eac3to - audio conversion tool

[zeropc]

It can't be done in the first pass. A 2nd pass is necessary. It's not done automatically because sometimes it might be preferred to not fix the detected gaps/overlaps.

thanks for the clarification :)

[Boulder]

Please note that I'm not sure whether SSRC or r8brain is better for resampling. r8brain is a lot slower than SSRC, so hopefully it's a little bit better, but you be the judge. Because of the dramatic speed difference the highest quality SSRC mode is now the default resampling mode (also used for PAL speedup/slowdown). SSRC is somewhat limited in which conversions it likes to do exactly, though. Some sample rate conversions might be declined by SSRC. If you stumble over such a case, just use the "-quality=ultra" option to switch to r8brain instead.Many seem to prefer SSRC over r8brain, for example http://src.infinitewave.ca/ has some nice test result graphs.

[bmnot]

Is eac3to able to decode this track? I think it's simply a broken/corrupted TrueHD track.

please look at the sample: http://www.mediafire.com/?giwnn4qgtm2

(it's actually the whole file off the disc untouched)

[madshi]

for example http://src.infinitewave.ca/ has some nice test result graphs.
Yep, that's a very nice site. Back when I decided to use r8brain the site already existed but I think it didn't have SSRC test results at the time. Anyway, I find it hard to judge final quality by these test graphs. Just one example: The default behaviour seems to be linear phase. However, some filters intentionally use a different phase design, which seems to have both advantages and disadvantages. So which is better now? Also the help of that site says: "It is not always possible to judge subjective quality from the presented measurement results."

Specifically "r8brain Free" is better than "SSRC High Precision" in the first two test graphs, but worse (or different?) in the other test graphs. So which sounds better now? SSRC seems to be better than almost any other resample in the "Passband" and "Transition" graphs. But I'm wondering: Is this due to a better design? Or do the other resamplers intentionally take a hit in these two things with the intention to improve something else?

Many seem to prefer SSRC over r8brain
I've searched for comparisons but found none, except the infinitewave site you linked to. Where did you read that people prefer SSRC over r8brain?

Thanks!

please look at the sample: http://www.mediafire.com/?giwnn4qgtm2

(it's actually the whole file off the disc untouched)
Hmmmm... Have you tried reripping the file, just to be sure? There's no obvious corruption in the file, but it's still possible that there's a bit fault hidden somewhere. So reripping might still help. If you're 100% sure that the file is correctly ripped we might have to submit the file to the libav guys so they can check what's wrong. FWIW, the Nero Audio Decoder seems to be able to decode the TrueHD track just fine. However, it's limited to 5.1 output. Now I'm not sure if there's a bug in the libav decoder or whether the file is actually broken in some way. It's possible that the Nero decoder is just more forgiving compared to libav...

[ACrowley]

eac3to v2.69 released
* fixed: Sonic Decoder was incorrectly assumed to decode XXCh DTS files to 6.1



What does it mean ? Down it mean DTS ES 6.1 was not decoded properly by SonicDecoder ?

Ah, Implementing DIRAC would be superb...its one of the best Algor. fro Timestretching

[madshi]

What does it mean ? Down it mean DTS ES 6.1 was not decoded properly by SonicDecoder ?
There are some DTS and DTS-HD 6.1 tracks which the Sonic Decoder correctly decodes as 6.1. But there are other such tracks which the Sonic decoder only outputs as 5.1. It depends on which DTS extension is used for the additional channel. There are two different extensions, namely "XCh" and "XXCh". Most 6.1 tracks use "XCh" which the Sonic decoder can decode. But some are using "XXCh" which the Sonic decoder does not support. Older eac3to versions thought that Sonic would be able to decode every 6.1 track as 6.1. This caused eac3to to abort processing in specific situations when the Sonic decoder unexpectedly only decoded 5.1. Now the new eac3to version knows exactly which tracks are decoded as 5.1 and which as 6.1 by the Sonic decoder. So eac3to doesn't get surprised, anymore...

Ah, Implementing DIRAC would be superb...its one of the best Algor. fro Timestretching
Good to hear that it's a good algo. I plan on implementing DIRAC support soon. The free lib version looks easy enough to add.

[nwg]

Can this program convert a TrueHD track to a DTS one?

[nautilus7]

If you have a dts encoder, yes.

[nwg]

If you have a dts encoder, yes.

I just realised I have the Surcode one. Will that do and if so how would I do it? I am not too good at the command line bit as I prefer the Eac3to GUI and that seems to have not been updated for a while.

[nautilus7]

eac3to source.thd dest.dts

[nwg]

Thanks.

[nautilus7]

I have a question for DTS-ES Matrix tracks. Those tracks have 5.1 discrete channels and one (BC) matrixed into SL and SR, right? An A/V receiver capable of DTS-ES decoding outputs 6.1, correct? Shouldn't eac3to do the same? Are Sonic or Arcsoft decoders able to decode DTS-ES Matrix?

[kurt]

Fixed in v2.69.


thx, working now :)

D:\Movies\xxx>eac3to deutsch1.dts deutsch.ac3 -448
DTS-ES, 6.1 channels, 2:08:29, 16 bits, 1536kbps, 48khz
AC3 encoding doesn't support back channels. Will mix them into the surround.
Remapping channels...
Mixing surround channels...
Loading white noise (needed for dithering)...
Decoding with ArcSoft DTS Decoder...
Patching bitdepth to 24 bits...
Encoding AC3 <448kbps> with libAften...
Creating file "deutsch.ac3"...
-------------

btw: anyone tried hibernating windows during conversion? is it safe? managed to get good results with megui. here too?

[madshi]

I have a question for DTS-ES Matrix tracks. Those tracks have 5.1 discrete channels and one (BC) matrixed into SL and SR, right? An A/V receiver capable of DTS-ES decoding outputs 6.1, correct?
Yes.

Shouldn't eac3to do the same?
What specific purpose would you need that for? Anyway, I don't know how to do that.

Are Sonic or Arcsoft decoders able to decode DTS-ES Matrix?
Sonic no. ArcSoft don't know. I can ask ArcSoft to output 6.1 and it does it - even for real 5.1 tracks. But I don't know which algorithms ArcSoft is using for its channel transformations. So I feel safer asking ArcSoft to output the native format. Which is 5.1 for "6.1 matrix" tracks.

[nautilus7]

Is the BC matrixed into SL, SR the same way Center and Surround channels were matrixed into Left and Right channels in the analog Dolby ProLogic format (I think this was done with phase shifting)?
EDIT: the answer is yes to the above.

Basically, what i'm interested in is to be able to maintain 6.1 channels if i transcode a DTS-ES Matrix track to AC3 or whatever. Will i get a Dolby Digital EX 6.1 (matrixed) then? I suppose i won't because DTS and Dolby aren't the same...

[madshi]

Aren't DTS and Dolby Matrix algorithms similar to each other? I don't really know, to be honest.

[nautilus7]

I don't know as well. I am confused... Maybe tebasuna51 can answer that.

I will try making some DTS-ES tracks and then converting to DD-EX to see what happens.

[tebasuna51]

I don't know as well. I am confused... Maybe tebasuna51 can answer that.

I will try making some DTS-ES tracks and then converting to DD-EX to see what happens.
I don't know very much about 6.1. Only the samples I make with DTS-ES encoder:

If I call FL,FR,FC,LF,BL,BR,BC my source 6.1 channels to encode, and FL,FR,FC,LF,SL,SR the 5.1 channels after decode with ArcSoft or libav (1) seems:

SL = j x BL + 0.707 x BC
SR = -j x BR + 0.707 x BC

where j is +90º phase shift and -j is -90º phase shift.

(1) the output of libav seems identical to ArcSoft but -0.5 dB

I can ask ArcSoft to output 6.1 and it does it - even for real 5.1 tracks. But I don't know which algorithms ArcSoft is using for its channel transformations. So I feel safer asking ArcSoft to output the native format. Which is 5.1 for "6.1 matrix" tracks.
Maybe we can know something about the algorithms if we ask ArcSoft to output 6.1 using the samples from the DTS-ES encoder previously mentioned.

Would love to, but I don't know how to do it. Does anybody know any LGPL libraries or source code which can do high quality phase shifts? (if possible floating point in/out)
Only know the FooBar2000 plugins:

- foo_dsp_downmix (http://www.hydrogenaudio.org/forums/index.php?showtopic=52887): make the dpl II downmix with or without phase shift.

