[FFmpeg-devel] [PATCH] avformat/demux: Add allow_codec_changes option to AVFormatContext (v2)

[FFmpeg-devel] [PATCH] avformat/demux: Add allow_codec_changes option to AVFormatContext (v2)

[Pavel Koshevoy]

Make runtime AVStream.codecpar codec_id updates optional and disabled
by default, so that avformat API clients can enable this feature explicitly
when they add support for runtime codec changes.

Accordingly, codec_close should not assume that the codec_id cannot change.
This fixes 'ffprobe 1_poc.mp4' segfault introduced with
commit 0021484d05f9b0f032fa319399de6e24eea0c04f.
---
 doc/APIchanges              |  3 +++
 doc/formats.texi            |  6 ++++++
 libavformat/avformat.h      | 10 ++++++++++
 libavformat/demux.c         |  8 +++++++-
 libavformat/mpegts.c        |  4 +++-
 libavformat/options_table.h |  1 +
 libavformat/version.h       |  2 +-
 tests/fate/demux.mak        |  2 +-
 8 files changed, 32 insertions(+), 4 deletions(-)

diff --git a/doc/APIchanges b/doc/APIchanges
index 91710bb27d..29b3389094 100644
--- a/doc/APIchanges
+++ b/doc/APIchanges
@@ -2,6 +2,9 @@ The last version increases of all libraries were on 2025-03-28
 
 API changes, most recent first:
 
+2025-06-17 - xxxxxxxxxx - lavf 62.2.100 - avformat.h
+  Add AVFormatContext.allow_codec_changes AVOption.
+
 2025-05-21 - xxxxxxxxxx - lavu 60.3.100 - avassert.h
   Add av_unreachable() and av_assume() macros.
 
diff --git a/doc/formats.texi b/doc/formats.texi
index 876a9e92b3..d18613f8bd 100644
--- a/doc/formats.texi
+++ b/doc/formats.texi
@@ -228,6 +228,12 @@ would require too many resources due to a large number of streams.
 Skip estimation of input duration if it requires an additional probing for PTS at end of file.
 At present, applicable for MPEG-PS and MPEG-TS.
 
+@item allow_codec_changes @var{bool} (@emph{input})
+Allow AVStream.codecpar codec_type and codec_id to change at runtime
+(for example when MPEG-TS PMT ES stream_type changes at runtime).
+This is disabled by default, because most clients of avformat API
+do not support random mid-stream codec changes.
+
 @item duration_probesize @var{integer} (@emph{input})
 Set probing size, in bytes, for input duration estimation when it actually requires
 an additional probing for PTS at end of file (at present: MPEG-PS and MPEG-TS).
diff --git a/libavformat/avformat.h b/libavformat/avformat.h
index b6c63e2237..c6956626b5 100644
--- a/libavformat/avformat.h
+++ b/libavformat/avformat.h
@@ -1884,6 +1884,16 @@ typedef struct AVFormatContext {
      * @see skip_estimate_duration_from_pts
      */
     int64_t duration_probesize;
+
+    /**
+     * Allow AVStream.codecpar codec_type and codec_id to change at runtime
+     * (for example when MPEG-TS PMT ES stream_type changes at runtime).
+     * This is disabled by default, because most clients of avformat API
+     * do not support random mid-stream codec changes.
+     * - muxing: unused
+     * - demuxing: set by user
+     */
+    int allow_codec_changes;
 } AVFormatContext;
 
 /**
diff --git a/libavformat/demux.c b/libavformat/demux.c
index ecd4f40da9..3749ab67a3 100644
--- a/libavformat/demux.c
+++ b/libavformat/demux.c
@@ -1292,9 +1292,15 @@ static int codec_close(FFStream *sti)
 {
     AVCodecContext *avctx_new = NULL;
     AVCodecParameters *par_tmp = NULL;
+    const AVCodec *new_codec = NULL;
     int ret;
 
