View Full Version : FFmpeg/Mplayer supports E-AC-3 format now...
madshi
3rd April 2008, 09:57
I've been working feverishly to get eac-3 decoding to work in windows
How about "eac3to source.eac3 dest.ac3 -libav"? It doesn't get much easier than that.
Gannjunior
3rd April 2008, 12:06
exactly phhangover.
jruggle, tonight i will check my decoder version but i think i have used the last one.
ciao!
sl1pkn07
4th April 2008, 01:04
one noob question..
Which compiles before? ffmpeg, mplayer or is indifferent (I suppose that compiles the ffmpeg)
in case update ffmpeg... is needed recompile mplayer? or mplayer adds automatically changes on ffmpeg
sorry my bad english
greetings
pbjr
10th April 2008, 05:52
Hello, another noob here!
I can svn eac3 and compile ffmpeg no problem.
Whats the next step to get this version of ffmpeg with eac3 working with mplayer svn? Just copy the libav* directories over the ones that come with mplayer and compile? It looks like the svn ffmpeg is newer than the one that svn mplayer grabs.
How do I combine the two?
I'm grateful for anyones help. I've just blow 4 hours on trying to figure this out. Any chance of a new patch? :)
Thanks,
-PBjr.
Blue_MiSfit
17th April 2008, 02:50
How nice would it be to have eac3 decoding patched into ffdshow or AC3Filter :D
~MiSfit
jruggle
17th April 2008, 03:06
How nice would it be to have eac3 decoding patched into ffdshow or AC3Filter :D
That will probably be more likely once it is in the main FFmpeg branch, which is proving slower than expected. 7.1ch decoding is a bit difficult to fit into FFmpeg's current AC3 frame parsing framework. But it will happen eventually.
zonyl
18th April 2008, 21:34
I was able to assemble a working rev of ffmpeg for Kurtnoise's MLP patch and it compiled, however, I cannot get a clean rev of mplayer at the specified revision (external HEAD reference to libavcodec causes patch to fail).
Does someone have a tarball they can share of mplayer with MLP TrueHD support?
Kurtnoise
19th April 2008, 08:44
just wait...Here is an extract from the FFmpeg svn log:
rev 12896 - part 1 of EAC3 support - by Michael Niedermayer.
rev 12897 - Part 2 of EAC3 support, this is still disabled as it breaks regressions due to bugs elsewhere - by Michael Niedermayer.
concerning MLP, what is the problem exactly ?
Taktaal
22nd April 2008, 21:32
Hello. I linked to this thread from a guide and people are complaining that there's no more Windows version to download. Did you remove that?
Kurtnoise
23rd April 2008, 10:33
FFmpeg or mplayer ?
Most FFmpeg related stuff are included in the eac3to tool now. So...
About mplayer, yes I'm lazy to build every day/week a new build. Just hoping that all this stuff will be included soon in the svn tree...
madshi
16th May 2008, 16:19
That will probably be more likely once it is in the main FFmpeg branch, which is proving slower than expected. 7.1ch decoding is a bit difficult to fit into FFmpeg's current AC3 frame parsing framework. But it will happen eventually.
Hi Justin,
any news? :) Sorry, don't want to put any pressure on you. Just wondering...
Maybe it would make sense to skip 7.1 decoding for now and get 5.1 decoding into SVN first? I mean there's no real E-AC3 7.1 content available, anyway. All movies are only 5.1. The only 7.1 samples we have are from a Dolby demo disc... ;) And now with the death of HD DVD I don't think we'll see real 7.1 E-AC3 content anytime soon. It seems to me that Blu-Ray studios don't plan to ever make serious use of E-AC3. The consumers are crying too loud for lossless audio. So while 7.1 decoding would be nice to have, all we really need right now IMHO is 5.1 decoding...
jruggle
16th May 2008, 17:36
any news? :) Sorry, don't want to put any pressure on you. Just wondering...
Hi Madshi,
I've been somewhat busy with mentoring for this year's GSoC, among other things. Since it seems that Bartek has also moved on to working on his project for this summer, I'll go ahead and get back to finishing up E-AC3. Your suggestion makes sense. I'll get 5.1 into SVN first, then try to get those 7.1 samples working later.
madshi
16th May 2008, 17:53
Thank you for the heads up - and for your continued work on libav. It's much appreciated... :)
ACrowley
17th May 2008, 08:53
Is there any already Patched (EAC3, THD)mplayer Build ?
jruggle
2nd June 2008, 03:35
I've been somewhat busy with mentoring for this year's GSoC, among other things. Since it seems that Bartek has also moved on to working on his project for this summer, I'll go ahead and get back to finishing up E-AC3. Your suggestion makes sense. I'll get 5.1 into SVN first, then try to get those 7.1 samples working later.
Update: The version in svn.mplayerhq.hu/soc/eac3 should be pretty close to the final version which will be in FFmpeg svn. I might find something during the merging process, and I will probably have to make some changes during the final devel review as well.
