[FFmpeg-devel] [PATCH] fftools/ffmpeg_demux: gracefully ignore mismatching channel layouts for -channel_layout option

Marton Balint cus at passwd.hu
Tue Jun 4 00:48:47 EEST 2024


The very old behaviour of -channel_layout was to simply warn the user about
channel layouts which does not have a matching channel count, and ignore them,
instead of reporting an error.

The recent fix re-added support for overriding -channel_layout, but it rejected
mismatching layouts. There is no easy way for the user to specify a channel
layout only for streams with matching number of channels, so this patch
restores the very old behaviour of ignoring mismatching layouts. See the
discussion in ticket #11016.

Signed-off-by: Marton Balint <cus at passwd.hu>
---
 fftools/ffmpeg_demux.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/fftools/ffmpeg_demux.c b/fftools/ffmpeg_demux.c
index 1ca8d804ae..2e77210237 100644
--- a/fftools/ffmpeg_demux.c
+++ b/fftools/ffmpeg_demux.c
@@ -1399,13 +1399,14 @@ static int ist_add(const OptionsContext *o, Demuxer *d, AVStream *st)
                 av_channel_layout_uninit(&par->ch_layout);
                 par->ch_layout = ch_layout;
             } else {
-                av_log(ist, AV_LOG_ERROR,
-                    "Specified channel layout '%s' has %d channels, but input has %d channels.\n",
+                av_log(ist, AV_LOG_WARNING,
+                    "Specified channel layout '%s' has %d channels, but input has %d channels. Ignoring specified layout.\n",
                     ch_layout_str, ch_layout.nb_channels, par->ch_layout.nb_channels);
                 av_channel_layout_uninit(&ch_layout);
-                return AVERROR(EINVAL);
+                ch_layout_str = NULL;
             }
-        } else {
+        }
+        if (!ch_layout_str) {
             int guess_layout_max = INT_MAX;
             MATCH_PER_STREAM_OPT(guess_layout_max, i, guess_layout_max, ic, st);
             guess_input_channel_layout(ist, par, guess_layout_max);
-- 
2.35.3



More information about the ffmpeg-devel mailing list