eac3to - audio conversion tool [Archive] - Page 62

madshi

27th January 2008, 16:40

eac3to v2.20 released

http://madshi.net/eac3to.zip

* changed VC-1 muxing method to fix problems with several movies, e.g.
- Unforgiven
- Phantom of the Opera
- Million Dollar Baby
- Fantastic Four 2
* fps value is now also added to MKV header when muxing raw VC-1 stream
* added new "-skip" option to skip corruption in the beginning of an EVO file
* added extra handling which fixes some EVO authoring bugs
This release should fix all remaining VC-1 related problems including the mentioned movies (see above).

Found out what the problem with some of the movies was. For those who are interested in the technical details, here's the explanation: VC-1 video streams (like most video streams) consist not only of frames, but also contain additional information records. With most VC-1 movies we've seen so far these information records stayed the same throughout the whole movie. Now Haali's VC-1 muxing drops all those information records and stores only one instance of those records at one place in the MKV. This works fine for all those movies where those information records don't change. But with Unforgiven, Million Dollar Baby and Fantastic Four 2 (and probably also with Phantom of the Opera USA) the information records are changing all the time throughout the movie!! Now eac3to v2.20 stores all those information records into the MKV. Basically if you demux the VC-1 stream from a v2.20 muxed MKV you'll get the full original VC-1 stream back. This wasn't the case with any VC-1 MKVs you've seen until to day (to my best knowledge). Storing those information records into the MKV solved most problems. Also it improved seeking stability. With older VC-1 MKV muxes when I was seeking sometimes the image broke up shortly and then recovered again. With v2.20 muxes this doesn't seem to happen anymore at all.

There's one problem which still remains: If the EVO is really heavily damaged, this might still screw up the VC-1 decoders. This happens with the Searchers and Million Dollar Baby samples I've received from you guys. With the Million Dollar Baby sample the corruption is very early in the EVO only, so you can use the new "-skip" option to jump over the corruption. For the sample I received "-skip2" does the trick. When using "-skip" you need to manually correct audio sync, though. With the Searchers sample I received there's corruption in the middle of the movie. Sorry, no way for me to handle this properly. The MS VC-1 decoder totally goes crazy if the VC-1 stream is really heavily damaged and it doesn't seem to be able to recover from such a situation.

For all intends and purposes, I believe eac3to v2.20's VC-1 handling should be pretty perfect now.

nautilus7

27th January 2008, 17:40

Chumbo

27th January 2008, 17:51

@madshi,
Wow and thank you. You keep outdoing yourself. :)

madshi

27th January 2008, 17:57

Thanks again madshi for your efforts!

I will test with million dollar baby. Regarding the audio sync you mention... Using -skip2 (skips the first 2 VC-1 sequence headers according to the 1st post) how can i figure the delay? How many frames are removed? I mean by knowing the # of frames removed and the fps (frame duration) i can find it out, right?

How is the original evo delay handled in this case (when skip is used)?
When the EVO is corrupted, audio sync is problematic, anyway. eac3to will even flat out refuse to demux corrupted audio at all. So you'll have to use EvoDemux to demux the audio, run the demuxed audio file through delaycut to clean it up and then finally you'll have to manually find the right delay to get audio in sync. It's not nice, but it's doable. I currently do not plan to invest any more time in fixing or working around damaged/corrupted EVOs.

@madshi,
Wow and thank you. You keep outdoing yourself. :)
:D

Richum

27th January 2008, 18:44

Transformers was never really broken and the problem was not, to my knowledge, ever with eac3to. I changed video and audio decoders and it now plays back flawlessly with Windows Media Player 11. Still has problems playing in Video Lan C and Mplayer however.

For those interested in what worked for me, from Graphedit;

Transformer MKV ------>video-->CoreAVC Video Decoder---->VMR input0/Video Renderer

Transformer MKV ------>Audio-->Intervideo Audio Decoder--->Default Direct Sound Device

Now on to 12 monkeys, I am using 2.20 and the -skip2 option to see if I can make it play properly.

Thunderbolt8

27th January 2008, 18:59

:thanks:

nautilus7

27th January 2008, 18:59

Now on to 12 monkeys, I am using 2.20 and the -skip2 option to see if I can make it play properly.Just to make it clear, in case you haven't read the 1st post...

Skip option skips the first 1..99 VC-1 sequence headers, so in your case (12 monkeys) maybe -skip3 or skip6 or skip?? does the trick.

But how do we know what to choose? Does it have to do the point at which the gaps/overlaps occur? How many frames are removed by removing one sequence header? :confused:

Madshi (or anyone) explain us please.

Thunderbolt8

27th January 2008, 19:21

madshi

27th January 2008, 20:02

But how do we know what to choose? Does it have to do the point at which the gaps/overlaps occur? How many frames are removed by removing one sequence header? :confused:
Obviously if remuxing works without skipping than you shouldn't skip. If the EVO is damaged right at the beginning, try "-skip1" first. If that doesn't fix the problem, try "-skip2" and so on until the problem is solved. But before you start using the "-skip" option, first try to rerip the movie. Because of course not using "-skip" is much better than using it.

madshi

27th January 2008, 20:04

btw. do you think that a similar problem could be the cause for all the h264 problems (rainbows etc), that some muxers just drop data, which is not clearly video data, as with vc-1, but which is still necessary?
Current focus is on VC-1. Once you guys can confirm that it works fine now I can move on to MPEG2 and h264. I will not answer any questions about h264 right now. One step at a time, please.

rory83

27th January 2008, 21:16

intomed

27th January 2008, 21:26

Madshi,

YOU ARE THE MAN!!!

Version 2.20 fixed the POTO problems with the video mkv files. Later today I will package it all together with the audio and let you know.

If you will excuse my ignorance, once again, and allow me one more question. When I add the audio (I guess the correct term is remux, right) to the mkv, can I add more than 1 audio track? Also, the flac file that was created, do I need to do something special to it for me to use it. I guess what I'm asking, is there something that needs to be done to it to enable it to be surround sound (discrete channels)?

Thanks for all your help and hard work. It is greatly appreciated. :thanks:

madshi

27th January 2008, 21:26

I don't suppose one additional option could be to write the mkvs in smaller sized chunks could it please? I'm ending up with 15GB .mkvs which sadly my NAS won't accept since it hits its 4GB max. I could run it through mkvmerge, but I don't know whether that would use its own splitter internally and get back to the same issues that Madshi's internal splitter now fixes.
You can use mkvtoolnix to split big MKVs into smaller pieces. Don't know how well playback of such splitted pieces works, though.

I could run it through mkvmerge, but I don't know whether that would use its own splitter internally and get back to the same issues that Madshi's internal splitter now fixes.
mkvmerge shouldn't destroy what eac3to has built, as long as you don't rewrite the timestamps. But in the end you just need to try out how well it works for you. I'm not sure if Haali's Media Splitter (which is the state of the art filter to playback MKV files) supports splitted files.

madshi

27th January 2008, 21:32

Version 2.20 fixed the POTO problems with the video mkv files.
Finally!! :) Thanks for testing...

If you will excuse my ignorance, once again, and allow me one more question. When I add the audio (I guess the correct term is remux, right) to the mkv, can I add more than 1 audio track?
You can add as many audio tracks and subtitles as you like. "Remux" usually means converting video and/or audio tracks from one container to another. What you want to do is add audio tracks to an already existing container. I'd not call that remuxing but rather muxing or simply adding. But as long as I understand what you're talking about using exact terms doesn't matter that much to me.

Also, the flac file that was created, do I need to do something special to it for me to use it. I guess what I'm asking, is there something that needs to be done to it to enable it to be surround sound (discrete channels)?
I'd suggest to use madFlac as a decode filter. Apart from that you don't need anything special. Of course you need to make sure that your transport method to your receiver supports uncompressed multi channel audio. The easiest way for that would be analog connection at this point in time. SPDIF doesn't support multi channel uncompressed audio transport, unfortunately...

DeepBeepMeep

27th January 2008, 22:07

Great! my previous observations regarding Eac3to hanging is no longer true. Now when I try to remux my EVO file to a MKV file I get a "the track is not clean. process aborted".

So I tried instead to avoid demuxing the audio tracks:
Eac3to source.evo 1: target.mkv
This works very well since the end result can be remuxed with the audio tracks obtained with Evodemux. And then I got a perfect playback !

However, there must be something wrong with the remuxing: my dual core CPU E6600 has 100% CPU usage during all the process which takes 1 hour 10 minutes. I would be great if before you get to H264 remuxing you might think of optimizing this one as usually VC1 remuxing takes 10 minutes for EVO files.

Anyway, thanks for solving the mystery regarding the glitched EVO VC1 files. It seems the Haali filter had the same issue (that is discarding extra info that kept changing) with EVO and M2TS H264 files but Haali has fixed it since it had been reported to him.

