", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str = "1024x512"}, 0, 0, FLAGS }, + { "mode", "set display mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=LINE}, LINEAR, NB_MODES-1, FLAGS, "mode" }, + { "line", "show lines", 0, AV_OPT_TYPE_CONST, {.i64=LINE}, 0, 0, FLAGS, "mode" }, + { "bar", "show bars", 0, AV_OPT_TYPE_CONST, {.i64=BAR}, 0, 0, FLAGS, "mode" }, + { "dot", "show dots", 0, AV_OPT_TYPE_CONST, {.i64=DOT}, 0, 0, FLAGS, "mode" }, + { "scale", "set display scale", OFFSET(scale), AV_OPT_TYPE_INT, {.i64=SQRT}, LINEAR, NB_SCALES-1, FLAGS, "scale" }, + { "sqrt", "square root", 0, AV_OPT_TYPE_CONST, {.i64=SQRT}, 0, 0, FLAGS, "scale" }, + { "cbrt", "cubic root", 0, AV_OPT_TYPE_CONST, {.i64=CBRT}, 0, 0, FLAGS, "scale" }, + { "log", "logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=LOG}, 0, 0, FLAGS, "scale" }, + { "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=LINEAR}, 0, 0, FLAGS, "scale" }, + { "win_func", "set window function", OFFSET(win_func), AV_OPT_TYPE_INT, {.i64 = WFUNC_HANN}, 0, NB_WFUNC-1, FLAGS, "win_func" }, + { "hann", "Hann window", 0, AV_OPT_TYPE_CONST, {.i64 = WFUNC_HANN}, 0, 0, FLAGS, "win_func" }, + { "hamming", "Hamming window", 0, AV_OPT_TYPE_CONST, {.i64 = WFUNC_HAMMING}, 0, 0, FLAGS, "win_func" }, + { "blackman", "Blackman window", 0, AV_OPT_TYPE_CONST, {.i64 = WFUNC_BLACKMAN}, 0, 0, FLAGS, "win_func" }, + { NULL } +}; + +AVFILTER_DEFINE_CLASS(showfreqs); + +static int query_formats(AVFilterContext *ctx) +{ + AVFilterFormats *formats = NULL; + AVFilterChannelLayouts *layouts = NULL; + AVFilterLink *inlink = ctx->inputs[0]; + AVFilterLink *outlink = ctx->outputs[0]; + static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_NONE }; + static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_RGBA, AV_PIX_FMT_NONE }; + + /* set input audio formats */ + formats = ff_make_format_list(sample_fmts); + if (!formats) + return AVERROR(ENOMEM); + ff_formats_ref(formats, &inlink->out_formats); + + layouts = ff_all_channel_layouts(); + if (!layouts) + return AVERROR(ENOMEM); + ff_channel_layouts_ref(layouts, &inlink->out_channel_layouts); + + formats = ff_all_samplerates(); + if (!formats) + return AVERROR(ENOMEM); + ff_formats_ref(formats, &inlink->out_samplerates); + + /* set output video format */ + formats = ff_make_format_list(pix_fmts); + if (!formats) + return AVERROR(ENOMEM); + ff_formats_ref(formats, &outlink->in_formats); + + return 0; +} + +static int config_output(AVFilterLink *outlink) +{ + AVFilterContext *ctx = outlink->src; + AVFilterLink *inlink = ctx->inputs[0]; + ShowFreqsContext *s = ctx->priv; + unsigned win_size; + int i, rdft_bits; + + outlink->w = s->w; + outlink->h = s->h; + + /* RDFT window size (precision) according to the requested output frame height */ + for (rdft_bits = 1; 1 << rdft_bits < 2 * outlink->w; rdft_bits++); + win_size = 1 << rdft_bits; + + /* (re-)configuration if the video output changed (or first init) */ + if (rdft_bits != s->rdft_bits) { + size_t rdft_size, rdft_listsize; + + av_rdft_end(s->rdft); + s->rdft = av_rdft_init(rdft_bits, DFT_R2C); + if (!s->rdft) { + av_log(ctx, AV_LOG_ERROR, "Unable to create RDFT context. " + "The window size might be too high.\n"); + return AVERROR(EINVAL); + } + s->rdft_bits = rdft_bits; + + /* RDFT buffers: x2 for each (display) channel buffer. + * Note: we use free and malloc instead of a realloc-like function to + * make sure the buffer is aligned in memory for the FFT functions. */ + for (i = 0; i < s->nb_channels; i++) + av_freep(&s->rdft_data[i]); + av_freep(&s->rdft_data); + s->nb_channels = inlink->channels; + + if (av_size_mult(sizeof(*s->rdft_data), + s->nb_channels, &rdft_listsize) < 0) + return AVERROR(EINVAL); + if (av_size_mult(sizeof(**s->rdft_data), + win_size, &rdft_size) < 0) + return AVERROR(EINVAL); + s->rdft_data = av_malloc(rdft_listsize); + if (!s->rdft_data) + return AVERROR(ENOMEM); + for (i = 0; i < s->nb_channels; i++) { + s->rdft_data[i] = av_malloc(rdft_size); + if (!s->rdft_data[i]) + return AVERROR(ENOMEM); + } + + /* pre-calc windowing function */ + s->window_func_lut = + av_realloc_f(s->window_func_lut, win_size, + sizeof(*s->window_func_lut)); + if (!s->window_func_lut) + return AVERROR(ENOMEM); + switch (s->win_func) { + case WFUNC_NONE: + for (i = 0; i < win_size; i++) + s->window_func_lut[i] = 1.; + break; + case WFUNC_HANN: + for (i = 0; i < win_size; i++) + s->window_func_lut[i] = .5f * (1 - cos(2*M_PI*i / (win_size-1))); + break; + case WFUNC_HAMMING: + for (i = 0; i < win_size; i++) + s->window_func_lut[i] = .54f - .46f * cos(2*M_PI*i / (win_size-1)); + break; + case WFUNC_BLACKMAN: { + for (i = 0; i < win_size; i++) + s->window_func_lut[i] = .42f - .5f*cos(2*M_PI*i / (win_size-1)) + .08f*cos(4*M_PI*i / (win_size-1)); + break; + } + default: + av_assert0(0); + } + + outlink->sample_aspect_ratio = (AVRational){1,1}; + } + + outlink->frame_rate = av_make_q(inlink->sample_rate, win_size); + + inlink->min_samples = inlink->max_samples = inlink->partial_buf_size = + win_size; + + av_log(ctx, AV_LOG_VERBOSE, "s:%dx%d RDFT window size:%d\n", + s->w, s->h, win_size); + return 0; +} + +static int plot_freqs(AVFilterLink *inlink, AVFrame *insamples) +{ + AVFilterContext *ctx = inlink->dst; + AVFilterLink *outlink = ctx->outputs[0]; + ShowFreqsContext *s = ctx->priv; + AVFrame *out = ff_get_video_buffer(outlink, outlink->w, outlink->h); + + /* nb_freq contains the power of two superior or equal to the output image + * height (or half the RDFT window size) */ + const int nb_freq = 1 << (s->rdft_bits - 1); + const int win_size = nb_freq << 1; + const double w = 1. / (nb_freq * 32768.); + int ch, n, x; + + /* fill RDFT input with the number of samples available */ + for (ch = 0; ch < s->nb_channels; ch++) { + const int16_t *p = (int16_t *)insamples->extended_data[ch]; + + for (n = 0; n < insamples->nb_samples; n++) + s->rdft_data[ch][n] = p[n] * s->window_func_lut[n]; + for (; n < win_size; n++) + s->rdft_data[ch][n] = 0; + } + + /* run RDFT on each samples set */ + for (ch = 0; ch < s->nb_channels; ch++) + av_rdft_calc(s->rdft, s->rdft_data[ch]); + +#define RE(x, ch) s->rdft_data[ch][2 * (x) + 0] +#define IM(x, ch) s->rdft_data[ch][2 * (x) + 1] +#define MAGNITUDE(x, ch) hypot(RE(x, ch), IM(x, ch)) + + for (ch = 0; ch < s->nb_channels; ch++) { + for (x = 0; x < outlink->w; x++) { + double a = w * MAGNITUDE(x, ch); + int y, prev_y, i; + + switch(s->scale) { + case SQRT: + a = 1.0 - sqrt(a); + break; + case CBRT: + a = 1.0 - cbrt(a); + break; + case LOG: + a = log(FFMAX(FFMIN(1, a), 1e-6)) / log(1e-6); // zero = -120dBFS + break; + case LINEAR: + a = 1.0 - a; + break; + } + y = a * (outlink->h - 1); + + switch(s->mode) { + case LINE: + if (x == 0) { + prev_y = y; + } + if (y <= prev_y) { + for (i = y; i <= prev_y; i++) + AV_WL32(out->data[0] + i * out->linesize[0] + x * 4, 0xffffffff); + } else { + for (i = prev_y; i <= y; i++) + AV_WL32(out->data[0] + i * out->linesize[0] + x * 4, 0xffffffff); + } + prev_y = y; + break; + case BAR: + for (; y < outlink->h; y++) + AV_WL32(out->data[0] + y * out->linesize[0] + x * 4, 0xffffffff); + break; + case DOT: + AV_WL32(out->data[0] + y * out->linesize[0] + x * 4, 0xffffffff); + break; + } + } + } + + out->pts = insamples->pts; + return ff_filter_frame(outlink, out); +} + +static int filter_frame(AVFilterLink *inlink, AVFrame *insamples) +{ + AVFilterContext *ctx = inlink->dst; + ShowFreqsContext *s = ctx->priv; + unsigned win_size = 1 << s->rdft_bits; + int ret; + + av_assert0(insamples->nb_samples <= win_size); + ret = plot_freqs(inlink, insamples); + + av_frame_free(&insamples); + return ret; +} + +static av_cold void uninit(AVFilterContext *ctx) +{ + ShowFreqsContext *s = ctx->priv; + int i; + + av_rdft_end(s->rdft); + for (i = 0; i < s->nb_channels; i++) + av_freep(&s->rdft_data[i]); + av_freep(&s->rdft_data); + av_freep(&s->window_func_lut); +} + +static const AVFilterPad showfreqs_inputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_AUDIO, + .filter_frame = filter_frame, + }, + { NULL } +}; + +static const AVFilterPad showfreqs_outputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + .config_props = config_output, + }, + { NULL } +}; + +AVFilter ff_avf_showfreqs = { + .name = "showfreqs", + .description = NULL_IF_CONFIG_SMALL("Convert input audio to a frequencies video output."), + .uninit = uninit, + .query_formats = query_formats, + .priv_size = sizeof(ShowFreqsContext), + .inputs = showfreqs_inputs, + .outputs = showfreqs_outputs, + .priv_class = &showfreqs_class, +}; -- 2.4.5 _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel