eac3to - audio conversion tool

[madshi]

@madshi,
If I may request a wish list item as you continue to update this please. Would you consider adding the number of frames on the video codec line please, along with the codec, fps, etc.? Many thanks.
Do you mean the number of frames that should be in the video if the runtime information is correct (which might be right or wrong)? Or do you mean the number of frames that are *actually* in there? The latter information is only available after having scanned through the whole file which eac3to does only if you do some real action on the EVO file(s). eac3to already posts the actual number of frames, but only after having done real processing on the EVO file(s).

[Chumbo]

Do you mean the number of frames that should be in the video if the runtime information is correct (which might be right or wrong)? Or do you mean the number of frames that are *actually* in there? The latter information is only available after having scanned through the whole file which eac3to does only if you do some real action on the EVO file(s). eac3to already posts the actual number of frames, but only after having done real processing on the EVO file(s).
The latter would be great. :)

[EDIT]madshi, I know evodemux has two frames values: calculated and counted. I normally don't use the Sonic stuff unless there are problems with the video stream and I'm forced to. When I have, the calculated number is what I've used and it's always worked fine. I'm assuming the counted evodemux uses is the actual after running through the whole stream. It may be valuable to also have something similar to calculated in addition to the actual. What do you think?

For those who use the Sonic stuff regulary, do you consistently use calcualted or counted from evodemux? Thanks.

[madshi]

The latter would be great. :)
You mean the number which is actually in there? As I said, eac3to cannot possibly know this number without reading and parsing the full source file(s) from the first to last bit.

[kabal223]

http://img528.imageshack.us/img528/4598/sshot3he5.th.png (http://img528.imageshack.us/my.php?image=sshot3he5.png)

Hello, nice test there!!

What is the name of that program to measure dts and dd tracks? any links to buy?

Thanks!!

[madshi]

What is the name of that program to measure dts and dd tracks? any links to buy?
That program is named Audacity and is freeware. However, it can't handle DTS and DD tracks. You need to decode those tracks to WAV by using eac3to. Audacity can then show those nice graphcs for the WAV files.

[kabal223]

That program is named Audacity and is freeware. However, it can't handle DTS and DD tracks. You need to decode those tracks to WAV by using eac3to. Audacity can then show those nice graphcs for the WAV files.

Thanks, very helpful!

Just one more question, if I recode the audio to wav the dinamic range remains? I just want to compare the dinamic range in tracks of the reencoded HD DVD/Blu-Ray audio with the original DVD's audio tracks.

Sorry for my bad english.

[nautilus7]

It depends on the decoder you use. eac3to can do it correctly.

This is how eac3to begun. The talking in the first pages is about this. eac3to uses reference decoders and "forces" them (when necessary) to output the audio data without any processing (DRC & DN). In other words using eac3to in combination with the decoders described in the 1st post, you get max quality.

[Jaja1]

Why don't you ask in the appropriate sub forum?I was so busy to prepare a evo sample for eac3to that this felt the natural place to ask. Silly of course and you're right, it doesn't belong here.

Edit: Is there any movie which is available in VC-1 on both HD DVD and Blu-Ray, *and* which shows these total garbling in HD DVD but not in Blu-Ray (after having remuxed both to MKV)? If so, maybe we could find out what the difference is exactly...
As I said before, the problems I see in VC1 HD DVD using Haali splitter, go from tiny stutters to a garbled mess. Unfortunately, in the only example I have in both evo and m2ts, all the evo has got are these tiny stutters. I've converted both to mkv, the evo using eac3to, the m2ts using gsdemux. Look at the moment Harry Potter drops the glass. When it shatters and slides away, the evo hangs for a very small moment. The m2ts does not. Moreover, the orginal evo doesn't hang using PowerDVD. So, what's up doc?
http://www.sendspace.com/file/fu2n45 and http://www.sendspace.com/file/8f9uy2


The other sample I have is an VC1 evo, that has distorted audio when played back using Haali splitter+VMR9 and becomes a green mess using Haali splitter + Haali video renderer. Converting the eac3 track to ac3 using eac3to -sonic, the ac3 plays back fine except for two annoying distortions. Here again, the sample has no distortions in PowerDVD.
http://www.sendspace.com/file/er2ql7