It is quite exciting, we are getting closer and closer of perfect tool that will do all the job to turn a HDDVD/BluRay into a mkv file. Keep on the good work!

Richum

27th January 2008, 22:10

Just to make it clear, in case you haven't read the 1st post...

Skip option skips the first 1..99 VC-1 sequence headers, so in your case (12 monkeys) maybe -skip3 or skip6 or skip?? does the trick.

But how do we know what to choose? Does it have to do the point at which the gaps/overlaps occur? How many frames are removed by removing one sequence header? :confused:

Madshi (or anyone) explain us please.

Nautilus7,

I understand that -skip is to be used in increments, however from my previous experience with 12 Monkeys using FFmpeg, I remembered that it had some serious audio sync issues, such that the audio is several seconds ahead.

So I just guessed that skip2 would be a good starting place, and it was. 12 Monkeys has now been successfully converted with no stutter but it still has very small amount of audio sync issue, I am going to try a bit of delay in MKVmerge, about 500 milliseconds should do it.

Great work Madshi!

rickardk

27th January 2008, 22:14

Million Dollar Baby plays perfect!!!!!
Wow, you're good!!!

madshi

27th January 2008, 22:16

However, there must be something wrong with the remuxing: my dual core CPU E6600 has 100% CPU usage during all the process which takes 1 hour 10 minutes. I would be great if before you get to H264 remuxing you might think of optimizing this one as usually VC1 remuxing takes 10 minutes for EVO files.
Can you please post the eac3to log here? Where is that 10 minutes number from?

madshi

27th January 2008, 22:17

Million Dollar Baby plays perfect!!!!!
Wow, you're good!!!
Good news!

Is that with or without using the "-skip" option? The sample that I have from Million Dollar Baby needs "-skip2" but it could be a damaged/corrupted rip.

madshi

27th January 2008, 22:18

I understand that -skip is to be used in increments, however from my previous experience with 12 Monkeys using FFmpeg, I remembered that it had some serious audio sync issues, such that the audio is several seconds ahead.

So I just guessed that skip2 would be a good starting place, and it was. 12 Monkeys has now been successfully converted with no stutter but it still has very small amount of audio sync issue, I am going to try a bit of delay in MKVmerge, about 500 milliseconds should do it.
Do you need to use "-skip" at all? Have you tried without it?

I mean now you cleaned up your DirectShow filters and got Transformers working correctly, maybe 12 Monkeys works just fine for you now, too? Without having to use "-skip"? At least I can say that "-skip" is not about fixing audio sync, it's about fixing rainbow (that is corrupted) video frames.

madshi

27th January 2008, 22:22

Does anybody have the "The Searchers" HD DVD and can do a rerip and check whether there are still problems? It's the only VC-1 sample I have that I cannot get to work at all. I want to believe that it's a bad rip. However, the audio streams seem to be 100% ok. So I'm not really sure about the bad rip theory...

nautilus7

27th January 2008, 22:30

Good news!

Is that with or without using the "-skip" option? The sample that I have from Million Dollar Baby needs "-skip2" but it could be a damaged/corrupted rip.

I did it (million dollar baby) with -skip2 and everything looks fine! Would you like to do it again without the skip option (i can't re-rip though, i can try the one with corrupted frames).

rickardk

27th January 2008, 22:35

Tried 2 diffrent sources. Same exact problem. Used -skip2. But that does not cut anything out from the movie so its ok.

madshi
Should I redo all my VC-1 HD DVD with v2.20? To preserve the complete VC-1 stream and don't rewrite timestamps.

And maybe try demuxing VC-1 Blu-rays and muxing them with eac3to?

madshi

27th January 2008, 22:45

Tried 2 diffrent sources. Same exact problem. Used -skip2. But that does not cut anything out from the movie so its ok.
Ok, so it seems to be an authoring error. Pretty sad...

Should I redo all my VC-1 HD DVD with v2.20? To preserve the complete VC-1 stream and don't rewrite timestamps.
I'm not sure if it's worth it. Do v2.20 muxes play smoother for you than your old (v2.14 and older) muxes? If so, then I'd redo all VC-1 HD DVD movies. If no, then I'm not sure what to recommend. Personally, I'm going to redo all my VC-1 HD DVD movies sooner or later. (BTW, did you have a chance to retest ATI vs. NVidia with a new v2.20 remux yet?)

