[FFmpeg-user] AC3 encoding problem - wrong audio tag?

Andy Furniss andyqos at ukfsn.org
Sat Jul 14 16:58:04 CEST 2012

Carl Eugen Hoyos wrote:
> Stefan Gofferje <ffmpeg-user <at> home.gofferje.net> writes:
>> What can I do...? :)
> Did you test if my interpretation of your problem is correct at all?
> (Iirc you did not post complete, uncut FFmpeg console output
> or did you?)
> Note that it is extremely simple to force FFmpeg to produce
> invalid files, just look at line 452 of libavformat/riff.c
> (make it channels > 20 or something similar)
>> The ffmpeg-developers seem to utilize an interpretation of a standard
>> which seems to be not much widely shared
> I googled for waveformatextensible yesterday and any other
> (less strict) interpretation of the standard seems truly absurd to me.
> After testing that this is indeed the problem, could you add more
> information about your setup? It seems extremely unlikely to me that
> a Dolby decoder cares at all about a riff / wav header...

May be taking the thread OT, but I don't know.

I do have access to an xbox if you want someone to test something.

One thing I notice is that if I make ac3 with ffmpeg then mediainfo 
shows the channels like -

Channel positions                        : Front: L C R, Back: L R, LFE
Channel positions                        : 3/0/2.1

I know mediainfo is not rated here :-) but the one thing I find it handy 
for is ac3. mediainfo -f will instantly show me if some "kind" person 
has ripped out drc/other metainfo. I don't know a quick way to see that 
with ffmpeg.

Testing dvds ac3 mediainfo shows -

Channel positions                        : Front: L C R, Side: L R, LFE
Channel positions                        : 3/2/0.1

ffmpeg -i says side for both cases - made with ffmpeg -

  Stream #0:1: Audio: ac3 ([0] [0][0] / 0x2000), 48000 Hz, 5.1(side), 
s16, 448 kb/s

and for a dvd

Stream #0:1[0x80]: Audio: ac3, 48000 Hz, 5.1(side), s16, 448 kb/s

More information about the ffmpeg-user mailing list