[madshi]

As I said before, the problems I see in VC1 HD DVD using Haali splitter, go from tiny stutters to a garbled mess. Unfortunately, in the only example I have in both evo and m2ts, all the evo has got are these tiny stutters. I've converted both to mkv, the evo using eac3to
Thanks for those samples. Could you please upload them as EVO/m2ts files? I can do the conversion to mkv myself. Having the original container format is much more helpful to me cause only that way I can try to find a solution to improve the conversion to MKV. If you only upload the MKVs then I can confirm that there's a problem but I can't find out how to avoid the problem in the first place.

Anyway, did you turn off timecode rewriting? The evo -> mkv file you sent me does not have rewritten timecodes. If you let eac3to rewrite the timecodes the tiny stutters will be gone.

So, what's up doc?
It's quite simple: The Haali Media Splitter parses EVO files in such a way that timecodes are sometimes a bit "funny". This can result in such tiny stutters. That's why eac3to is by default rewriting timecodes by using mkvtoolnix as a 2nd step after having muxed the EVO file to MKV by using the Haali filters.

The other sample I have is an VC1 evo, that has distorted audio when played back using Haali splitter+VMR9 and becomes a green mess using Haali splitter + Haali video renderer. Converting the eac3 track to ac3 using eac3to -sonic, the ac3 plays back fine except for two annoying distortions. Here again, the sample has no distortions in PowerDVD.
http://www.sendspace.com/file/er2ql7
Thanks, will check that out later, don't have time right now.

[nesNYC]

Firstly, I tried all your recommendations and nothing was working. So this is what I did to fix the problem:

1. took the original dtshd track and converted to *.wavs using eac3to
2. manually encoded the wavs with the SurCode encoder back to DTS.
3. Muxed the new DTS file and now everything is in sync.

Surprisingly, once the DTSHD file was converted to WAVs, the WAVs themselves were of the proper time length down to the millisecond. So I feel, the DTSHD file must have a "corrected" time in it's header telling the audio apps how fast to play?!?!? Could be a simple case of editing this header?

Now I know re-encoding might have degraded the sound a bit but really, I can't tell the difference AND, it sounds 100x better than the E-AC3 anyway so that's good enough for me :)

So madshi, if you get a hold of the Chronicles or Riddick, you may want to see how to edit the header and see if correction is possible this way. What it boils down to is that the original DTSHD file on the disk was probably corrected to play in real time from devices and once ripped the correction remains but is no longer needed giving a speeded up version.

Here's a screen dump showing the resluts:
http://bbs.undergroundnyc.com/dts-convert.jpg

Hope this helps others :D

First of all please demux with eac3to and not with EvoDemux. Not sure whether it makes a difference, but there are some movies where EvoDemux corrupts demuxed tracks. Next: How are you playing that DTS track? Are you playing the DTS track as an external file? If so, which DTS source filter are you using? If the DTS track is DTS-HD Master Audio I'd recommend to transcode that to FLAC. I'd recommend to transcode it to AC3 or FLAC for testing purposes, anyway. If the transcoded AC3 or FLAC doesn't have any audio sync issues then you know that the DTS file itself is alright and just the source filter makes trouble. If the transcoded AC3/FLAC file shows the same sync problems then the DTS-HD track is faulty.

[Chumbo]

You mean the number which is actually in there? As I said, eac3to cannot possibly know this number without reading and parsing the full source file(s) from the first to last bit.
Yes, that's fine because you can provide it when we scan the whole thing anyway. Although, per my added post above, I think the calculated number of frames would be useful too. :)

[Thunderbolt8]

theres a new pre-build version of mkvtoolnix available, which has the following bug fixed:

2007-12-31 Moritz Bunkus <moritz@bunkus.org>

* mkvmerge: bug fix: Rewrote the timecode application
code. Additionally force the "previous cluster timecodes" that
libmatroska uses to the current timecode. This seems to get rid of
libmatroska's assertions about the local timecode being to
small/big to fit into an int16. It also seems to get rid of some
of mkvmerge's errors about the packet queue not being empty.

this is no official release yet, but maybe with this change it already solves some muxing problems or also improves the compability to haali (had no time for tests so far).
http://www.bunkus.org/videotools/mkvtoolnix/win32/pre/

[madshi]

Firstly, I tried all your recommendations and nothing was working.
Have you tried the modified DTS source filter posted somewhere earlier in this thread?