And maybe try demuxing VC-1 Blu-rays and muxing them with eac3to?
You can do that. But if you ask me, I'd be happy if you could save the Blu-Rays for later. I mean I still need you for testing next year (just joking) when eac3to has learned how to remux Blu-Ray directly... :p

BTW, do your Blu-Ray VC-1 remuxes (if you have some) play totally smoothly? Or do you have some minor stuttering in them, too? In other words: Is there something to gain from redoing them right now?

madshi

27th January 2008, 22:47

I did it (million dollar baby) with -skip2 and everything looks fine! Would you like to do it again without the skip option (i can't re-rip though, i can try the one with corrupted frames).
No need for redoing the test as rickardk has already confirmed that "-skip2" is necessary in any case. It seems that the standalone players must be pretty robust against corrupted EVO files!

DeepBeepMeep

27th January 2008, 22:59

Can you please post the eac3to log here? Where is that 10 minutes number from?

The command line is:
"eac3to g:\p.evo 1: d:\p.mkv"

and the correspondig log:
"eac3to v2.20
eac3to g:\p.evo 1: d:\p.mkv
EVO, 1 video track, 2 audio tracks, 2:21:06
1: VC-1, 1080p24 /1.001
2: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
3: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Muxing video to Matroska...
Video has a gap of 7 frames at playtime 0:00:01.
Added fps value to MKV header.
Video track 1 contains 202971 frames.
eac3to processing took 1 hour, 6 minutes.
Done."

100% CPU on a dual core during all the remuxing

10 to 20 minutes is usally what it takes to remux an EVO file using Haali GSdemux. Of course for this particular file, Haali filter won't work since it doesn't support the ever changing VC1 info you have identified.

Note that if I use the following command line
"eac3to g:\p.evo d:\p.mkv"

instead I get:

"EVO, 1 video track, 2 audio tracks, 2:21:06
1: VC-1, 1080p24 /1.001
2: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
3: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Extracting audio track number 3...
Removing dialog normalization...
Extracting audio track number 2...
Removing dialog normalization...
This EVO file contains authoring faults. Will try to work around that.
This track is not clean. Processing aborted.
This track is not clean. Processing aborted.
Please clean the track with delaycut and then retry eac3to.
Please clean the track with delaycut and then retry eac3to."

rickardk

27th January 2008, 22:59

Ok, so it seems to be an authoring error. Pretty sad...

I'm not sure if it's worth it. Do v2.20 muxes play smoother for you than your old (v2.14 and older) muxes? If so, then I'd redo all VC-1 HD DVD movies. If no, then I'm not sure what to recommend. Personally, I'm going to redo all my VC-1 HD DVD movies sooner or later. (BTW, did you have a chance to retest ATI vs. NVidia with a new v2.20 remux yet?)

You can do that. But if you ask me, I'd be happy if you could save the Blu-Rays for later. I mean I still need you for testing next year (just joking) when eac3to has learned how to remux Blu-Ray directly... :p

BTW, do your Blu-Ray VC-1 remuxes (if you have some) play totally smoothly? Or do you have some minor stuttering in them, too? In other words: Is there something to gain from redoing them right now?

The Island and Goodfellas I think are the only titles on Blu-ray I encountered with micro stutters.

Ok, I will wait with Blu-ray. Im also going to wait with H264 HD DVDs (because diffrent muxing methods makes ffdshow reports diffrent fps even though codec, timecode and header are the same).

Going to redo all HD DVD VC-1 titles for now...

I will test nvidia vs ati again..next weekend

nautilus7

27th January 2008, 23:15

@ Madshi

Can eac3to downmix to 2.0 a 4.1 track? See here (http://forum.doom9.org/showthread.php?t=134197), please.

Richum

27th January 2008, 23:37

Do you need to use "-skip" at all? Have you tried without it?

I mean now you cleaned up your DirectShow filters and got Transformers working correctly, maybe 12 Monkeys works just fine for you now, too? Without having to use "-skip"? At least I can say that "-skip" is not about fixing audio sync, it's about fixing rainbow (that is corrupted) video frames.

Either your 2.20 version or the -skip option fixed the problem, I will try it without the -skip and see if it has issues that way. I suppose we need to know. That would explain a lot, to me anyway.

Thunderbolt8

27th January 2008, 23:54

btw. I also did 12 monkeys few days ago (could have been with 2.14 though) and there are no issues of any kind for me in it. so it could be a problem of your rip.

