Re: [FFmpeg-devel] [PATCH v2] ffprobe: add -codec:<media_spec> option
On date Friday 2025-06-27 10:23:56 +0200, ffmpeg-devel Mailing List wrote:
From: Maryla Ustarroz-Calonge <maryla@google.com> Subject: [PATCH v2] ffprobe: add -codec:<media_spec> option Date: Fri, 27 Jun 2025 10:23:56 +0200 To: ffmpeg-devel@ffmpeg.org X-Mailer: git-send-email 2.50.0.727.gbf7dc18ff4-goog
And -c:<media_spec> variant. Update ffprobe.texi. opt_codec() is mostly copied over from ffplay.c
Signed-off-by: Maryla Ustarroz-Calonge <maryla@google.com> --- Changelog | 2 +- doc/ffprobe.texi | 8 ++++ fftools/ffprobe.c | 117 +++++++++++++++++++++++++++++++++++++++++----- 3 files changed, 114 insertions(+), 13 deletions(-)
diff --git a/Changelog b/Changelog index 4217449438..ae73611222 100644 --- a/Changelog +++ b/Changelog @@ -18,7 +18,7 @@ version <next>: - APV encoding support through a libopenapv wrapper - VVC decoder supports all content of SCC (Screen Content Coding): IBC (Inter Block Copy), Palette Mode and ACT (Adaptive Color Transform - +- ffprobe -codec option
version 7.1: - Raw Captions with Time (RCWT) closed caption demuxer diff --git a/doc/ffprobe.texi b/doc/ffprobe.texi index 8834df8d35..4ce0c8b583 100644 --- a/doc/ffprobe.texi +++ b/doc/ffprobe.texi @@ -368,6 +368,14 @@ Read @var{input_url}. Write output to @var{output_url}. If not specified, the output is sent to stdout.
+@item -c:@var{media_specifier} @var{codec_name} +Force a specific decoder implementation for the stream identified by +@var{media_specifier}, which can assume the values @code{a} (audio), +@code{v} (video), @code{s} (subtitle), and @code{d} (data). + +@item -codec:@var{media_specifier} @var{codec_name} +Alias for -c:@var{media_specifier}.
Nit: I'd put the long version before the short one, but keep as is if this is consistent with the other docs.
+ @end table @c man end
diff --git a/fftools/ffprobe.c b/fftools/ffprobe.c index 1346ed33c5..14ba4d3fb0 100644 --- a/fftools/ffprobe.c +++ b/fftools/ffprobe.c @@ -36,6 +36,7 @@ #include "libavutil/ambient_viewing_environment.h" #include "libavutil/avassert.h" #include "libavutil/avstring.h" +#include "libavutil/avutil.h" #include "libavutil/bprint.h" #include "libavutil/channel_layout.h" #include "libavutil/display.h" @@ -130,6 +131,11 @@ static int use_byte_value_binary_prefix = 0; static int use_value_sexagesimal_format = 0; static int show_private_data = 1;
+static const char *audio_codec_name = NULL; +static const char *data_codec_name = NULL; +static const char *subtitle_codec_name = NULL; +static const char *video_codec_name = NULL; + #define SHOW_OPTIONAL_FIELDS_AUTO -1 #define SHOW_OPTIONAL_FIELDS_NEVER 0 #define SHOW_OPTIONAL_FIELDS_ALWAYS 1 @@ -2284,6 +2290,64 @@ static void show_error(AVTextFormatContext *tfc, int err) avtext_print_section_footer(tfc); }
+static int get_decoder_by_name(const char *codec_name, const AVCodec **codec) +{ + if (codec_name == NULL) + return 0; + + *codec = avcodec_find_decoder_by_name(codec_name); + if (*codec == NULL) { + av_log(NULL, AV_LOG_ERROR, + "No codec could be found with name '%s'\n", codec_name); + return AVERROR(EINVAL); + } + return 0; +} + +static int set_decoders(AVFormatContext *fmt_ctx) +{ + int ret;
+ ret = get_decoder_by_name(audio_codec_name, &fmt_ctx->audio_codec); + if (ret < 0) return ret;
nit++: could be turned to a macro to capture the pattern: #define GET_DECODER(type_) \ ret = get_decoder_by_name(type_##_codec_name, &fmt_ctx->type_##_codec);\ if (ret < 0) return ret; GET_DECODER(audio); GET_DECODER(data); ...
+ ret = get_decoder_by_name(data_codec_name, &fmt_ctx->data_codec); + if (ret < 0) return ret; + ret = get_decoder_by_name(subtitle_codec_name, &fmt_ctx->subtitle_codec); + if (ret < 0) return ret; + ret = get_decoder_by_name(video_codec_name, &fmt_ctx->video_codec); + if (ret < 0) return ret; + return 0; +} +
Looks good apart from the minor nits, you can update the patch with the extended commit description otherwise I'd push this in a few days if there are no more comments. Thanks.
On Thu, Jul 31, 2025 at 12:01 AM Stefano Sabatini <stefasab@gmail.com> wrote:
On date Friday 2025-06-27 10:23:56 +0200, ffmpeg-devel Mailing List wrote:
From: Maryla Ustarroz-Calonge <maryla@google.com> Subject: [PATCH v2] ffprobe: add -codec:<media_spec> option Date: Fri, 27 Jun 2025 10:23:56 +0200 To: ffmpeg-devel@ffmpeg.org X-Mailer: git-send-email 2.50.0.727.gbf7dc18ff4-goog
And -c:<media_spec> variant. Update ffprobe.texi. opt_codec() is mostly copied over from ffplay.c
Signed-off-by: Maryla Ustarroz-Calonge <maryla@google.com> --- Changelog | 2 +- doc/ffprobe.texi | 8 ++++ fftools/ffprobe.c | 117 +++++++++++++++++++++++++++++++++++++++++----- 3 files changed, 114 insertions(+), 13 deletions(-)
diff --git a/Changelog b/Changelog index 4217449438..ae73611222 100644 --- a/Changelog +++ b/Changelog @@ -18,7 +18,7 @@ version <next>: - APV encoding support through a libopenapv wrapper - VVC decoder supports all content of SCC (Screen Content Coding): IBC (Inter Block Copy), Palette Mode and ACT (Adaptive Color Transform - +- ffprobe -codec option
version 7.1: - Raw Captions with Time (RCWT) closed caption demuxer diff --git a/doc/ffprobe.texi b/doc/ffprobe.texi index 8834df8d35..4ce0c8b583 100644 --- a/doc/ffprobe.texi +++ b/doc/ffprobe.texi @@ -368,6 +368,14 @@ Read @var{input_url}. Write output to @var{output_url}. If not specified, the output is sent to stdout.
+@item -c:@var{media_specifier} @var{codec_name} +Force a specific decoder implementation for the stream identified by +@var{media_specifier}, which can assume the values @code{a} (audio), +@code{v} (video), @code{s} (subtitle), and @code{d} (data). + +@item -codec:@var{media_specifier} @var{codec_name} +Alias for -c:@var{media_specifier}.
Nit: I'd put the long version before the short one, but keep as is if this is consistent with the other docs.
I just sent a v3 addressing your comments. In the texi file, I merged the two entries for -c and -codec with -c first, as is done in ffmpeg.texi Note that in the flag descriptions in ffprobe.c, -codec is marked as being an alias of -c, which seems backwards to me, but I did it this way to be consistent with ffmpeg.c
+ @end table @c man end
diff --git a/fftools/ffprobe.c b/fftools/ffprobe.c index 1346ed33c5..14ba4d3fb0 100644 --- a/fftools/ffprobe.c +++ b/fftools/ffprobe.c @@ -36,6 +36,7 @@ #include "libavutil/ambient_viewing_environment.h" #include "libavutil/avassert.h" #include "libavutil/avstring.h" +#include "libavutil/avutil.h" #include "libavutil/bprint.h" #include "libavutil/channel_layout.h" #include "libavutil/display.h" @@ -130,6 +131,11 @@ static int use_byte_value_binary_prefix = 0; static int use_value_sexagesimal_format = 0; static int show_private_data = 1;
+static const char *audio_codec_name = NULL; +static const char *data_codec_name = NULL; +static const char *subtitle_codec_name = NULL; +static const char *video_codec_name = NULL; + #define SHOW_OPTIONAL_FIELDS_AUTO -1 #define SHOW_OPTIONAL_FIELDS_NEVER 0 #define SHOW_OPTIONAL_FIELDS_ALWAYS 1 @@ -2284,6 +2290,64 @@ static void show_error(AVTextFormatContext *tfc, int err) avtext_print_section_footer(tfc); }
+static int get_decoder_by_name(const char *codec_name, const AVCodec **codec) +{ + if (codec_name == NULL) + return 0; + + *codec = avcodec_find_decoder_by_name(codec_name); + if (*codec == NULL) { + av_log(NULL, AV_LOG_ERROR, + "No codec could be found with name '%s'\n", codec_name); + return AVERROR(EINVAL); + } + return 0; +} + +static int set_decoders(AVFormatContext *fmt_ctx) +{ + int ret;
+ ret = get_decoder_by_name(audio_codec_name, &fmt_ctx->audio_codec); + if (ret < 0) return ret;
nit++: could be turned to a macro to capture the pattern:
#define GET_DECODER(type_) \ ret = get_decoder_by_name(type_##_codec_name, &fmt_ctx->type_##_codec);\ if (ret < 0) return ret;
GET_DECODER(audio); GET_DECODER(data); ...
