from:"Dmitry Gumenyuk"

Re: [FFmpeg-devel] [PATCH v2] avfilter: add dumpwave filter.

2018-01-12 Thread Dmitry Gumenyuk


> On 12 Jan 2018, at 08:36, Kyle Swanson  wrote:
> 
> Hi,
> 
> Make sure you go back and read our comments. There were several things you
> didn't address in your most recent patch.
Sorry, seems I missed few emails. Sure
> Thanks,
> Kyle
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel


smime.p7s
Description: S/MIME cryptographic signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

2018-01-11 Thread Dmitry Gumenyuk



--
Best regards,
Dmytro

> On 11 Jan 2018, at 23:57, Paul B Mahol <one...@gmail.com> wrote:
> 
>> On 1/11/18, Dmitry Gumenyuk <dmitry.gumen...@gmail.com> wrote:
>> 
>>>> On 11 Jan 2018, at 23:02, Paul B Mahol <one...@gmail.com> wrote:
>>>> 
>>>> On 1/11/18, Dmitry Gumenyuk <dmitry.gumen...@gmail.com> wrote:
>>>> 
>>>>>> On 11 Jan 2018, at 22:41, Paul B Mahol <one...@gmail.com> wrote:
>>>>>> 
>>>>>> On 1/11/18, Dmitry Gumenyuk <dmitry.gumen...@gmail.com> wrote:
>>>>>> Hello
>>>>>> 
>>>>>> 2018-01-10 11:33 GMT+01:00 Moritz Barsnick <barsn...@gmx.net>:
>>>>>>>> On Wed, Jan 10, 2018 at 08:58:05 +0100, dmitry.gumen...@gmail.com
>>>>>>>> wrote:
>>>>>>>> 
>>>>>>>> +@table @option
>>>>>>>> +@item d
>>>>>>>> +Dimensions @code{WxH}.
>>>>>>>> +@code{W} - number of data values in json, values will be scaled
>>>>>>>> according to @code{H}.
>>>>>>>> +The default value is @var{640x480}
>>>>>>>> +
>>>>>>>> +@item s
>>>>>>>> +Samples count per value per channel
>>>>>>> 
>>>>>>> In most other filters/filtersources, s+size is used for dimensions,
>>>>>>> d+duration for time, and n for an amount/number of frames/samples or
>>>>>>> so. Might be a good idea to align with this. And use aliases (i.e.
>>>>>>> implement both "s" and "size").
>>>>>>> 
>>>>>>>> +static const AVOption dumpwave_options[] = {
>>>>>>>> +{ "d", "set width and height", OFFSET(w),
>>>>>>>> AV_OPT_TYPE_IMAGE_SIZE,
>>>>>>>> {.str = "640x480"}, 0, 0, FLAGS },
>>>>>>>> +{ "s", "set number of samples per value per channel",
>>>>>>>> OFFSET(s),
>>>>>>>> AV_OPT_TYPE_INT64,  {.i64 = 0}, 0, INT64_MAX, FLAGS },
>>>>>>>> +{ "json", "set json dump file", OFFSET(json),
>>>>>>>> AV_OPT_TYPE_STRING,
>>>>>>>> {
>>>>>>>> .str = NULL }, 0, 0, FLAGS },
>>>>>>>> +{ NULL }
>>>>>>> [...]
>>>>>>>> +static av_cold int init(AVFilterContext *ctx)
>>>>>>>> +{
>>>>>>>> +DumpWaveContext *dumpwave = ctx->priv;
>>>>>>>> +if(!dumpwave->s) {
>>>>>>>> +dumpwave->is_disabled = 1;
>>>>>>>> +av_log(ctx, AV_LOG_ERROR, "Invalid samples per value
>>>>>>>> config\n");
>>>>>>>> +}
>>>>>>>> +return 0;
>>>>>>> 
>>>>>>> I don't like the idea of having to provide the "s" parameter. Is
>>>>>>> there
>>>>>>> really no useful default?
>>>>>>> 
>>>>>>> And now, if s=0, what does the filter do? Just sit around? Couldn't
>>>>>>> it
>>>>>>> fail instead?
>>>>>>> 
>>>>>>> Apart from that:
>>>>>>> "if (" is the bracket style ffmpeg uses.
>>>>>>> 
>>>>>>>> +if (dumpwave->json && !(dump_fp =
>>>>>>>> av_fopen_utf8(dumpwave->json,
>>>>>>>> "w")))
>>>>>>>> +av_log(ctx, AV_LOG_WARNING, "Flushing dump failed\n");
>>>>>>> 
>>>>>>> I would appreciate evaluation of errno and printing the appropriate
>>>>>>> string (using av_strerror(), I believe).
>>>>>>> 
>>>>>>>> +switch (inlink->format) {
>>>>>>>> +case AV_SAMPLE_FMT_DBLP:
>>>>>>> 
>>>>>>> As Kyle hinted: Can this not force a conversion (implicit insertion
>>>>>>> of
>>>>>>> aformat filter) to e.g. double by only supporting this format, and
>>>>>>> reducing this switch to one or two cases? (The filter's output isn't
>>>>>>> really meant for transparent reuse anyway? af_volumedetect e.g. also
>>>>>>> supports only one, meaning its output can be a different format than
>>>>>>> its input.) It's also really hard to go through and later to
>>>>>>> maintain.
>>>>>>> It the big switch/case remains, one or two code macros would be
>>>>>>> appropriate.
>>>>>> 
>>>>>> I checked solution used in volumedetect and couldn't find a way to
>>>>>> read across formats.
>>>>> 
>>>>> I do not understand what you are trying to do.
>>>> Sorry, I'm trying to add support for 8, 16, 24, 32, 64 bit sample
>>>> formats
>>>>>> How would you implement such macros? Since version 3.2 astats filter
>>>>>> uses same approach for reading different formats and as far as I know
>>>>>> macros harder to debug
>>>>> 
>>>>> astats is using all formats because of numerous reasons. astats uses
>>>>> raw
>>>>> values,
>>>>> your filter just convert each raw value to float representation.
>>>> Is this wrong, as I'd like to have high precision?
>>> 
>>> For rendering to small size image?
>> Data can be used for analysis as well. Any size I would say as user may
>> define size
> 
> Than use floats.
Ok, will give it a try
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel


smime.p7s
Description: S/MIME cryptographic signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

2018-01-11 Thread Dmitry Gumenyuk


> On 11 Jan 2018, at 23:02, Paul B Mahol <one...@gmail.com> wrote:
> 
>> On 1/11/18, Dmitry Gumenyuk <dmitry.gumen...@gmail.com> wrote:
>> 
>>>> On 11 Jan 2018, at 22:41, Paul B Mahol <one...@gmail.com> wrote:
>>>> 
>>>> On 1/11/18, Dmitry Gumenyuk <dmitry.gumen...@gmail.com> wrote:
>>>> Hello
>>>> 
>>>> 2018-01-10 11:33 GMT+01:00 Moritz Barsnick <barsn...@gmx.net>:
>>>>>> On Wed, Jan 10, 2018 at 08:58:05 +0100, dmitry.gumen...@gmail.com
>>>>>> wrote:
>>>>>> 
>>>>>> +@table @option
>>>>>> +@item d
>>>>>> +Dimensions @code{WxH}.
>>>>>> +@code{W} - number of data values in json, values will be scaled
>>>>>> according to @code{H}.
>>>>>> +The default value is @var{640x480}
>>>>>> +
>>>>>> +@item s
>>>>>> +Samples count per value per channel
>>>>> 
>>>>> In most other filters/filtersources, s+size is used for dimensions,
>>>>> d+duration for time, and n for an amount/number of frames/samples or
>>>>> so. Might be a good idea to align with this. And use aliases (i.e.
>>>>> implement both "s" and "size").
>>>>> 
>>>>>> +static const AVOption dumpwave_options[] = {
>>>>>> +{ "d", "set width and height", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE,
>>>>>> {.str = "640x480"}, 0, 0, FLAGS },
>>>>>> +{ "s", "set number of samples per value per channel",  OFFSET(s),
>>>>>> AV_OPT_TYPE_INT64,  {.i64 = 0}, 0, INT64_MAX, FLAGS },
>>>>>> +{ "json", "set json dump file", OFFSET(json), AV_OPT_TYPE_STRING,
>>>>>> {
>>>>>> .str = NULL }, 0, 0, FLAGS },
>>>>>> +{ NULL }
>>>>> [...]
>>>>>> +static av_cold int init(AVFilterContext *ctx)
>>>>>> +{
>>>>>> +DumpWaveContext *dumpwave = ctx->priv;
>>>>>> +if(!dumpwave->s) {
>>>>>> +dumpwave->is_disabled = 1;
>>>>>> +av_log(ctx, AV_LOG_ERROR, "Invalid samples per value
>>>>>> config\n");
>>>>>> +}
>>>>>> +return 0;
>>>>> 
>>>>> I don't like the idea of having to provide the "s" parameter. Is there
>>>>> really no useful default?
>>>>> 
>>>>> And now, if s=0, what does the filter do? Just sit around? Couldn't it
>>>>> fail instead?
>>>>> 
>>>>> Apart from that:
>>>>> "if (" is the bracket style ffmpeg uses.
>>>>> 
>>>>>> +if (dumpwave->json && !(dump_fp =
>>>>>> av_fopen_utf8(dumpwave->json,
>>>>>> "w")))
>>>>>> +av_log(ctx, AV_LOG_WARNING, "Flushing dump failed\n");
>>>>> 
>>>>> I would appreciate evaluation of errno and printing the appropriate
>>>>> string (using av_strerror(), I believe).
>>>>> 
>>>>>> +switch (inlink->format) {
>>>>>> +case AV_SAMPLE_FMT_DBLP:
>>>>> 
>>>>> As Kyle hinted: Can this not force a conversion (implicit insertion of
>>>>> aformat filter) to e.g. double by only supporting this format, and
>>>>> reducing this switch to one or two cases? (The filter's output isn't
>>>>> really meant for transparent reuse anyway? af_volumedetect e.g. also
>>>>> supports only one, meaning its output can be a different format than
>>>>> its input.) It's also really hard to go through and later to maintain.
>>>>> It the big switch/case remains, one or two code macros would be
>>>>> appropriate.
>>>> 
>>>> I checked solution used in volumedetect and couldn't find a way to
>>>> read across formats.
>>> 
>>> I do not understand what you are trying to do.
>> Sorry, I'm trying to add support for 8, 16, 24, 32, 64 bit sample formats
>>>> How would you implement such macros? Since version 3.2 astats filter
>>>> uses same approach for reading different formats and as far as I know
>>>> macros harder to debug
>>> 
>>> astats is using all formats because of numerous reasons. astats uses raw
>>> values,
>>> your filter just convert each raw value to float representation.
>> Is this wrong, as I'd like to have high precision?
> 
> For rendering to small size image?
Data can be used for analysis as well. Any size I would say as user may define 
size
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel


smime.p7s
Description: S/MIME cryptographic signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

2018-01-11 Thread Dmitry Gumenyuk


> On 11 Jan 2018, at 22:41, Paul B Mahol <one...@gmail.com> wrote:
> 
>> On 1/11/18, Dmitry Gumenyuk <dmitry.gumen...@gmail.com> wrote:
>> Hello
>> 
>> 2018-01-10 11:33 GMT+01:00 Moritz Barsnick <barsn...@gmx.net>:
>>>> On Wed, Jan 10, 2018 at 08:58:05 +0100, dmitry.gumen...@gmail.com wrote:
>>>> 
>>>> +@table @option
>>>> +@item d
>>>> +Dimensions @code{WxH}.
>>>> +@code{W} - number of data values in json, values will be scaled
>>>> according to @code{H}.
>>>> +The default value is @var{640x480}
>>>> +
>>>> +@item s
>>>> +Samples count per value per channel
>>> 
>>> In most other filters/filtersources, s+size is used for dimensions,
>>> d+duration for time, and n for an amount/number of frames/samples or
>>> so. Might be a good idea to align with this. And use aliases (i.e.
>>> implement both "s" and "size").
>>> 
>>>> +static const AVOption dumpwave_options[] = {
>>>> +{ "d", "set width and height", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE,
>>>> {.str = "640x480"}, 0, 0, FLAGS },
>>>> +{ "s", "set number of samples per value per channel",  OFFSET(s),
>>>> AV_OPT_TYPE_INT64,  {.i64 = 0}, 0, INT64_MAX, FLAGS },
>>>> +{ "json", "set json dump file", OFFSET(json), AV_OPT_TYPE_STRING, {
>>>> .str = NULL }, 0, 0, FLAGS },
>>>> +{ NULL }
>>> [...]
>>>> +static av_cold int init(AVFilterContext *ctx)
>>>> +{
>>>> +DumpWaveContext *dumpwave = ctx->priv;
>>>> +if(!dumpwave->s) {
>>>> +dumpwave->is_disabled = 1;
>>>> +av_log(ctx, AV_LOG_ERROR, "Invalid samples per value config\n");
>>>> +}
>>>> +return 0;
>>> 
>>> I don't like the idea of having to provide the "s" parameter. Is there
>>> really no useful default?
>>> 
>>> And now, if s=0, what does the filter do? Just sit around? Couldn't it
>>> fail instead?
>>> 
>>> Apart from that:
>>> "if (" is the bracket style ffmpeg uses.
>>> 
>>>> +if (dumpwave->json && !(dump_fp = av_fopen_utf8(dumpwave->json,
>>>> "w")))
>>>> +av_log(ctx, AV_LOG_WARNING, "Flushing dump failed\n");
>>> 
>>> I would appreciate evaluation of errno and printing the appropriate
>>> string (using av_strerror(), I believe).
>>> 
>>>> +switch (inlink->format) {
>>>> +case AV_SAMPLE_FMT_DBLP:
>>> 
>>> As Kyle hinted: Can this not force a conversion (implicit insertion of
>>> aformat filter) to e.g. double by only supporting this format, and
>>> reducing this switch to one or two cases? (The filter's output isn't
>>> really meant for transparent reuse anyway? af_volumedetect e.g. also
>>> supports only one, meaning its output can be a different format than
>>> its input.) It's also really hard to go through and later to maintain.
>>> It the big switch/case remains, one or two code macros would be
>>> appropriate.
>> 
>> I checked solution used in volumedetect and couldn't find a way to
>> read across formats.
> 
> I do not understand what you are trying to do.
Sorry, I'm trying to add support for 8, 16, 24, 32, 64 bit sample formats
>> How would you implement such macros? Since version 3.2 astats filter
>> uses same approach for reading different formats and as far as I know
>> macros harder to debug
> 
> astats is using all formats because of numerous reasons. astats uses raw 
> values,
> your filter just convert each raw value to float representation.
Is this wrong, as I'd like to have high precision?
> This filter does not need support for so many formats, it could
> support just s16.
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel


smime.p7s
Description: S/MIME cryptographic signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

2018-01-11 Thread Dmitry Gumenyuk

Hello

2018-01-10 11:33 GMT+01:00 Moritz Barsnick :
> On Wed, Jan 10, 2018 at 08:58:05 +0100, dmitry.gumen...@gmail.com wrote:
>
>> +@table @option
>> +@item d
>> +Dimensions @code{WxH}.
>> +@code{W} - number of data values in json, values will be scaled according 
>> to @code{H}.
>> +The default value is @var{640x480}
>> +
>> +@item s
>> +Samples count per value per channel
>
> In most other filters/filtersources, s+size is used for dimensions,
> d+duration for time, and n for an amount/number of frames/samples or
> so. Might be a good idea to align with this. And use aliases (i.e.
> implement both "s" and "size").
>
>> +static const AVOption dumpwave_options[] = {
>> +{ "d", "set width and height", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str 
>> = "640x480"}, 0, 0, FLAGS },
>> +{ "s", "set number of samples per value per channel",  OFFSET(s), 
>> AV_OPT_TYPE_INT64,  {.i64 = 0}, 0, INT64_MAX, FLAGS },
>> +{ "json", "set json dump file", OFFSET(json), AV_OPT_TYPE_STRING, { 
>> .str = NULL }, 0, 0, FLAGS },
>> +{ NULL }
> [...]
>> +static av_cold int init(AVFilterContext *ctx)
>> +{
>> +DumpWaveContext *dumpwave = ctx->priv;
>> +if(!dumpwave->s) {
>> +dumpwave->is_disabled = 1;
>> +av_log(ctx, AV_LOG_ERROR, "Invalid samples per value config\n");
>> +}
>> +return 0;
>
> I don't like the idea of having to provide the "s" parameter. Is there
> really no useful default?
>
> And now, if s=0, what does the filter do? Just sit around? Couldn't it
> fail instead?
>
> Apart from that:
> "if (" is the bracket style ffmpeg uses.
>
>> +if (dumpwave->json && !(dump_fp = av_fopen_utf8(dumpwave->json, 
>> "w")))
>> +av_log(ctx, AV_LOG_WARNING, "Flushing dump failed\n");
>
> I would appreciate evaluation of errno and printing the appropriate
> string (using av_strerror(), I believe).
>
>> +switch (inlink->format) {
>> +case AV_SAMPLE_FMT_DBLP:
>
> As Kyle hinted: Can this not force a conversion (implicit insertion of
> aformat filter) to e.g. double by only supporting this format, and
> reducing this switch to one or two cases? (The filter's output isn't
> really meant for transparent reuse anyway? af_volumedetect e.g. also
> supports only one, meaning its output can be a different format than
> its input.) It's also really hard to go through and later to maintain.
> It the big switch/case remains, one or two code macros would be
> appropriate.

I checked solution used in volumedetect and couldn't find a way to
read across formats.
How would you implement such macros? Since version 3.2 astats filter
uses same approach for reading different formats and as far as I know
macros harder to debug

>> +for (i = 0; i < buf->nb_samples; i++) {
>> +for (c = 0; c < channels; c++, src++)
>> +calc_db_rms(dumpwave, *src);
>> +}}
>
> I believe the curly brackets need to be placed differently.
>
>> +.description   = NULL_IF_CONFIG_SMALL("Dumps RMS amplitude to JSON 
>> file"),
>
> Imperative, i.e. "Dump RMS ...".
>
>> +FATE_AFILTER-$(call FILTERDEMDEC, DUMPWAVE, WAV, PCM_S16LE) += 
>> fate-filter-dumpwave
>> +fate-filter-dumpwave: SRC = $(TARGET_PATH)/tests/data/asynth-44100-2.wav
>> +fate-filter-dumpwave: CMD = ffmpeg -i $(SRC) -af 
>> dumpwave=d=1800x140:s=147:json=$(TARGET_PATH)/tests/data/fate/filter-dumpwave.out
>>  -f null - && cat $(TARGET_PATH)/tests/data/fate/filter-dumpwave.out
>> +
>
> If you need to implement all formats, you might want to test them all.
> Just a thought.
>
> Moritz
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

2018-01-09 Thread dmitry . gumenyuk

From: Dmytro Humeniuk 

Signed-off-by: Dmytro Humeniuk 
---
 Changelog  |   1 +
 doc/filters.texi   |  23 
 libavfilter/Makefile   |   1 +
 libavfilter/af_dumpwave.c  | 285 +
 libavfilter/allfilters.c   |   1 +
 libavfilter/version.h  |   4 +-
 tests/fate/filter-audio.mak|   5 +
 tests/ref/fate/filter-dumpwave |   1 +
 8 files changed, 319 insertions(+), 2 deletions(-)
 create mode 100644 libavfilter/af_dumpwave.c
 create mode 100644 tests/ref/fate/filter-dumpwave

diff --git a/Changelog b/Changelog
index 61075b3392..40fd624449 100644
--- a/Changelog
+++ b/Changelog
@@ -38,6 +38,7 @@ version :
 - Removed the ffserver program
 - Removed the ffmenc and ffmdec muxer and demuxer
 - VideoToolbox HEVC encoder and hwaccel
+- dumpwave audio filter
 
 
 version 3.4:
diff --git a/doc/filters.texi b/doc/filters.texi
index d29c40080f..98e54aec6e 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -2529,6 +2529,29 @@ Optional. It should have a value much less than 1 (e.g. 
0.05 or 0.02) and is
 used to prevent clipping.
 @end table
 
+@section dumpwave
+Dumps RMS amplitude to JSON file.
+Converts samples to decibels and calculates RMS (Root-Mean-Square) audio power 
scaled to desired values.
+
+@table @option
+@item d
+Dimensions @code{WxH}.
+@code{W} - number of data values in json, values will be scaled according to 
@code{H}.
+The default value is @var{640x480}
+
+@item s
+Samples count per value per channel
+
+@item json
+Path to json file
+@end table
+
+For example, to generate RMS amplitude for 44.1 kHz 6 seconds length audio
+with dimensions @var{1800x140}, samples count @code{44100*6/1800=147} and 
store it to @var{/tmp/out.json}, you might use:
+@example
+dumpwave=d=1800x140:s=147:json=/tmp/out.json
+@end example
+
 @section dynaudnorm
 Dynamic Audio Normalizer.
 
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index ef4729dd3f..2ffbc9497a 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -87,6 +87,7 @@ OBJS-$(CONFIG_COMPENSATIONDELAY_FILTER)  += 
af_compensationdelay.o
 OBJS-$(CONFIG_CROSSFEED_FILTER)  += af_crossfeed.o
 OBJS-$(CONFIG_CRYSTALIZER_FILTER)+= af_crystalizer.o
 OBJS-$(CONFIG_DCSHIFT_FILTER)+= af_dcshift.o
+OBJS-$(CONFIG_DUMPWAVE_FILTER)   += af_dumpwave.o
 OBJS-$(CONFIG_DYNAUDNORM_FILTER) += af_dynaudnorm.o
 OBJS-$(CONFIG_EARWAX_FILTER) += af_earwax.o
 OBJS-$(CONFIG_EBUR128_FILTER)+= f_ebur128.o
diff --git a/libavfilter/af_dumpwave.c b/libavfilter/af_dumpwave.c
new file mode 100644
index 00..35b970af6c
--- /dev/null
+++ b/libavfilter/af_dumpwave.c
@@ -0,0 +1,285 @@
+/*
+ * Copyright (c) 2017 Dmytro Humeniuk
+ *
+ * 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
+ * waveform audio filter – dumps RMS amplitude to JSON file
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/avstring.h"
+#include "libavutil/channel_layout.h"
+#include "libavutil/opt.h"
+#include "libavutil/parseutils.h"
+#include "avfilter.h"
+#include "formats.h"
+#include "audio.h"
+#include "internal.h"
+
+typedef struct DumpWaveContext {
+const AVClass *class;   /**< class for AVOptions */
+int w;  /**< number of data values in json */
+int h;  /**< values will be scaled according to provided */
+int is_disabled;/**< disable filter in case it's misconfigured */
+int i;  /**< index of value */
+char *json; /**< path to json */
+char *str;  /**< comma separated values */
+double *values; /**< scaling factors */
+int64_t s;  /**< samples per value per channel */
+int64_t n;  /**< current number of samples counted */
+int64_t max_samples;/**< samples per value */
+double sum; /**< sum of the squared samples per value */
+} DumpWaveContext;
+
+#define OFFSET(x) offsetof(DumpWaveContext, x)
+#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+
+static const AVOption dumpwave_options[] = {
+{ "d", "set width and

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

2018-01-09 Thread Dmitry Gumenyuk

Hi,
>Can you provide some context for this? Please add docs as well.
filter converts samples to decibels and calculates RMS
(Root-Mean-Square) audio power scaled to desired values,
it supposed to generate something like SoundCloud does
http://plnkr.co/edit/WeppGcEPQiOyKDsA0C4Y
I added docs.
>Can't really test this because it segfaults for me:  ./ffmpeg -f lavfi
>-i anoisesrc -af dumpwave -f null -
it expected some config - fixed

>Just set the sample format to something like AV_SAMPLE_FMT_DBL in
>query_format. There's no need for this big switch statement.
it handles samples in different way, you may look at the astats filter

2018-01-08 23:58 GMT+01:00 Kyle Swanson :
> Hi,
>
> Can you provide some context for this? Please add docs as well.
> Can't really test this because it segfaults for me:  ./ffmpeg -f lavfi
> -i anoisesrc -af dumpwave -f null -
>
> On Sun, Jan 7, 2018 at 4:36 PM,  wrote:
>>
>> From: Dmytro Humeniuk 
>>
>> Signed-off-by: Dmytro Humeniuk 
>> ---
>>  Changelog  |   1 +
>>  libavfilter/Makefile   |   1 +
>>  libavfilter/af_dumpwave.c  | 273 
>> +
>>  libavfilter/allfilters.c   |   1 +
>>  libavfilter/version.h  |   4 +-
>>  tests/fate/filter-audio.mak|   5 +
>>  tests/ref/fate/filter-dumpwave |   1 +
>>  7 files changed, 284 insertions(+), 2 deletions(-)
>>  create mode 100644 libavfilter/af_dumpwave.c
>>  create mode 100644 tests/ref/fate/filter-dumpwave
>>
>> diff --git a/Changelog b/Changelog
>> index 61075b3392..40fd624449 100644
>> --- a/Changelog
>> +++ b/Changelog
>> @@ -38,6 +38,7 @@ version :
>>  - Removed the ffserver program
>>  - Removed the ffmenc and ffmdec muxer and demuxer
>>  - VideoToolbox HEVC encoder and hwaccel
>> +- dumpwave audio filter
>>
>>
>>  version 3.4:
>> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
>> index 256dfabd66..020d90ee01 100644
>> --- a/libavfilter/Makefile
>> +++ b/libavfilter/Makefile
>> @@ -122,6 +122,7 @@ OBJS-$(CONFIG_TREMOLO_FILTER)+= 
>> af_tremolo.o
>>  OBJS-$(CONFIG_VIBRATO_FILTER)+= af_vibrato.o 
>> generate_wave_table.o
>>  OBJS-$(CONFIG_VOLUME_FILTER) += af_volume.o
>>  OBJS-$(CONFIG_VOLUMEDETECT_FILTER)   += af_volumedetect.o
>> +OBJS-$(CONFIG_DUMPWAVE_FILTER)   += af_dumpwave.o
>>
>>  OBJS-$(CONFIG_AEVALSRC_FILTER)   += aeval.o
>>  OBJS-$(CONFIG_ANOISESRC_FILTER)  += asrc_anoisesrc.o
>> diff --git a/libavfilter/af_dumpwave.c b/libavfilter/af_dumpwave.c
>> new file mode 100644
>> index 00..493b5b7ff2
>> --- /dev/null
>> +++ b/libavfilter/af_dumpwave.c
>> @@ -0,0 +1,273 @@
>> +/*
>> + * Copyright (c) 2017 Dmytro Humeniuk
>> + *
>> + * 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
>> + * waveform audio filter – dumps RMS amplitude to JSON file like SoundCloud 
>> does
>> + */
>> +
>> +#include "libavutil/avassert.h"
>> +#include "libavutil/avstring.h"
>> +#include "libavutil/channel_layout.h"
>> +#include "libavutil/opt.h"
>> +#include "libavutil/parseutils.h"
>> +#include "avfilter.h"
>> +#include "formats.h"
>> +#include "audio.h"
>> +#include "internal.h"
>> +
>> +typedef struct DumpWaveContext {
>> +const AVClass *class;
>> +int w, h;
>> +AVRational rate;
>> +int col;
>> +char *json;
>> +char *str;
>> +double *values;
>> +int64_t c, n, max_samples;
>> +double sum; /* sum of the squared samples per segment */
>> +} DumpWaveContext;
>> +
>> +#define OFFSET(x) offsetof(DumpWaveContext, x)
>> +#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
>> +
>> +static const AVOption dumpwave_options[] = {
>> +{ "s","set dump size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str = 
>> "600x240"}, 0, 0, FLAGS },
>> +{ "c", "set number of samples per item",  OFFSET(c), AV_OPT_TYPE_INT64, 
>>  {.i64 = 0}, 0, INT64_MAX, FLAGS },
>
> Use more descriptive parameter names. Also, per item? What is an item?
>
>> +{ "json", "set dump file", OFFSET(json), AV_OPT_TYPE_STRING, { .str = 
>> NULL }, 0, 0, FLAGS },
>> +{ NULL

[FFmpeg-devel] [PATCH] add dumpwave filter

2018-01-09 Thread dmitry . gumenyuk

From: Dmytro Humeniuk 

Signed-off-by: Dmytro Humeniuk 
---
 Changelog  |   1 +
 doc/filters.texi   |  23 
 libavfilter/Makefile   |   1 +
 libavfilter/af_dumpwave.c  | 285 +
 libavfilter/allfilters.c   |   1 +
 libavfilter/version.h  |   4 +-
 tests/fate/filter-audio.mak|   5 +
 tests/ref/fate/filter-dumpwave |   1 +
 8 files changed, 319 insertions(+), 2 deletions(-)
 create mode 100644 libavfilter/af_dumpwave.c
 create mode 100644 tests/ref/fate/filter-dumpwave

diff --git a/Changelog b/Changelog
index 61075b3392..40fd624449 100644
--- a/Changelog
+++ b/Changelog
@@ -38,6 +38,7 @@ version :
 - Removed the ffserver program
 - Removed the ffmenc and ffmdec muxer and demuxer
 - VideoToolbox HEVC encoder and hwaccel
+- dumpwave audio filter
 
 
 version 3.4:
diff --git a/doc/filters.texi b/doc/filters.texi
index d29c40080f..98e54aec6e 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -2529,6 +2529,29 @@ Optional. It should have a value much less than 1 (e.g. 
0.05 or 0.02) and is
 used to prevent clipping.
 @end table
 
+@section dumpwave
+Dumps RMS amplitude to JSON file.
+Converts samples to decibels and calculates RMS (Root-Mean-Square) audio power 
scaled to desired values.
+
+@table @option
+@item d
+Dimensions @code{WxH}.
+@code{W} - number of data values in json, values will be scaled according to 
@code{H}.
+The default value is @var{640x480}
+
+@item s
+Samples count per value per channel
+
+@item json
+Path to json file
+@end table
+
+For example, to generate RMS amplitude for 44.1 kHz 6 seconds length audio
+with dimensions @var{1800x140}, samples count @code{44100*6/1800=147} and 
store it to @var{/tmp/out.json}, you might use:
+@example
+dumpwave=d=1800x140:s=147:json=/tmp/out.json
+@end example
+
 @section dynaudnorm
 Dynamic Audio Normalizer.
 
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index ef4729dd3f..2ffbc9497a 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -87,6 +87,7 @@ OBJS-$(CONFIG_COMPENSATIONDELAY_FILTER)  += 
af_compensationdelay.o
 OBJS-$(CONFIG_CROSSFEED_FILTER)  += af_crossfeed.o
 OBJS-$(CONFIG_CRYSTALIZER_FILTER)+= af_crystalizer.o
 OBJS-$(CONFIG_DCSHIFT_FILTER)+= af_dcshift.o
+OBJS-$(CONFIG_DUMPWAVE_FILTER)   += af_dumpwave.o
 OBJS-$(CONFIG_DYNAUDNORM_FILTER) += af_dynaudnorm.o
 OBJS-$(CONFIG_EARWAX_FILTER) += af_earwax.o
 OBJS-$(CONFIG_EBUR128_FILTER)+= f_ebur128.o
diff --git a/libavfilter/af_dumpwave.c b/libavfilter/af_dumpwave.c
new file mode 100644
index 00..a1aa33d090
--- /dev/null
+++ b/libavfilter/af_dumpwave.c
@@ -0,0 +1,285 @@
+/*
+ * Copyright (c) 2017 Dmytro Humeniuk
+ *
+ * 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
+ * waveform audio filter – dumps RMS amplitude to JSON file like SoundCloud 
does
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/avstring.h"
+#include "libavutil/channel_layout.h"
+#include "libavutil/opt.h"
+#include "libavutil/parseutils.h"
+#include "avfilter.h"
+#include "formats.h"
+#include "audio.h"
+#include "internal.h"
+
+typedef struct DumpWaveContext {
+const AVClass *class;   /**< class for AVOptions */
+int w;  /**< number of data values in json */
+int h;  /**< values will be scaled according to provided */
+int is_disabled;/**< disable filter in case it's misconfigured */
+int i;  /**< index of value */
+char *json; /**< path to json */
+char *str;  /**< comma separated values */
+double *values; /**< scaling factors */
+int64_t s;  /**< samples per value per channel */
+int64_t n;  /**< current number of samples counted */
+int64_t max_samples;/**< samples per value */
+double sum; /**< sum of the squared samples per value */
+} DumpWaveContext;
+
+#define OFFSET(x) offsetof(DumpWaveContext, x)
+#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+
+static const AVOption dumpwave_options[] = {
+{ "d",

[FFmpeg-devel] dumpwave audio filter

2018-01-07 Thread dmitry . gumenyuk

Let me introduce dumpwave waveform audio filter – dumps RMS amplitude to JSON 
file just like SoundCloud does
As there is no duration knowledge, samples count should be precalculated
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