Yraen

28th January 2008, 00:15

madshi,

Does "The last (E-)AC3 frame is incomplete and thus gets skipped." mean anything important? This popped up in 2.20 where it wasn't there in 2.19. This 2.19 was done after the error in 2.20. At this point in the audio there is only silence so nothing has been lost. It was just weird to get an error where I had never gotten one before.

eac3to v2.19
"D:\editing apps\eac3to\2.19\eac3to.exe" "H:\HD\THE_THING_HDDVD\HVDVD_TS\PEVOB_1.EVO"+"H:\HD\THE_THING_HDDVD\HVDVD_TS\PEVOB_2.EVO" 3: "h:\thing\219audio.ac3" -640
EVO, 1 video track, 3 audio tracks, 1:48:42
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: E-AC3, 5.1 channels, 1536kbit/s, 48khz, dialnorm: -27dB
4: E-AC3, 2.0 channels, 448kbit/s, 48khz, dialnorm: -27dB
5: E-AC3, 2.0 channels, 384kbit/s, 48khz, dialnorm: -27dB
Extracting audio track number 3...
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...
DirectShow reports 5.1 channels, 24 bits, 48khz
Encoding AC3...
Creating/writing file "h:\thing\219audio.ac3"...
Video track 2 contains 156382 frames.
eac3to processing took 42 minutes, 21 seconds.
Done.
eac3to v2.20
"D:\editing apps\eac3to\2.20\eac3to.exe" "H:\HD\THE_THING_HDDVD\HVDVD_TS\PEVOB_1.EVO"+"H:\HD\THE_THING_HDDVD\HVDVD_TS\PEVOB_2.EVO" 2: "h:\thing\video.mkv" -strippulldown 3: "h:\thing\audio.ac3" -640
EVO, 1 video track, 3 audio tracks, 1:48:42
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: E-AC3, 5.1 channels, 1536kbit/s, 48khz, dialnorm: -27dB
4: E-AC3, 2.0 channels, 448kbit/s, 48khz, dialnorm: -27dB
5: E-AC3, 2.0 channels, 384kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Muxing video to Matroska...
Extracting audio track number 3...
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...
DirectShow reports 5.1 channels, 24 bits, 48khz
Encoding AC3...
Creating/writing file "h:\thing\audio.ac3"...
The last (E-)AC3 frame is incomplete and thus gets skipped.
Added fps value to MKV header.
Video track 2 contains 156382 frames.
eac3to processing took 45 minutes, 59 seconds.
Done.

Richum

28th January 2008, 00:23

btw. I also did 12 monkeys few days ago (could have been with 2.14 though) and there are no issues of any kind for me in it. so it could be a problem of your rip.

I suppose it could be, but have I ripped about 5 or so HD DVDs and it and Transformers were the only ones with issues. Transformers is H264 and 12 Monkeys is VC1, Transformers was not right because I did not have the correct chain of video decoder---> directshow filter.

12Monkeys has been ripped multiple times and so attributing a bad rip to only one out 5 HD DVDs seems unlikely to me. More likely is my video codecs are faulty or HD DVD disk itself has problems. I know for certain they can be, as I have returned one HD DVD that would not play about 5 or 6 months ago.

I am going to rip 12 monkeys one more time and run it through eac3to v.2.20 as a test without the -skip option as Madshi has suggested/requested.

nautilus7

28th January 2008, 00:36

Does "The last (E-)AC3 frame is incomplete and thus gets skipped." mean anything important? This popped up in 2.20 where it wasn't there in 2.19. This 2.19 was done after the error in 2.20. At this point in the audio there is only silence so nothing has been lost. It was just weird to get an error where I had never gotten one before.
Skipping the last frame means the last ~5ms of the audio are removed. As you said there is usually silence there, so nothing to worry about, but it's really confusing why 2.19 worked without complains about this, while 2.20 not. :confused:

Laffer

28th January 2008, 02:10

First off I'm no noob :) Need to downsample a DTS track. I have tried everything to get AC3TO working (now on 4 different machines including two fresh installs of XP SP2, one virtual) I keep getting the "not working correctly" on Nero (7) and Sonic (4.3.0.192) audio decoders.

I actually want to use the sonic decoder. All the software is legit.

Can anyone take me through getting this to work, say from a freshly installed XP SP2 (including all updates)?? So any codec packs used any software that needs to be installed etc etc.

Short of some voodoo ritual - I have run out of Ideas.....

Thanks,

Laffer.

