diff mbox series

[FFmpeg-devel] avfilter: add audio signal to distortion ratio filter

Message ID 20210912184401.18782-1-onemda@gmail.com
State New
Headers show
Series [FFmpeg-devel] avfilter: add audio signal to distortion ratio filter
Related show

Checks

Context Check Description
andriy/make_x86 success Make finished
andriy/make_fate_x86 success Make fate finished
andriy/make_ppc success Make finished
andriy/make_fate_ppc success Make fate finished

Commit Message

Paul B Mahol Sept. 12, 2021, 6:44 p.m. UTC
Signed-off-by: Paul B Mahol <onemda@gmail.com>
---
 doc/filters.texi         |   7 ++
 libavfilter/Makefile     |   1 +
 libavfilter/af_asdr.c    | 237 +++++++++++++++++++++++++++++++++++++++
 libavfilter/allfilters.c |   1 +
 4 files changed, 246 insertions(+)
 create mode 100644 libavfilter/af_asdr.c

Comments

Nicolas George Sept. 12, 2021, 7:06 p.m. UTC | #1
Paul B Mahol (12021-09-12):
> Signed-off-by: Paul B Mahol <onemda@gmail.com>
> ---
>  doc/filters.texi         |   7 ++
>  libavfilter/Makefile     |   1 +
>  libavfilter/af_asdr.c    | 237 +++++++++++++++++++++++++++++++++++++++
>  libavfilter/allfilters.c |   1 +
>  4 files changed, 246 insertions(+)
>  create mode 100644 libavfilter/af_asdr.c
> 
> diff --git a/doc/filters.texi b/doc/filters.texi
> index 8f20ccf8c6..6af7344820 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -2531,6 +2531,13 @@ noise removed from input signal.
>  
>  This filter supports the all above options as @ref{commands}.
>  
> +@section asdr
> +Measure Audio Signal-to-Distortion Ratio.
> +
> +This filter takes two audio streams for input, and outputs first
> +audio stream.
> +Results are in dB per channel at end of either input.
> +
>  @section asetnsamples
>  
>  Set the number of samples per each output audio frame.
> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> index 76c65c3f42..865252ef3f 100644
> --- a/libavfilter/Makefile
> +++ b/libavfilter/Makefile
> @@ -82,6 +82,7 @@ OBJS-$(CONFIG_AREALTIME_FILTER)              += f_realtime.o
>  OBJS-$(CONFIG_ARESAMPLE_FILTER)              += af_aresample.o
>  OBJS-$(CONFIG_AREVERSE_FILTER)               += f_reverse.o
>  OBJS-$(CONFIG_ARNNDN_FILTER)                 += af_arnndn.o
> +OBJS-$(CONFIG_ASDR_FILTER)                   += af_asdr.o
>  OBJS-$(CONFIG_ASEGMENT_FILTER)               += f_segment.o
>  OBJS-$(CONFIG_ASELECT_FILTER)                += f_select.o
>  OBJS-$(CONFIG_ASENDCMD_FILTER)               += f_sendcmd.o
> diff --git a/libavfilter/af_asdr.c b/libavfilter/af_asdr.c
> new file mode 100644
> index 0000000000..220da74a85
> --- /dev/null
> +++ b/libavfilter/af_asdr.c
> @@ -0,0 +1,237 @@
> +/*
> + * Copyright (c) 2021 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/audio_fifo.h"
> +#include "libavutil/channel_layout.h"
> +#include "libavutil/common.h"
> +#include "libavutil/opt.h"
> +
> +#include "audio.h"
> +#include "avfilter.h"
> +#include "formats.h"
> +#include "filters.h"
> +#include "internal.h"
> +
> +typedef struct AudioSDRContext {
> +    int channels;
> +    int64_t pts;
> +    double *sum_u;
> +    double *sum_uv;
> +
> +    AVFrame *cache[2];
> +
> +    AVAudioFifo *fifo[2];
> +} AudioSDRContext;
> +
> +static int query_formats(AVFilterContext *ctx)
> +{
> +    static const enum AVSampleFormat sample_fmts[] = {
> +        AV_SAMPLE_FMT_DBLP,
> +        AV_SAMPLE_FMT_NONE
> +    };
> +    int ret = ff_set_common_all_channel_counts(ctx);
> +    if (ret < 0)
> +        return ret;
> +
> +    ret = ff_set_common_formats_from_list(ctx, sample_fmts);
> +    if (ret < 0)
> +        return ret;
> +
> +    return ff_set_common_all_samplerates(ctx);
> +}
> +
> +static void sdr(AVFilterContext *ctx, const AVFrame *u, const AVFrame *v)
> +{
> +    AudioSDRContext *s = ctx->priv;
> +
> +    for (int ch = 0; ch < u->channels; ch++) {
> +        const double *const us = (double *)u->extended_data[ch];
> +        const double *const vs = (double *)v->extended_data[ch];
> +        double sum_uv = s->sum_uv[ch];
> +        double sum_u = s->sum_u[ch];
> +
> +        for (int n = 0; n < u->nb_samples; n++) {
> +            sum_u  += us[n] * us[n];
> +            sum_uv += (us[n] - vs[n]) * (us[n] - vs[n]);
> +        }
> +
> +        s->sum_uv[ch] = sum_uv;
> +        s->sum_u[ch]  = sum_u;
> +    }
> +}
> +
> +static int activate(AVFilterContext *ctx)
> +{
> +    AudioSDRContext *s = ctx->priv;
> +    AVFrame *frame = NULL;
> +    int ret, status;
> +    int available;
> +    int64_t pts;
> +
> +    FF_FILTER_FORWARD_STATUS_BACK_ALL(ctx->outputs[0], ctx);
> +