+ ret = get_decoder_by_name(data_codec_name, &fmt_ctx->data_codec); + if (ret < 0) return ret; + ret = get_decoder_by_name(subtitle_codec_name, &fmt_ctx->subtitle_codec); + if (ret < 0) return ret; + ret = get_decoder_by_name(video_codec_name, &fmt_ctx->video_codec); + if (ret < 0) return ret; + return 0; +} +
Looks good apart from the minor nits, you can update the patch with the extended commit description otherwise I'd push this in a few days if there are no more comments.
Thanks.
On date Thursday 2025-07-31 13:54:52 +0200, ffmpeg-devel Mailing List wrote:
On Thu, Jul 31, 2025 at 12:01 AM Stefano Sabatini <stefasab@gmail.com> wrote:
On date Friday 2025-06-27 10:23:56 +0200, ffmpeg-devel Mailing List wrote:
From: Maryla Ustarroz-Calonge <maryla@google.com> Subject: [PATCH v2] ffprobe: add -codec:<media_spec> option Date: Fri, 27 Jun 2025 10:23:56 +0200 To: ffmpeg-devel@ffmpeg.org X-Mailer: git-send-email 2.50.0.727.gbf7dc18ff4-goog
And -c:<media_spec> variant. Update ffprobe.texi. opt_codec() is mostly copied over from ffplay.c
Signed-off-by: Maryla Ustarroz-Calonge <maryla@google.com> --- Changelog | 2 +- doc/ffprobe.texi | 8 ++++ fftools/ffprobe.c | 117 +++++++++++++++++++++++++++++++++++++++++----- 3 files changed, 114 insertions(+), 13 deletions(-)
diff --git a/Changelog b/Changelog index 4217449438..ae73611222 100644 --- a/Changelog +++ b/Changelog @@ -18,7 +18,7 @@ version <next>: - APV encoding support through a libopenapv wrapper - VVC decoder supports all content of SCC (Screen Content Coding): IBC (Inter Block Copy), Palette Mode and ACT (Adaptive Color Transform - +- ffprobe -codec option
version 7.1: - Raw Captions with Time (RCWT) closed caption demuxer diff --git a/doc/ffprobe.texi b/doc/ffprobe.texi index 8834df8d35..4ce0c8b583 100644 --- a/doc/ffprobe.texi +++ b/doc/ffprobe.texi @@ -368,6 +368,14 @@ Read @var{input_url}. Write output to @var{output_url}. If not specified, the output is sent to stdout.
+@item -c:@var{media_specifier} @var{codec_name} +Force a specific decoder implementation for the stream identified by +@var{media_specifier}, which can assume the values @code{a} (audio), +@code{v} (video), @code{s} (subtitle), and @code{d} (data). + +@item -codec:@var{media_specifier} @var{codec_name} +Alias for -c:@var{media_specifier}.
Nit: I'd put the long version before the short one, but keep as is if this is consistent with the other docs.
I just sent a v3 addressing your comments. In the texi file, I merged the two entries for -c and -codec with -c first, as is done in ffmpeg.texi Note that in the flag descriptions in ffprobe.c, -codec is marked as being an alias of -c, which seems backwards to me, but I did it this way to be consistent with ffmpeg.c
I just pushed version 3, thanks!
participants (2)
-
Maryla Ustarroz -
Stefano Sabatini