Richum

28th January 2008, 02:53

OK, from a fresh rip, using v.2.20 and no options save -640 12 monkeys is perfect. Video and Audio in sync and no stutter or rainbows whatsoever. Seems v.2.20 is handling the 12 Monkeys correctly now or, I finally got the elusive "good rip" :}

Good job Madshi!!!

rickardk

28th January 2008, 03:20

I also get "The last (E-)AC3 frame is incomplete and thus gets skipped" on some titles that gave no errors before. Is eac3to more accurate now? Or is it a bogus message?

I have a couple of ts files with VC-1 in them that I would like to get into mkvs. I guess (with latest improvements of eac3to) it would be better do demux the VC-1 stream and mux it into mkv with eac3to then to do the old remux method with gdsmux?

madshi

28th January 2008, 08:26

The command line is:
"eac3to g:\p.evo 1: d:\p.mkv"

and the correspondig log:
"eac3to v2.20
eac3to g:\p.evo 1: d:\p.mkv
EVO, 1 video track, 2 audio tracks, 2:21:06
1: VC-1, 1080p24 /1.001
2: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
3: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Muxing video to Matroska...
Video has a gap of 7 frames at playtime 0:00:01.
Added fps value to MKV header.
Video track 1 contains 202971 frames.
eac3to processing took 1 hour, 6 minutes.
Done."

100% CPU on a dual core during all the remuxing

10 to 20 minutes is usally what it takes to remux an EVO file using Haali GSdemux. Of course for this particular file, Haali filter won't work since it doesn't support the ever changing VC1 info you have identified.

Note that if I use the following command line
"eac3to g:\p.evo d:\p.mkv"

instead I get:

"EVO, 1 video track, 2 audio tracks, 2:21:06
1: VC-1, 1080p24 /1.001
2: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
3: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Extracting audio track number 3...
Removing dialog normalization...
Extracting audio track number 2...
Removing dialog normalization...
This EVO file contains authoring faults. Will try to work around that.
This track is not clean. Processing aborted.
This track is not clean. Processing aborted.
Please clean the track with delaycut and then retry eac3to.
Please clean the track with delaycut and then retry eac3to."
Thanks, I'll check out if I can improve performance. Should be possible to do if gdsmux is so much faster. This might take a while, though. Probably next weekend.

madshi

28th January 2008, 08:29

@ Madshi

Can eac3to downmix to 2.0 a 4.1 track? See here (http://forum.doom9.org/showthread.php?t=134197), please.
All "odd" formats like 3.0, 4.0, 4.1 and 5.0 are only "somewhat" supported. They all have funny channel mapping. I think I'd need to add full channel mapping information throughout the whole processing chain to make this work properly. This should be done somewhen as the current solution is not fully clean in this aspect. The current solution is really targetted at only handling 1.0, 2.0, 5.1, 6.1 and 7.1 correctly. But since the needed changes are not really small I won't do this soon.

madshi

28th January 2008, 08:30

Does "The last (E-)AC3 frame is incomplete and thus gets skipped." mean anything important? This popped up in 2.20 where it wasn't there in 2.19.
Could you please compare the length of the demuxed E-AC3 files v2.19 vs v2.20? I'm wondering whether v2.20 demuxes more or less than v2.19. Currently don't have time to look into this. Anyway, the difference should be minimal, so nothing to worry about...

madshi

28th January 2008, 08:31

First off I'm no noob :) Need to downsample a DTS track. I have tried everything to get AC3TO working (now on 4 different machines including two fresh installs of XP SP2, one virtual) I keep getting the "not working correctly" on Nero (7) and Sonic (4.3.0.192) audio decoders.

I actually want to use the sonic decoder. All the software is legit.

Can anyone take me through getting this to work, say from a freshly installed XP SP2 (including all updates)?? So any codec packs used any software that needs to be installed etc etc.
Not sure where the problem is coming from. eac3to simply checks whether a specific decoder can be made use of. In your code this for whatever reason fails. I don't know why.

madshi

28th January 2008, 08:34

I have a couple of ts files with VC-1 in them that I would like to get into mkvs. I guess (with latest improvements of eac3to) it would be better do demux the VC-1 stream and mux it into mkv with eac3to then to do the old remux method with gdsmux?
Depends on whether there are gaps in the video track or not. Usually there should be no gaps. In that case demuxing and muxing with eac3to should give you better (more exact) timestamps than gdsmux. Furthermore eac3to will mux the whole VC-1 data while gdsmux will drop some VC-1 information records. So I'd suggest to try using eac3to. But make sure that you check audio sync in both the beginning and end of the movie.