> +    for (int i = 0; i < 2; i++) {
> +        ret = ff_inlink_consume_frame(ctx->inputs[i], &frame);
> +        if (ret > 0) {
> +            if (s->pts == AV_NOPTS_VALUE)
> +                s->pts = frame->pts;
> +            ret = av_audio_fifo_write(s->fifo[i], (void **)frame->extended_data,
> +                                      frame->nb_samples);
> +            av_frame_free(&frame);
> +            if (ret < 0)
> +                return ret;
> +        }
> +    }
> +
> +    available = FFMIN(av_audio_fifo_size(s->fifo[0]), av_audio_fifo_size(s->fifo[1]));

Since you are using activate, the links already contain a FIFO that you
can access, you should not copy the samples into a second FIFO.

Instead, apply the FFMIN to ff_inlink_queued_samples(), and
ff_inlink_consume_samples() on the result.

Also, call ff_inlink_request_frame() if one link has samples available
but not the other.

> +    if (available > 0) {
> +        AVFrame *out;
> +
> +        if (!s->cache[0] || s->cache[0]->nb_samples < available) {
> +            av_frame_free(&s->cache[0]);
> +            s->cache[0] = ff_get_audio_buffer(ctx->outputs[0], available);
> +            if (!s->cache[0])
> +                return AVERROR(ENOMEM);
> +        }
> +
> +        if (!s->cache[1] || s->cache[1]->nb_samples < available) {
> +            av_frame_free(&s->cache[1]);
> +            s->cache[1] = ff_get_audio_buffer(ctx->outputs[0], available);
> +            if (!s->cache[1])
> +                return AVERROR(ENOMEM);
> +        }
> +
> +        ret = av_audio_fifo_peek(s->fifo[0], (void **)s->cache[0]->extended_data, available);
> +        if (ret < 0)
> +            return ret;
> +
> +        ret = av_audio_fifo_peek(s->fifo[1], (void **)s->cache[1]->extended_data, available);
> +        if (ret < 0)
> +            return ret;
> +
> +        sdr(ctx, s->cache[0], s->cache[1]);
> +
> +        out = av_frame_clone(s->cache[0]);
> +        out->nb_samples = available;
> +        out->pts = s->pts;
> +        s->pts += available;
> +
> +        av_audio_fifo_drain(s->fifo[0], available);
> +        av_audio_fifo_drain(s->fifo[1], available);
> +
> +        return ff_filter_frame(ctx->outputs[0], out);
> +    }
> +
> +    for (int 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 (av_audio_fifo_size(s->fifo[0]) > 0 &&
> +        av_audio_fifo_size(s->fifo[1]) > 0) {
> +        ff_filter_set_ready(ctx, 10);
> +        return 0;
> +    }
> +
> +    if (ff_outlink_frame_wanted(ctx->outputs[0])) {
> +        for (int i = 0; i < 2; i++) {
> +            if (av_audio_fifo_size(s->fifo[i]) > 0)
> +                continue;
> +            ff_inlink_request_frame(ctx->inputs[i]);
> +            return 0;
> +        }
> +    }
> +
> +    return FFERROR_NOT_READY;
> +}
> +
> +static int config_output(AVFilterLink *outlink)
> +{