So I feel, the DTSHD file must have a "corrected" time in it's header telling the audio apps how fast to play?!?!? Could be a simple case of editing this header?
A DTS file does not have a header. It consists of a multitude of DTS frames each of which have a little header on their own.

[madshi]

Yes, that's fine because you can provide it when we scan the whole thing anyway.
Well, eac3to v2.14 already does that.

Although, per my added post above, I think the calculated number of frames would be useful too. :)
The actual number of frames is better than the calculated number of frames, right? So why bother with the calculated number of frames? The number of frames can only make any sense if you demux/remux the video, right? And in that case eac3to already shows the actual number of frames today. So I don't really see what the calculated number of frames would be good for.

[madshi]

theres a new pre-build version of mkvtoolnix available, which has the following bug fixed:

[...]

this is no official release yet, but maybe with this change it already solves some muxing problems or also improves the compability to haali (had no time for tests so far).
http://www.bunkus.org/videotools/mkvtoolnix/win32/pre/
It does improve the compatability to Haali, I had already tested that some days ago. E.g. it was not possible to rewrite the Equilibrium timestamps before when having remuxed Equilibrium to MKV by using Haali's filters. Now with the new version it works. Unfortunately seeking in Equilibrium is still a pain even with the rewritten timestamps - simply because the GOPs are monstrous.

[napalm-187]

hi

i have problems to decode an dts hd stream with eac3 2.14 and nero

eac3 to brings errors and i will fix this with delaycut, after this step is the same, errors and eac3 abort the progress.

the second prob i have is an 6.1 dts es stream to decode to wave

with nero decode only 6 waves but not the 7

is here sonic 4.3 audiodecoder recommed for this?

thx

[Thunderbolt8]

so if we use this pre version of mkvtoolnix already, then eac3to will also use this code?

and regarding equilibrium, is it possible to say that the rewriting of timestamps with mkvmkerge is 100% fine now in this case and the remaining seeking problems are only due to the movie's internal structure, or could it still be that its because of both, that there are still some minor issues and its not 100% fine in that case and these issues still contribute a small part which makes the seeking difficult (=is making a remux a safe solution now to get an identical working backup of the movie or would it still be better to wait for another update)?

[Chumbo]

Well, eac3to v2.14 already does that.
Very cool, sorry I had not noticed it.

The actual number of frames is better than the calculated number of frames, right? So why bother with the calculated number of frames? The number of frames can only make any sense if you demux/remux the video, right? And in that case eac3to already shows the actual number of frames today. So I don't really see what the calculated number of frames would be good for.
This one is mostly because I've noticed many posts that say to use the calculated frames number from EVOdemux as opposed to the counted. I'm not sure why, but it seems that the calculated is the right number in that case although it seems backwards to me, i.e., you'd think the counted would be correct after you've processed the entire file.

[madshi]

i have problems to decode an dts hd stream with eac3 2.14 and nero

eac3 to brings errors and i will fix this with delaycut, after this step is the same, errors and eac3 abort the progress.
Maybe it's too corrupt to be fully fixed, don't know. Just rerip the audio stream.

the second prob i have is an 6.1 dts es stream to decode to wave

with nero decode only 6 waves but not the 7

is here sonic 4.3 audiodecoder recommed for this?
Yes. See first post of this thread.

[madshi]

so if we use this pre version of mkvtoolnix already, then eac3to will also use this code?
What do you mean with "this code"? If you install the mkvtoolnix beta, eac3to will use it, of course.

and regarding equilibrium, is it possible to say that the rewriting of timestamps with mkvmkerge is 100% fine now in this case and the remaining seeking problems are only due to the movie's internal structure, or could it still be that its because of both, that there are still some minor issues and its not 100% fine in that case and these issues still contribute a small part which makes the seeking difficult (=is making a remux a safe solution now to get an identical working backup of the movie or would it still be better to wait for another update)?
To be honest: I don't know yet.