-    avctx_new = avcodec_alloc_context3(sti->avctx->codec);
+    new_codec =
+      (sti->avctx->codec_id != sti->pub.codecpar->codec_id) ?
+      avcodec_find_decoder(sti->pub.codecpar->codec_id) :
+      sti->avctx->codec;
+
+    avctx_new = avcodec_alloc_context3(new_codec);
     if (!avctx_new) {
         ret = AVERROR(ENOMEM);
         goto fail;
diff --git a/libavformat/mpegts.c b/libavformat/mpegts.c
index deb69a0548..0c78e870bb 100644
--- a/libavformat/mpegts.c
+++ b/libavformat/mpegts.c
@@ -2510,7 +2510,9 @@ static void pmt_cb(MpegTSFilter *filter, const uint8_t *section, int section_len
         if (!st)
             goto out;
 
-        if (pes && pes->stream_type != stream_type)
+        if (pes && (!pes->stream_type ||
+                    (pes->stream_type != stream_type &&
+                     ts->stream->allow_codec_changes)))
             mpegts_set_stream_info(st, pes, stream_type, prog_reg_desc);
 
         add_pid_to_program(prg, pid);
diff --git a/libavformat/options_table.h b/libavformat/options_table.h
index e2e690fd2a..e949d60c3e 100644
--- a/libavformat/options_table.h
+++ b/libavformat/options_table.h
@@ -106,6 +106,7 @@ static const AVOption avformat_options[] = {
 {"skip_estimate_duration_from_pts", "skip duration calculation in estimate_timings_from_pts", OFFSET(skip_estimate_duration_from_pts), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1, D},
 {"max_probe_packets", "Maximum number of packets to probe a codec", OFFSET(max_probe_packets), AV_OPT_TYPE_INT, { .i64 = 2500 }, 0, INT_MAX, D },
 {"duration_probesize", "Maximum number of bytes to probe the durations of the streams in estimate_timings_from_pts", OFFSET(duration_probesize), AV_OPT_TYPE_INT64, {.i64 = 0 }, 0, INT64_MAX, D},
+{"allow_codec_changes", "Allow AVStream.codecpar codec_type and codec_id to change at runtime", OFFSET(allow_codec_changes), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1, D},
 {NULL},
 };
 
diff --git a/libavformat/version.h b/libavformat/version.h
index a7c80dc564..904e7f06aa 100644
--- a/libavformat/version.h
+++ b/libavformat/version.h
@@ -31,7 +31,7 @@
 
 #include "version_major.h"
 
-#define LIBAVFORMAT_VERSION_MINOR   1
+#define LIBAVFORMAT_VERSION_MINOR   2
 #define LIBAVFORMAT_VERSION_MICRO 100
 
 #define LIBAVFORMAT_VERSION_INT AV_VERSION_INT(LIBAVFORMAT_VERSION_MAJOR, \
diff --git a/tests/fate/demux.mak b/tests/fate/demux.mak
index ead5ad4b10..4a39e950b9 100644
--- a/tests/fate/demux.mak
+++ b/tests/fate/demux.mak
@@ -158,7 +158,7 @@ FATE_SAMPLES_DEMUX-$(CONFIG_XWMA_DEMUXER) += fate-xwma-demux
 fate-xwma-demux: CMD = crc -i $(TARGET_SAMPLES)/xwma/ergon.xwma -c:a copy
 
 FATE_FFPROBE_DEMUX-$(CONFIG_MPEGTS_DEMUXER) += fate-ts-demux
-fate-ts-demux: CMD = ffprobe_demux $(TARGET_SAMPLES)/ac3/mp3ac325-4864-small.ts
+fate-ts-demux: CMD = ffprobe_demux $(TARGET_SAMPLES)/ac3/mp3ac325-4864-small.ts -allow_codec_changes 1
 
 FATE_FFPROBE_DEMUX-$(CONFIG_MPEGTS_DEMUXER) += fate-ts-timed-id3-demux
 fate-ts-timed-id3-demux: CMD = ffprobe_demux $(TARGET_SAMPLES)/mpegts/id3.ts
-- 
2.43.0

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH] avformat/demux: Add allow_codec_changes option to AVFormatContext (v2)

[Marton Balint]

On Tue, 17 Jun 2025, Pavel Koshevoy wrote:

> Make runtime AVStream.codecpar codec_id updates optional and disabled
> by default, so that avformat API clients can enable this feature explicitly
> when they add support for runtime codec changes.
>
> Accordingly, codec_close should not assume that the codec_id cannot change.
> This fixes 'ffprobe 1_poc.mp4' segfault introduced with
> commit 0021484d05f9b0f032fa319399de6e24eea0c04f.
> ---
> doc/APIchanges              |  3 +++
> doc/formats.texi            |  6 ++++++
> libavformat/avformat.h      | 10 ++++++++++
> libavformat/demux.c         |  8 +++++++-
> libavformat/mpegts.c        |  4 +++-
> libavformat/options_table.h |  1 +
> libavformat/version.h       |  2 +-
> tests/fate/demux.mak        |  2 +-
> 8 files changed, 32 insertions(+), 4 deletions(-)
>
> diff --git a/doc/APIchanges b/doc/APIchanges
> index 91710bb27d..29b3389094 100644
> --- a/doc/APIchanges
> +++ b/doc/APIchanges
> @@ -2,6 +2,9 @@ The last version increases of all libraries were on 2025-03-28
>
> API changes, most recent first:
>
> +2025-06-17 - xxxxxxxxxx - lavf 62.2.100 - avformat.h
> +  Add AVFormatContext.allow_codec_changes AVOption.

Considering that this is a simple boolean option, maybe you should use 
AVFormatContext->flags instead, so no new field need to be added, only a 
new flag.