> +    AVFilterContext *ctx = outlink->src;
> +    AVFilterLink *inlink = ctx->inputs[0];
> +    AudioSDRContext *s = ctx->priv;
> +
> +    s->pts = AV_NOPTS_VALUE;
> +
> +    s->channels = inlink->channels;
> +    outlink->format = inlink->format;
> +    outlink->channels = inlink->channels;
> +
> +    s->fifo[0] = av_audio_fifo_alloc(outlink->format, outlink->channels, 1024);
> +    s->fifo[1] = av_audio_fifo_alloc(outlink->format, outlink->channels, 1024);
> +    if (!s->fifo[0] || !s->fifo[1])
> +        return AVERROR(ENOMEM);
> +
> +    s->sum_u  = av_calloc(outlink->channels, sizeof(*s->sum_u));
> +    s->sum_uv = av_calloc(outlink->channels, sizeof(*s->sum_uv));
> +    if (!s->sum_u || !s->sum_uv)
> +        return AVERROR(ENOMEM);
> +
> +    return 0;
> +}
> +
> +static av_cold void uninit(AVFilterContext *ctx)
> +{
> +    AudioSDRContext *s = ctx->priv;
> +
> +    for (int ch = 0; ch < s->channels; ch++)
> +        av_log(ctx, AV_LOG_INFO, "SDR ch%d: %g dB\n", ch, 20. * log10(s->sum_u[ch] / s->sum_uv[ch]));
> +
> +    av_audio_fifo_free(s->fifo[0]);
> +    av_audio_fifo_free(s->fifo[1]);
> +
> +    av_frame_free(&s->cache[0]);
> +    av_frame_free(&s->cache[1]);
> +
> +    av_freep(&s->sum_u);
> +    av_freep(&s->sum_uv);
> +}
> +
> +static const AVFilterPad inputs[] = {
> +    {
> +        .name = "input0",
> +        .type = AVMEDIA_TYPE_AUDIO,
> +    },
> +    {
> +        .name = "input1",
> +        .type = AVMEDIA_TYPE_AUDIO,
> +    },
> +};
> +
> +static const AVFilterPad outputs[] = {
> +    {
> +        .name         = "default",
> +        .type         = AVMEDIA_TYPE_AUDIO,
> +        .config_props = config_output,
> +    },
> +};
> +
> +const AVFilter ff_af_asdr = {
> +    .name           = "asdr",
> +    .description    = NULL_IF_CONFIG_SMALL("Measure Audio Signal-to-Distortion Ratio."),
> +    .priv_size      = sizeof(AudioSDRContext),
> +    .query_formats  = query_formats,
> +    .activate       = activate,
> +    .uninit         = uninit,
> +    FILTER_INPUTS(inputs),
> +    FILTER_OUTPUTS(outputs),
> +};
> diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> index 73a0bf9c44..7234ca6dbe 100644
> --- a/libavfilter/allfilters.c
> +++ b/libavfilter/allfilters.c
> @@ -75,6 +75,7 @@ extern const AVFilter ff_af_arealtime;
>  extern const AVFilter ff_af_aresample;
>  extern const AVFilter ff_af_areverse;
>  extern const AVFilter ff_af_arnndn;
> +extern const AVFilter ff_af_asdr;
>  extern const AVFilter ff_af_asegment;
>  extern const AVFilter ff_af_aselect;
>  extern const AVFilter ff_af_asendcmd;

Regards,
diff mbox series

Patch

diff --git a/doc/filters.texi b/doc/filters.texi
index 8f20ccf8c6..6af7344820 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -2531,6 +2531,13 @@  noise removed from input signal.
 
 This filter supports the all above options as @ref{commands}.
 
