Bug 85407 – ffmpeg with amr support requires arch testing

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Chris White (RETIRED) 2005-03-15 13:09:15 UTC]

I have ffmpeg marked for the following arches:

~sparc ~alpha ~amd64 ~ia64 ~ppc64 ~arm ~mips ~hppa

and a new feature of ffmpeg is the ability to handle amr (adaptive multi-rate) decoding, commonly used in telephony and phone video situations.  As I do not know the clear endian nature or drawbacks on other arches, I'd like you all to confirm that it works on your arch.  If it appears horribly broken, please add "amr_nb" and "amr_wb" to your arch based use mask.  Once everyone has confirmed whether or not it works, I will commit the ebuild to the portage tree.  I've already tested x86 with ppc on the way.  Thanks ahead of time.

[Comment 1 Chris White (RETIRED) 2005-03-15 13:10:15 UTC]

Created attachment 53559 [details]
ffmpeg-0.4.9_p20050226-r2.ebuild

The ebuild.

[Comment 2 Chris White (RETIRED) 2005-03-15 13:27:30 UTC]

Test case would help...

You can find some videos with amr audio here:

http://mplayerhq.hu/MPlayer/samples/mobileVideo_3gp/

[Comment 3 Chris White (RETIRED) 2005-03-15 21:43:42 UTC]

*** Bug 71485 has been marked as a duplicate of this bug. ***

[Comment 4 Chris White (RETIRED) 2005-03-16 08:11:06 UTC]

Changing priority to enhancement so I know not to recognize this bug as one that prevents me from stable marking ffmpeg..

[Comment 5 Jason Wever (RETIRED) 2005-03-25 18:36:59 UTC]

Compilation (regardless of the new CFLAGS) dies with the following on SPARC64;

gcc -O3 -Wall -Wno-switch -mcpu=ultrasparc3 -O2 -pipe -I. -I'/var/tmp/portage/ffmpeg-0.4.9_p20050226-r3/work/ffmpeg-0.4.9-p20050226' -I'/var/tmp/portage/ffmpeg-0.4.9_p20050226-r3/work/ffmpeg-0.4.9-p20050226'/libavcodec -I'/var/tmp/portage/ffmpeg-0.4.9_p20050226-r3/work/ffmpeg-0.4.9-p20050226'/libavformat -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_GNU_SOURCE -c -o ffmpeg.o ffmpeg.c 
gcc -O3 -Wall -Wno-switch -mcpu=ultrasparc3 -O2 -pipe -I. -I'/var/tmp/portage/ffmpeg-0.4.9_p20050226-r3/work/ffmpeg-0.4.9-p20050226' -I'/var/tmp/portage/ffmpeg-0.4.9_p20050226-r3/work/ffmpeg-0.4.9-p20050226'/libavcodec -I'/var/tmp/portage/ffmpeg-0.4.9_p20050226-r3/work/ffmpeg-0.4.9-p20050226'/libavformat -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_GNU_SOURCE -c -o cmdutils.o cmdutils.c 
gcc -Wl,--warn-common -rdynamic -g  -o ffmpeg_g ffmpeg.o cmdutils.o -L./libavformat -lavformat -L./libavcodec -lavcodec -lm -lz -ldl -logg  -lvorbis -lvorbisenc -lfaac -lxvidcore
./libavcodec/libavcodec.so: undefined reference to `a52_syncinfo'
collect2: ld returned 1 exit status
make: *** [ffmpeg_g] Error 1


Note that the latest and greatest ffmpeg in the tree at the moment doesn't have this problem.

[Comment 6 chris roddy 2005-04-10 16:44:36 UTC]

except for the build error produced above (which can be avoided with USE="-dvd"), the build works just great for me on amd64. ffmpeg -formats reports "DEA   amr_nb".

audio encoded using the produced ffmpeg binaries plays flawlessly on my nokia 6230 telephone, but the produced ffplay binary cannot decode the audio, instead producing an ear-shattering BWAAAAAAAAAAAA!

thanks!
cmr

[Comment 7 Jeremy Huddleston (RETIRED) 2005-04-25 18:24:00 UTC]

closing

[Comment 8 chris roddy 2005-05-03 22:42:03 UTC]

I must admit, I don't understand why this is closed ... ?

[Comment 9 Simon Stelling (RETIRED) 2005-05-04 00:20:09 UTC]

-r3 is marked stable, that's what this bug was about

[Comment 10 chris roddy 2005-05-10 20:37:23 UTC]

in that case, should I open a new RFE for amr_nb support?