[madshi]

This one is mostly because I've noticed many posts that say to use the calculated frames number from EVOdemux as opposed to the counted. I'm not sure why, but it seems that the calculated is the right number in that case although it seems backwards to me, i.e., you'd think the counted would be correct after you've processed the entire file.
Don't know. FWIW, I've seen eac3to report slightly different counted frames than EvoDemux.

[nincollector]

havent had any experience with blu ray yet but im getting a drive today and was wondering if this tool would do the same / work with blu rays m2ts files?

[madshi]

havent had any experience with blu ray yet but im getting a drive today and was wondering if this tool would do the same / work with blu rays m2ts files?
No, not yet. This tool does do audio conversion for Blu-Ray audio tracks. But you need to demux the audio tracks by using xport or TsRemux.

[nincollector]

ok thanks for the great tool. im assuming by your "not yet" reply, it is in the works?

[madshi]

ok thanks for the great tool. im assuming by your "not yet" reply, it is in the works?
No, not yet... :)

[nincollector]

ok gotcha its not yet not yet in the works... :thanks:

[esoteradactyl]

Great Tool!! So far ive done a quite a few hd-dvd's and everything has came out perfect. Now its time to tackle my bluray discs. Just want to check with the pro's here if these are the correct steps for bluray discs.

POTC 3 is the flick im going try first

1. demux the pcm file with xport
2. run the demuxed pcm through eac3to and convert to flac
3. demux the video with either tsremux or gdsmux (does it matter which app?)
4. drop audio and video into mkvmerge and finish

sound about right?

[nautilus7]

The audio part is ok.
For the video, i believe is better to search around the forum.

[Thunderbolt8]

ok thanks for the great tool. im assuming by your "not yet" reply, it is in the works?
it will come sooner or later when all studios (unfortunately!!!) should have gone to blu-ray :P

[moshmothma]

Great Tool!! So far ive done a quite a few hd-dvd's and everything has came out perfect. Now its time to tackle my bluray discs. Just want to check with the pro's here if these are the correct steps for bluray discs.

POTC 3 is the flick im going try first

1. demux the pcm file with xport
2. run the demuxed pcm through eac3to and convert to flac
3. demux the video with either tsremux or gdsmux (does it matter which app?)
4. drop audio and video into mkvmerge and finish

sound about right?

Don't think u need to demux the video - just drag the m2ts into gdmux.

[esoteradactyl]

Don't think u need to demux the video - just drag the m2ts into gdmux.

unless i missed something, this makes no sense. wouldnt loading the file into gdsmux and selecting the video stream leave me with a demuxed video file in mkv?

[nincollector]

the tool works fine however the end MKV is very blocky and has weird color blocks in some areas? any idea what could cause this and or how to fix it?

[madshi]

unless i missed something, this makes no sense. wouldnt loading the file into gdsmux and selecting the video stream leave me with a demuxed video file in mkv?
It would leave you with a MKV file with the video in it. Just the same what eac3to does when remuxing EVOs to MKV.

However, gdsmux fails to work for some h264 Blu-Ray movies. You'll have to try if it works or not.

[madshi]

the tool works fine however the end MKV is very blocky and has weird color blocks in some areas? any idea what could cause this and or how to fix it?
Which movie is that and what exactly did you do?

[nincollector]

its the nine inch nails beside you in time hd dvd.

i used eac3to nin.evo nin.mkv

it all processed with no errors but when i play back the mkv it has blocks color distortion and stutters. possible my filters arent installed properly?

ive installed matroska splitter
mkvtoolnix
and matroska pack full 1.1.2

should i uninstall sonic?

if i demux the evo the VC1 plays fine.

[madshi]

its the nine inch nails beside you in time hd dvd.

i used eac3to nin.evo nin.mkv

it all processed with no errors but when i play back the mkv it has blocks color distortion and stutters.
What happens if you play the EVO file?

Can you upload a little sample with which I could reproduce this problem?

[napalm-187]

Maybe it's too corrupt to be fully fixed, don't know. Just rerip the audio stream.