+@section asdr
+Measure Audio Signal-to-Distortion Ratio.
+
+This filter takes two audio streams for input, and outputs first
+audio stream.
+Results are in dB per channel at end of either input.
+
 @section asetnsamples
 
 Set the number of samples per each output audio frame.
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 76c65c3f42..865252ef3f 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -82,6 +82,7 @@  OBJS-$(CONFIG_AREALTIME_FILTER)              += f_realtime.o
 OBJS-$(CONFIG_ARESAMPLE_FILTER)              += af_aresample.o
 OBJS-$(CONFIG_AREVERSE_FILTER)               += f_reverse.o
 OBJS-$(CONFIG_ARNNDN_FILTER)                 += af_arnndn.o
+OBJS-$(CONFIG_ASDR_FILTER)                   += af_asdr.o
 OBJS-$(CONFIG_ASEGMENT_FILTER)               += f_segment.o
 OBJS-$(CONFIG_ASELECT_FILTER)                += f_select.o
 OBJS-$(CONFIG_ASENDCMD_FILTER)               += f_sendcmd.o
diff --git a/libavfilter/af_asdr.c b/libavfilter/af_asdr.c
new file mode 100644
index 0000000000..220da74a85
--- /dev/null
+++ b/libavfilter/af_asdr.c
@@ -0,0 +1,237 @@ 
+/*
+ * Copyright (c) 2021 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/audio_fifo.h"
+#include "libavutil/channel_layout.h"
+#include "libavutil/common.h"
+#include "libavutil/opt.h"
+
+#include "audio.h"
+#include "avfilter.h"
+#include "formats.h"
+#include "filters.h"
+#include "internal.h"
+
+typedef struct AudioSDRContext {
+    int channels;
+    int64_t pts;
+    double *sum_u;
+    double *sum_uv;
+
+    AVFrame *cache[2];
+
+    AVAudioFifo *fifo[2];
+} AudioSDRContext;
+
+static int query_formats(AVFilterContext *ctx)
+{
+    static const enum AVSampleFormat sample_fmts[] = {
+        AV_SAMPLE_FMT_DBLP,
+        AV_SAMPLE_FMT_NONE
+    };
+    int ret = ff_set_common_all_channel_counts(ctx);
+    if (ret < 0)
+        return ret;
+
+    ret = ff_set_common_formats_from_list(ctx, sample_fmts);
+    if (ret < 0)
+        return ret;
+
+    return ff_set_common_all_samplerates(ctx);
+}
+
+static void sdr(AVFilterContext *ctx, const AVFrame *u, const AVFrame *v)
+{
+    AudioSDRContext *s = ctx->priv;
+
+    for (int ch = 0; ch < u->channels; ch++) {
+        const double *const us = (double *)u->extended_data[ch];
+        const double *const vs = (double *)v->extended_data[ch];
+        double sum_uv = s->sum_uv[ch];
+        double sum_u = s->sum_u[ch];
+
+        for (int n = 0; n < u->nb_samples; n++) {
+            sum_u  += us[n] * us[n];
+            sum_uv += (us[n] - vs[n]) * (us[n] - vs[n]);
+        }
+
+        s->sum_uv[ch] = sum_uv;
+        s->sum_u[ch]  = sum_u;
+    }
+}
+
+static int activate(AVFilterContext *ctx)
+{
+    AudioSDRContext *s = ctx->priv;
+    AVFrame *frame = NULL;
+    int ret, status;
+    int available;
+    int64_t pts;
+
+    FF_FILTER_FORWARD_STATUS_BACK_ALL(ctx->outputs[0], ctx);
+
+    for (int i = 0; i < 2; i++) {
+        ret = ff_inlink_consume_frame(ctx->inputs[i], &frame);
+        if (ret > 0) {
+            if (s->pts == AV_NOPTS_VALUE)
+                s->pts = frame->pts;
+            ret = av_audio_fifo_write(s->fifo[i], (void **)frame->extended_data,
+                                      frame->nb_samples);
+            av_frame_free(&frame);
+            if (ret < 0)
+                return ret;
+        }
+    }
+
+    available = FFMIN(av_audio_fifo_size(s->fifo[0]), av_audio_fifo_size(s->fifo[1]));
+    if (available > 0) {
+        AVFrame *out;
+
+        if (!s->cache[0] || s->cache[0]->nb_samples < available) {
+            av_frame_free(&s->cache[0]);
+            s->cache[0] = ff_get_audio_buffer(ctx->outputs[0], available);
+            if (!s->cache[0])
+                return AVERROR(ENOMEM);
+        }
+
+        if (!s->cache[1] || s->cache[1]->nb_samples < available) {