Hopefully this should all be within the next few days.
Since the last post, the only major thing I've added is error concealment. The goal was to prevent sync loss as much as possible with damaged streams. It's also updated to current FFmpeg svn (r13613).
madshi
2nd June 2008, 07:51
Thanks... :)
compunett
2nd June 2008, 21:30
thx justin, your work is much appreciated :)
Just for curiosity how is going to be the SoC 2008 :)
jruggle
2nd June 2008, 22:28
thx justin, your work is much appreciated :)
thank you.
Just for curiosity how is going to be the SoC 2008 :)
Wondeful, I hope! You can see FFmpeg's projects at
http://code.google.com/soc/2008/ffmpeg/about.html
I am mentoring an MLP encoder and an ALAC encoder.
compunett
6th June 2008, 00:44
hehe nice then :)
I only knew about multimedia.cx wiki site. Btw isn't it difficult to mentor students around the world (for student or for you) ?
jruggle
6th June 2008, 22:07
...
Btw isn't it difficult to mentor students around the world (for student or for you) ?
There are some obstacles, but it's not too bad. One student is in Brazil, and therefore only 1 hour ahead of me. The other student is in India, so we have to adapt a bit.
Email is fine for reviewing patches. For more involved discussions, IRC works well. For communicating with the student in India, there is 9 hours between us, so I just have to stay up a little later, and he has to get up a little earlier.
compunett
7th June 2008, 14:20
OK then, don't forget to rest a bit :)
compunett
9th June 2008, 20:23
As you are all waiting, I followed the process this weekend, so far 2 third of the patches have been applied, some patches are still being reviewed, for Q&A, etc.
That's my 2 cent on the progress being made. Greats to all the commiters ;)
Mercury_22
25th June 2008, 17:42
Any progress ? :helpful:
jruggle
25th June 2008, 23:26
Any progress ? :helpful:
The final patches have been reviewed, so it is now only a matter of me finding the time to make the necessary changes. I'm hoping to do so before I go out-of-town next Tuesday.
Mercury_22
26th June 2008, 08:57
The final patches have been reviewed, so it is now only a matter of me finding the time to make the necessary changes. I'm hoping to do so before I go out-of-town next Tuesday.
Great news ! I can barely wait for this! :thanks:
iamlindoro
18th July 2008, 19:23
Well, I noticed that the MLP decoder has been merged into ffmpeg SVN, so that's good news. MLP now more or less works out of the box in mplayer and ffmpeg. Any news on E-AC3 being merged in the near future?
Also, it appears the demuxers for EVO and m2ts might need some work as one still needs to specify the audio codec and demuxer for these formats. Any move towards a solution to pick up the default audio track from the container, and use the correct demuxer and codec so I don't need to determine the individual command line for each HD rip?
Thanks so much for your hard work on this, Justin.
jruggle
18th July 2008, 22:48
Well, I noticed that the MLP decoder has been merged into ffmpeg SVN, so that's good news. MLP now more or less works out of the box in mplayer and ffmpeg. Any news on E-AC3 being merged in the near future?
Thanks so much for your hard work on this, Justin.
Getting the MLP decoder merged was all due to the hard work of Ramiro and of the original author Ian.
As for E-AC3, the last sticking point was the 6-pt IDCT used in E-AC3's Adaptive Hybrid Transform. The version in SoC SVN is a really slow brute-force approach. I finally got an optimized version of the function working a couple days ago, which is about 6x faster. Now I just have to submit a final set patches.
compunett
19th July 2008, 11:41
Great news
I was following the newsgroup on gmane, and commiters are really active in a lot of area, thx to them :)
jruggle
5th August 2008, 05:26
Update: I made quite a lot of commits today to update the SoC SVN repo. I'll be submitting another round of patches for review tomorrow.
lnt4069
5th August 2008, 07:33
Thanks Justin for all your work! This update has finally let me get sound working on my evo and m2ts files again in mplayer. I had it working with an earlier version of the SoC patch, but updating broke it and I was never able to get it working again, until now. Thanks again!
Mercury_22
5th August 2008, 09:02
Yes Thanks Justin ! :D :thanks:
So tomorrow's patches will be the last one , (for now) ?
madshi
5th August 2008, 09:20
Update: I made quite a lot of commits today to update the SoC SVN repo. I'll be submitting another round of patches for review tomorrow.
Sounds great!
The one thing I'm waiting for is the license change from GPL to LGPL. Is that still in the cards?
Thanks! :)
jruggle
6th August 2008, 01:25
So tomorrow's patches will be the last one , (for now) ?
I can only hope.
The one thing I'm waiting for is the license change from GPL to LGPL. Is that still in the cards?
That depends on the liba52 maintainer. He has expressed willingness in the past to allow relicensing of FFmpeg's AC-3 decoder as LGPL once E-AC-3 support is added. I will not ask him again until the code is in FFmpeg SVN.