Yes. See first post of this thread.

hi i have demuxxed the dts hd stream via graphedit
filesource- sonic hd demuxxer-- hdtv dump filter

now is the file clean, i think evodemux make here an failure

thx

[nincollector]

will include a sample sometime 2morro. how do i cut a mkv file?

[madshi]

hi i have demuxxed the dts hd stream via graphedit
filesource- sonic hd demuxxer-- hdtv dump filter

now is the file clean, i think evodemux make here an failure
Why don't you use eac3to for demuxing!? :confused:

[madshi]

will include a sample sometime 2morro. how do i cut a mkv file?
I need a sample of the EVO file. A sample of the MKV file doesn't help me at all. You can cut EVO files e.g. by using a hexeditor.

[Jaja1]

Could you please upload them as EVO/m2ts files? I can do the conversion to mkv myself. Having the original container format is much more helpful to me cause only that way I can try to find a solution to improve the conversion to MKV.
The evo part was 150mb and I tried to upload it two times. Both times the upload to sendspace crashed after approx. 90 minutes, so I thought will try that later. It worked now, so if you're still interested:http://www.sendspace.com/file/yd2aj0 and http://www.sendspace.com/file/j3yomn

Anyone knows a free file host with better uploadspeed than sendspace?

Anyway, did you turn off timecode rewriting?
Nope, but I know what went wrong. I just cut of the main part of the evo with a file splitter (don't know of a better way) and eac3to/mkvtoolnix(?) doesn't like that. It completes normally, so I wasn't paying much attentention and didn't see the message "The DirectShow decoder thread seems to hang". So there was no rewrite of timestamps. After rewriting the timestamps myself with mkvmerge, the tiny stutter was gone, as you said. Awesome.

It's quite simple: The Haali Media Splitter parses EVO files in such a way that timecodes are sometimes a bit "funny". This can result in such tiny stutters. That's why eac3to is by default rewriting timecodes by using mkvtoolnix as a 2nd step after having muxed the EVO file to MKV by using the Haali filters.
Thanks for the explanation. Do you happen to know whether Haali is aware of this?

So if you decode an E-AC3 track and a libav request for sending a sample pops up, *please* take the time to upload a little sample, so that the E-AC3 decoder programmer gets a little help from us.
This eac3 file crashes libav.
http://www.sendspace.com/file/4w77n0

[napalm-187]

Why don't you use eac3to for demuxing!? :confused:


hi

@madshi

im a GUI man and i hate cmd lines
often i am to stupid to use cmd lines:stupid:
:D

[rickardk]

eac3to fails on Sum of all fears.

eac3to feature_1.evo+feature_2.evo
after that nothing...

[nautilus7]

Is the above command the exact you type? There's no output man!

[rickardk]

Is the above command the exact you type? There's no output man!

What I meant is that I don't even get information about the title.

Evodemux demuxed without problem.
eac3to can't demux/remux or show information about the title.

[nautilus7]

Sorry! :D

[Chumbo]

eac3to fails on Sum of all fears.

eac3to feature_1.evo+feature_2.evo
after that nothing...
What else did you try? What do these result in?
- eac3to feature_1.evo
- eac3to feature_2.evo
- eac3to feature_1.evo -test

[rickardk]

What else did you try? What do these result in?
- eac3to feature_1.evo
- eac3to feature_2.evo
- eac3to feature_1.evo -test

does not work at all with feature_1.evo, feature_2.evo or feature_1.evo+feature_2.evo


feature_1.evo(or feature_2.evo, rebuilt.evo) -test gives:

Nero Audio Decoder (Nero 7 or older) works fine
Sonic Audio Decoder (4.3.0.169) works fine
Haali Media Splitter (2007-11-18) is installed
Surcode DTS Encoder doesn't seem to be installed
MkvToolnix (v2.1.0) is installed

then it hangs/stops

[calinb]

hi
im a GUI man and i hate cmd lines
:Dhaha--I hate GUIs. Can't batch 'em. :angry: eac3to is sweet for batch files! :D

[nincollector]

eac3to fails on Sum of all fears.

eac3to feature_1.evo+feature_2.evo
after that nothing...

what are you trying to do? you havent given it a destination. if you want to demux you need to add -demux to the end