+            av_frame_free(&s->cache[1]);
+            s->cache[1] = ff_get_audio_buffer(ctx->outputs[0], available);
+            if (!s->cache[1])
+                return AVERROR(ENOMEM);
+        }
+
+        ret = av_audio_fifo_peek(s->fifo[0], (void **)s->cache[0]->extended_data, available);
+        if (ret < 0)
+            return ret;
+
+        ret = av_audio_fifo_peek(s->fifo[1], (void **)s->cache[1]->extended_data, available);
+        if (ret < 0)
+            return ret;
+
+        sdr(ctx, s->cache[0], s->cache[1]);
+
+        out = av_frame_clone(s->cache[0]);
+        out->nb_samples = available;
+        out->pts = s->pts;
+        s->pts += available;
+
+        av_audio_fifo_drain(s->fifo[0], available);
+        av_audio_fifo_drain(s->fifo[1], available);
+
+        return ff_filter_frame(ctx->outputs[0], out);
+    }
+
+    for (int 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 (av_audio_fifo_size(s->fifo[0]) > 0 &&
+        av_audio_fifo_size(s->fifo[1]) > 0) {
+        ff_filter_set_ready(ctx, 10);
+        return 0;
+    }
+
+    if (ff_outlink_frame_wanted(ctx->outputs[0])) {
+        for (int i = 0; i < 2; i++) {
+            if (av_audio_fifo_size(s->fifo[i]) > 0)
+                continue;
+            ff_inlink_request_frame(ctx->inputs[i]);
+            return 0;
+        }
+    }
+
+    return FFERROR_NOT_READY;
+}
+
+static int config_output(AVFilterLink *outlink)
+{
+    AVFilterContext *ctx = outlink->src;
+    AVFilterLink *inlink = ctx->inputs[0];
+    AudioSDRContext *s = ctx->priv;
+
+    s->pts = AV_NOPTS_VALUE;
+
+    s->channels = inlink->channels;
+    outlink->format = inlink->format;
+    outlink->channels = inlink->channels;
+
+    s->fifo[0] = av_audio_fifo_alloc(outlink->format, outlink->channels, 1024);
+    s->fifo[1] = av_audio_fifo_alloc(outlink->format, outlink->channels, 1024);
+    if (!s->fifo[0] || !s->fifo[1])
+        return AVERROR(ENOMEM);
+
+    s->sum_u  = av_calloc(outlink->channels, sizeof(*s->sum_u));
+    s->sum_uv = av_calloc(outlink->channels, sizeof(*s->sum_uv));
+    if (!s->sum_u || !s->sum_uv)
+        return AVERROR(ENOMEM);
+
+    return 0;
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    AudioSDRContext *s = ctx->priv;
+
+    for (int ch = 0; ch < s->channels; ch++)
+        av_log(ctx, AV_LOG_INFO, "SDR ch%d: %g dB\n", ch, 20. * log10(s->sum_u[ch] / s->sum_uv[ch]));
+
+    av_audio_fifo_free(s->fifo[0]);
+    av_audio_fifo_free(s->fifo[1]);
+
+    av_frame_free(&s->cache[0]);
+    av_frame_free(&s->cache[1]);
+
+    av_freep(&s->sum_u);
+    av_freep(&s->sum_uv);
+}
+
+static const AVFilterPad inputs[] = {
+    {
+        .name = "input0",
+        .type = AVMEDIA_TYPE_AUDIO,
+    },
+    {
+        .name = "input1",
+        .type = AVMEDIA_TYPE_AUDIO,
+    },
+};
+
+static const AVFilterPad outputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_AUDIO,
+        .config_props = config_output,
+    },
+};
+
+const AVFilter ff_af_asdr = {
+    .name           = "asdr",
+    .description    = NULL_IF_CONFIG_SMALL("Measure Audio Signal-to-Distortion Ratio."),
+    .priv_size      = sizeof(AudioSDRContext),
+    .query_formats  = query_formats,
+    .activate       = activate,
+    .uninit         = uninit,
+    FILTER_INPUTS(inputs),
+    FILTER_OUTPUTS(outputs),
+};
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 73a0bf9c44..7234ca6dbe 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -75,6 +75,7 @@  extern const AVFilter ff_af_arealtime;
 extern const AVFilter ff_af_aresample;
 extern const AVFilter ff_af_areverse;
 extern const AVFilter ff_af_arnndn;
+extern const AVFilter ff_af_asdr;
 extern const AVFilter ff_af_asegment;
 extern const AVFilter ff_af_aselect;
 extern const AVFilter ff_af_asendcmd;