Thanks,
Marton

> +
> 2025-05-21 - xxxxxxxxxx - lavu 60.3.100 - avassert.h
>   Add av_unreachable() and av_assume() macros.
>
> diff --git a/doc/formats.texi b/doc/formats.texi
> index 876a9e92b3..d18613f8bd 100644
> --- a/doc/formats.texi
> +++ b/doc/formats.texi
> @@ -228,6 +228,12 @@ would require too many resources due to a large number of streams.
> Skip estimation of input duration if it requires an additional probing for PTS at end of file.
> At present, applicable for MPEG-PS and MPEG-TS.
>
> +@item allow_codec_changes @var{bool} (@emph{input})
> +Allow AVStream.codecpar codec_type and codec_id to change at runtime
> +(for example when MPEG-TS PMT ES stream_type changes at runtime).
> +This is disabled by default, because most clients of avformat API
> +do not support random mid-stream codec changes.
> +
> @item duration_probesize @var{integer} (@emph{input})
> Set probing size, in bytes, for input duration estimation when it actually requires
> an additional probing for PTS at end of file (at present: MPEG-PS and MPEG-TS).
> diff --git a/libavformat/avformat.h b/libavformat/avformat.h
> index b6c63e2237..c6956626b5 100644
> --- a/libavformat/avformat.h
> +++ b/libavformat/avformat.h
> @@ -1884,6 +1884,16 @@ typedef struct AVFormatContext {
>      * @see skip_estimate_duration_from_pts
>      */
>     int64_t duration_probesize;
> +
> +    /**
> +     * Allow AVStream.codecpar codec_type and codec_id to change at runtime
> +     * (for example when MPEG-TS PMT ES stream_type changes at runtime).
> +     * This is disabled by default, because most clients of avformat API
> +     * do not support random mid-stream codec changes.
> +     * - muxing: unused
> +     * - demuxing: set by user
> +     */
> +    int allow_codec_changes;
> } AVFormatContext;
>
> /**
> diff --git a/libavformat/demux.c b/libavformat/demux.c
> index ecd4f40da9..3749ab67a3 100644
> --- a/libavformat/demux.c
> +++ b/libavformat/demux.c
> @@ -1292,9 +1292,15 @@ static int codec_close(FFStream *sti)
> {
>     AVCodecContext *avctx_new = NULL;
>     AVCodecParameters *par_tmp = NULL;
> +    const AVCodec *new_codec = NULL;
>     int ret;
>
> -    avctx_new = avcodec_alloc_context3(sti->avctx->codec);
> +    new_codec =
> +      (sti->avctx->codec_id != sti->pub.codecpar->codec_id) ?
> +      avcodec_find_decoder(sti->pub.codecpar->codec_id) :
> +      sti->avctx->codec;
> +
> +    avctx_new = avcodec_alloc_context3(new_codec);
>     if (!avctx_new) {
>         ret = AVERROR(ENOMEM);
>         goto fail;
> diff --git a/libavformat/mpegts.c b/libavformat/mpegts.c
> index deb69a0548..0c78e870bb 100644
> --- a/libavformat/mpegts.c
> +++ b/libavformat/mpegts.c
> @@ -2510,7 +2510,9 @@ static void pmt_cb(MpegTSFilter *filter, const uint8_t *section, int section_len
>         if (!st)
>             goto out;
>
> -        if (pes && pes->stream_type != stream_type)
> +        if (pes && (!pes->stream_type ||
> +                    (pes->stream_type != stream_type &&
> +                     ts->stream->allow_codec_changes)))
>             mpegts_set_stream_info(st, pes, stream_type, prog_reg_desc);
>
>         add_pid_to_program(prg, pid);
> diff --git a/libavformat/options_table.h b/libavformat/options_table.h
> index e2e690fd2a..e949d60c3e 100644
> --- a/libavformat/options_table.h
> +++ b/libavformat/options_table.h
> @@ -106,6 +106,7 @@ static const AVOption avformat_options[] = {
> {"skip_estimate_duration_from_pts", "skip duration calculation in estimate_timings_from_pts", OFFSET(skip_estimate_duration_from_pts), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1, D},
> {"max_probe_packets", "Maximum number of packets to probe a codec", OFFSET(max_probe_packets), AV_OPT_TYPE_INT, { .i64 = 2500 }, 0, INT_MAX, D },
> {"duration_probesize", "Maximum number of bytes to probe the durations of the streams in estimate_timings_from_pts", OFFSET(duration_probesize), AV_OPT_TYPE_INT64, {.i64 = 0 }, 0, INT64_MAX, D},
> +{"allow_codec_changes", "Allow AVStream.codecpar codec_type and codec_id to change at runtime", OFFSET(allow_codec_changes), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1, D},
> {NULL},
> };
>
> diff --git a/libavformat/version.h b/libavformat/version.h
> index a7c80dc564..904e7f06aa 100644
> --- a/libavformat/version.h
> +++ b/libavformat/version.h
> @@ -31,7 +31,7 @@
>
> #include "version_major.h"
>
> -#define LIBAVFORMAT_VERSION_MINOR   1
> +#define LIBAVFORMAT_VERSION_MINOR   2
> #define LIBAVFORMAT_VERSION_MICRO 100
>
> #define LIBAVFORMAT_VERSION_INT AV_VERSION_INT(LIBAVFORMAT_VERSION_MAJOR, \
> diff --git a/tests/fate/demux.mak b/tests/fate/demux.mak
> index ead5ad4b10..4a39e950b9 100644
> --- a/tests/fate/demux.mak
> +++ b/tests/fate/demux.mak
> @@ -158,7 +158,7 @@ FATE_SAMPLES_DEMUX-$(CONFIG_XWMA_DEMUXER) += fate-xwma-demux
> fate-xwma-demux: CMD = crc -i $(TARGET_SAMPLES)/xwma/ergon.xwma -c:a copy
>
> FATE_FFPROBE_DEMUX-$(CONFIG_MPEGTS_DEMUXER) += fate-ts-demux
> -fate-ts-demux: CMD = ffprobe_demux $(TARGET_SAMPLES)/ac3/mp3ac325-4864-small.ts
> +fate-ts-demux: CMD = ffprobe_demux $(TARGET_SAMPLES)/ac3/mp3ac325-4864-small.ts -allow_codec_changes 1
>
> FATE_FFPROBE_DEMUX-$(CONFIG_MPEGTS_DEMUXER) += fate-ts-timed-id3-demux
> fate-ts-timed-id3-demux: CMD = ffprobe_demux $(TARGET_SAMPLES)/mpegts/id3.ts
> -- 
> 2.43.0
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH] avformat/demux: Add allow_codec_changes option to AVFormatContext (v2)

[Pavel Koshevoy]

On Tue, Jun 17, 2025 at 12:42 PM Marton Balint <cus@passwd.hu> wrote:

>
>
> On Tue, 17 Jun 2025, Pavel Koshevoy wrote:
>
> > Make runtime AVStream.codecpar codec_id updates optional and disabled
> > by default, so that avformat API clients can enable this feature
> explicitly
> > when they add support for runtime codec changes.
> >
> > Accordingly, codec_close should not assume that the codec_id cannot
> change.
> > This fixes 'ffprobe 1_poc.mp4' segfault introduced with
> > commit 0021484d05f9b0f032fa319399de6e24eea0c04f.
> > ---
> > doc/APIchanges              |  3 +++
> > doc/formats.texi            |  6 ++++++
> > libavformat/avformat.h      | 10 ++++++++++
> > libavformat/demux.c         |  8 +++++++-
> > libavformat/mpegts.c        |  4 +++-
> > libavformat/options_table.h |  1 +
> > libavformat/version.h       |  2 +-
> > tests/fate/demux.mak        |  2 +-
> > 8 files changed, 32 insertions(+), 4 deletions(-)
> >
> > diff --git a/doc/APIchanges b/doc/APIchanges
> > index 91710bb27d..29b3389094 100644
> > --- a/doc/APIchanges
> > +++ b/doc/APIchanges
> > @@ -2,6 +2,9 @@ The last version increases of all libraries were on
> 2025-03-28
> >
> > API changes, most recent first:
> >
> > +2025-06-17 - xxxxxxxxxx - lavf 62.2.100 - avformat.h
> > +  Add AVFormatContext.allow_codec_changes AVOption.
>
> Considering that this is a simple boolean option, maybe you should use
> AVFormatContext->flags instead, so no new field need to be added, only a
> new flag.
>

Are you referring to AVFMT_FLAG_...  flags?
Are they controllable from the command line, as an AVOption?
I referenced AVFormatContext.skip_estimate_duration_from_pts as an example,
and that's not a flag.

I'd like a second opinion on this.

Thank you,
    Pavel.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH] avformat/demux: Add allow_codec_changes option to AVFormatContext (v2)

[Michael Niedermayer]

[-- Attachment #1.1: Type: text/plain, Size: 2105 bytes --]

On Tue, Jun 17, 2025 at 08:39:05PM +0200, Marton Balint wrote:
> 
> 
> On Tue, 17 Jun 2025, Pavel Koshevoy wrote:
> 
> > Make runtime AVStream.codecpar codec_id updates optional and disabled
> > by default, so that avformat API clients can enable this feature explicitly
> > when they add support for runtime codec changes.
> > 
> > Accordingly, codec_close should not assume that the codec_id cannot change.
> > This fixes 'ffprobe 1_poc.mp4' segfault introduced with
> > commit 0021484d05f9b0f032fa319399de6e24eea0c04f.
> > ---
> > doc/APIchanges              |  3 +++
> > doc/formats.texi            |  6 ++++++
> > libavformat/avformat.h      | 10 ++++++++++
> > libavformat/demux.c         |  8 +++++++-
> > libavformat/mpegts.c        |  4 +++-
> > libavformat/options_table.h |  1 +
> > libavformat/version.h       |  2 +-
> > tests/fate/demux.mak        |  2 +-
> > 8 files changed, 32 insertions(+), 4 deletions(-)
> > 
> > diff --git a/doc/APIchanges b/doc/APIchanges
> > index 91710bb27d..29b3389094 100644
> > --- a/doc/APIchanges
> > +++ b/doc/APIchanges
> > @@ -2,6 +2,9 @@ The last version increases of all libraries were on 2025-03-28
> > 
> > API changes, most recent first:
> > 
> > +2025-06-17 - xxxxxxxxxx - lavf 62.2.100 - avformat.h
> > +  Add AVFormatContext.allow_codec_changes AVOption.
> 
> Considering that this is a simple boolean option, maybe you should use
> AVFormatContext->flags instead, so no new field need to be added, only a new
> flag.

as a seperate value it could be extended to somethig like

-allow_codec_changes ALL
-allow_codec_changes NONE
-allow_codec_changes WITHIN_TYPE        (no type change like video to audio)

but not sure this makes sense or is even reasonable implementation wise

anyway i have nothing against making this a flag if people prefer
just wanted to provide an example where a flag would be at a disadvantage

thx

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

The worst form of inequality is to try to make unequal things equal.
-- Aristotle

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

[-- Attachment #2: Type: text/plain, Size: 251 bytes --]

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".