Mercury_22
14th August 2008, 09:18
So, is it done ? :thanks:
jruggle
18th August 2008, 00:32
So, is it done ? :thanks:
it is now pending review. i still have not gotten a full review at this point, so it may take more work to get it committed. i'll post here when it is.
Mercury_22
18th August 2008, 14:01
it is now pending review. i still have not gotten a full review at this point, so it may take more work to get it committed. i'll post here when it is.
Ok ! :thanks: I can't wait :p :)
madshi
18th August 2008, 14:51
There's a lot happening right now at libavcodec. The E-AC3 decoder, an AAC decoder + encoder and an MLP encoder. All very exciting projects! :)
jruggle
31st August 2008, 04:50
FFmpeg now supports E-AC-3 decoding as of SVN revision 15103. :)
Snowknight26
31st August 2008, 05:35
24-bit?
madshi
31st August 2008, 08:15
FFmpeg now supports E-AC-3 decoding as of SVN revision 15103. :)
Congrats - well done! :)
Sorry for going on your nerves, but has the time come to ask about that other AC3 guy about LGPL now? :o
Mercury_22
31st August 2008, 12:20
FFmpeg now supports E-AC-3 decoding as of SVN revision 15103. :)
YES ! Thanks !
:thanks:
liels
2nd September 2008, 20:52
FFmpeg now supports E-AC-3 decoding as of SVN revision 15103. :)
w00t!!! Thanks!!!
Mercury_22
4th September 2008, 12:56
@ Jruggle
Ok maybe this is too much BUT... :)
PLEASE CAN YOU HELP WITH IMPLEMENTATION OF FFmpeg's E-AC-3 IN FFDshow since nobody can / know and I have no idea how to... (http://forum.doom9.org/showthread.php?p=1177585#post1177585)? :helpful::stupid:
P.S. Because if the end user CAN'T BENEFIT from your work ..... :(
madshi
4th September 2008, 13:15
PLEASE CAN YOU HELP WITH IMPLEMENTATION OF FFmpeg's E-AC-3 IN FFDshow
Why are you shouting? That's not very polite...
As far as I can see ffdshow uses libav for many audio codecs (FLAC, WMA and Vorbis, just to name a few) - but for whatever strange reason not for AC3. The only thing they'd have to do is to add an option to use libav for AC3 decoding, too. Shouldn't be hard to do, and it should automatically make E-AC3 decoding work, too, with the latest libav SVN. But this is a request you should make in the ffdshow thread... ;)
Inventive Software
4th September 2008, 14:04
24 bit decoding AFAIK doesn't work properly in ffdshow, and more work in the underlying layer needs to be done.
Mercury_22
4th September 2008, 14:31
Why are you shouting? That's not very polite...
As far as I can see ffdshow uses libav for many audio codecs (FLAC, WMA and Vorbis, just to name a few) - but for whatever strange reason not for AC3. The only thing they'd have to do is to add an option to use libav for AC3 decoding, too. Shouldn't be hard to do, and it should automatically make E-AC3 decoding work, too, with the latest libav SVN. But this is a request you should make in the ffdshow thread... ;)
Sorry if it seems like I'm shouting ! :p But I've already made a "request in the ffdshow thread" (http://forum.doom9.org/showthread.php?p=1177551#post1177551) and they make it sound like a very hard thing to do So....! ? !:confused::helpful::thanks:
jruggle
9th September 2008, 00:46
Hi everyone.
Even though the E-AC-3 decoder is in FFmpeg SVN now, I'm continuing to work on it in SoC SVN. I was very fortunate to be sent a sample which uses Spectral Extension, so I have now implemented support for it in the decoder. The code will be added to FFmpeg SVN soon.
On that note, I'm still looking for E-AC-3 samples. The more I get, the better. If it fails or gives warnings with the decoder, I really really want it! The major hurdle in implementing the full spec is availability of samples which use the enhanced features. So far, the most promising samples seem to be low-bitrate.
Edit: I'm pretty sure 2 of the really good samples I have are from Warner Home Video HD DVD's. So if anyone has any of those besides Casablanca and Rio Bravo, it would be good to get E-AC-3 samples from some others.
nautilus7
9th September 2008, 00:59
Nice to hear, but what exactly is Spectral Extension?
jruggle
9th September 2008, 01:14
Nice to hear, but what exactly is Spectral Extension?
It is a tool which an E-AC-3 encoder can use to only encode minimal information about upper bandwidth frequencies. Then the decoder basically copies the lower frequency data to the upper bandwidth area, and applies scaling and noise blending using that small amount of information provided in the bitstream.
Snowknight26
9th September 2008, 03:33
On that note, I'm still looking for E-AC-3 samples. The more I get, the better. If it fails or gives warnings with the decoder, I really really want it! The major hurdle in implementing the full spec is availability of samples which use the enhanced features. So far, the most promising samples seem to be low-bitrate.
How big do these samples need to be? Is it ok if the last frame is damaged because of the way the samples are cut?
vBulletin® v3.8.11, Copyright ©2000-2025, vBulletin Solutions Inc.