2018-01-07 Thread dmitry . gumenyuk

From: Dmytro Humeniuk 

Signed-off-by: Dmytro Humeniuk 
---
 Changelog  |   1 +
 libavfilter/Makefile   |   1 +
 libavfilter/af_dumpwave.c  | 273 +
 libavfilter/allfilters.c   |   1 +
 libavfilter/version.h  |   4 +-
 tests/fate/filter-audio.mak|   5 +
 tests/ref/fate/filter-dumpwave |   1 +
 7 files changed, 284 insertions(+), 2 deletions(-)
 create mode 100644 libavfilter/af_dumpwave.c
 create mode 100644 tests/ref/fate/filter-dumpwave

diff --git a/Changelog b/Changelog
index 61075b3392..40fd624449 100644
--- a/Changelog
+++ b/Changelog
@@ -38,6 +38,7 @@ version :
 - Removed the ffserver program
 - Removed the ffmenc and ffmdec muxer and demuxer
 - VideoToolbox HEVC encoder and hwaccel
+- dumpwave audio filter
 
 
 version 3.4:
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 256dfabd66..020d90ee01 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -122,6 +122,7 @@ OBJS-$(CONFIG_TREMOLO_FILTER)+= af_tremolo.o
 OBJS-$(CONFIG_VIBRATO_FILTER)+= af_vibrato.o 
generate_wave_table.o
 OBJS-$(CONFIG_VOLUME_FILTER) += af_volume.o
 OBJS-$(CONFIG_VOLUMEDETECT_FILTER)   += af_volumedetect.o
+OBJS-$(CONFIG_DUMPWAVE_FILTER)   += af_dumpwave.o
 
 OBJS-$(CONFIG_AEVALSRC_FILTER)   += aeval.o
 OBJS-$(CONFIG_ANOISESRC_FILTER)  += asrc_anoisesrc.o
diff --git a/libavfilter/af_dumpwave.c b/libavfilter/af_dumpwave.c
new file mode 100644
index 00..493b5b7ff2
--- /dev/null
+++ b/libavfilter/af_dumpwave.c
@@ -0,0 +1,273 @@
+/*
+ * Copyright (c) 2017 Dmytro Humeniuk
+ *
+ * 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
+ * waveform audio filter – dumps RMS amplitude to JSON file like SoundCloud 
does
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/avstring.h"
+#include "libavutil/channel_layout.h"
+#include "libavutil/opt.h"
+#include "libavutil/parseutils.h"
+#include "avfilter.h"
+#include "formats.h"
+#include "audio.h"
+#include "internal.h"
+
+typedef struct DumpWaveContext {
+const AVClass *class;
+int w, h;
+AVRational rate;
+int col;
+char *json;
+char *str;
+double *values;
+int64_t c, n, max_samples;
+double sum; /* sum of the squared samples per segment */
+} DumpWaveContext;
+
+#define OFFSET(x) offsetof(DumpWaveContext, x)
+#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+
+static const AVOption dumpwave_options[] = {
+{ "s","set dump size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str = 
"600x240"}, 0, 0, FLAGS },
+{ "c", "set number of samples per item",  OFFSET(c), AV_OPT_TYPE_INT64,  
{.i64 = 0}, 0, INT64_MAX, FLAGS },
+{ "json", "set dump file", OFFSET(json), AV_OPT_TYPE_STRING, { .str = NULL 
}, 0, 0, FLAGS },
+{ NULL }
+};
+
+AVFILTER_DEFINE_CLASS(dumpwave);
+
+static int config_output(AVFilterLink *outlink)
+{
+DumpWaveContext *dumpwave = outlink->src->priv;
+const int ch_width = dumpwave->w;
+dumpwave->values = av_malloc(ch_width * sizeof(double));
+dumpwave->str = av_malloc(ch_width*sizeof(int));
+dumpwave->max_samples = dumpwave->c * outlink->channels;
+
+return 0;
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+DumpWaveContext *dumpwave = ctx->priv;
+const int ch_height = dumpwave->h;
+const int ch_width = dumpwave->w;
+char *result = dumpwave->str;
+FILE *dump_fp = NULL;
+
+if (dumpwave->json && !(dump_fp = av_fopen_utf8(dumpwave->json, "w")))
+av_log(ctx, AV_LOG_WARNING, "dump failed.\n");
+
+if (dump_fp) {
+fprintf(dump_fp, "{\"width\":%d,\"height\":%d,\"samples\":[%s]}", 
ch_width, ch_height, result);
+fclose(dump_fp);
+}
+av_freep(>str);
+av_freep(>values);
+}
+
+static int dumpwave_request_frame(AVFilterLink *outlink)
+{
+AVFilterContext *ctx = outlink->src;
+DumpWaveContext *dumpwave = ctx->priv;
+const int ch_height = dumpwave->h;
+const int ch_width = dumpwave->w;
+const double *values = dumpwave->values;
+char *result = dumpwave->str;
+
+AVFilterLink *inlink = ctx->inputs[0];
+int

Re: [FFmpeg-devel] [PATCH v2] avfilter: add dumpwave filter.

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

[FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

Re: [FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

[FFmpeg-devel] [PATCH] add dumpwave filter

[FFmpeg-devel] dumpwave audio filter

[FFmpeg-devel] [PATCH] avfilter: add dumpwave filter.

10 matches

Site Navigation

Mail list logo

Footer information