- foo_dsp_fsurround (http://www.hydrogenaudio.org/forums/index.php?showtopic=52235): recover the 5.1 channels from a stereo with dpl II

Both have the sources availables but need the libfftw3f-3.dll (Fast Fourier Transform library)

[3ngel]

Would love to, but I don't know how to do it. Does anybody know any LGPL libraries or source code which can do high quality phase shifts?

That would be just fantastic, so it coul be implemented the DLIIx algorithm wich uses a variable phase shift over the channels, and permits to correctly downmix 7.1 to 2.0.

[florinandrei]

Can't demux a file created by a Canon HF100 camcorder:

eac3to v2.68
command line: eac3to train.m2ts 1: video.mkv 2: audio.ac3
------------------------------------------------------------------------------
M2TS, 1 video track, 1 audio track, 0:00:25
1: h264/AVC, 1080i60 /1.001 (16:9)
2: AC3, 2.0 channels, 256kbps, 48khz, -67ms
[v01] Extracting video track number 1...
[a02] Extracting audio track number 2...
[a02] Applying (E-)AC3 delay...
[v01] Muxing video to Matroska...
Unfortunately the Haali Muxer cannot handle this source file.
It doesn't contain enough seek/recovery points.
Aborted at file position 32964608.

File is here, it's about 46 MB:

http://dl.getdropbox.com/u/29966/eac3to/train.m2ts


@florinandrei

eac3to train.m2ts 1: video.mkv -seekToIFrames 2: audio.ac3

Okay, that seems to pacify it. But there are some scary remarks about potential artifacts if I use this option. Under what conditions can this occur?

The source file is an honest-to-goodness AVCHD file generated by a Canon camcorder (http://www.usa.canon.com/consumer/controller?act=ModelInfoAct&fcategoryid=177&modelid=16187), it's probably clean and standards-compliant (the PS3 will happily play it, either directly or within a Blu-Ray structure), so I should not expect any surprises if I use -seekToIFrames - or should I?

[salehin]

Is it possible to edit/cut a ac3 stream using eac3to?

I need to cut about 0.5-1.0 sec of audio ( 6 channels AC3 Dolby audio at 384 kbps) to sync it with the BBC HD Video. Can you please advise the cmd line I should use.

The extra bit that I'm trying to get rid off is at the beginning.

Cheers :)

[odin24]

Is it possible to edit/cut a ac3 stream using eac3to?

I need to cut about 0.5-1.0 sec of audio ( 6 channels AC3 Dolby audio at 384 kbps) to sync it with the BBC HD Video. Can you please advise the cmd line I should use.

The extra bit that I'm trying to get rid off is at the beginning.

Cheers :)

You could apply a negative delay (audio starts early). I'm not sure of the command though, it might look something like this;

eac3to audio.ac3 -1000ms

[Darth Pinous]

Hi Madshi,

Don't know if someone as already reported on that, but doesn't seem so... Using latest versions of eac3to (2.6*), I encounter a problem when using the -down16 option on 24 bits soundtracks (WAV and DTS-HD, so far, not tested on trueHD or PCM).

downmixing to 16 bits is OK, but the audio is broken (weird sounds) as if the header was wrong. In fact, just passing the audio through wavewizard without any modifications fixes the problem...

Edit: just checked again with wavewizard.. the corrupt wav file says 16 bits per sample and 24 value bits per sample in its header. That may be the explanation...

[xkodi]

Sonic no. ArcSoft don't know. I can ask ArcSoft to output 6.1 and it does it - even for real 5.1 tracks. But I don't know which algorithms ArcSoft is using for its channel transformations. So I feel safer asking ArcSoft to output the native format. Which is 5.1 for "6.1 matrix" tracks.

madshi, please add temporary test switch that allow to ask Arcsoft to output 6.1 channels for DTS-ES, so i can run some tests.

i used the following test to find the parameters of different DTS-HD MA track before eac3to had support for such tracks:

http://forum.doom9.org/showthread.php?p=1147644#post1147644

i believe the same test could be used to find out what number of channels should be set for correct DTS-ES decoding with Arcsoft:

1. decode the DTS-ES 6.1 to WAVs with Sonic: outputs 5.1 channels

2. decode the DTS-ES 6.1 to WAVs with Arcsoft and set output channels to 5.1: outputs 5.1 channels

3. decode the DTS-ES 6.1 to WAVs with Arcsoft and set output channels to 6.1: outputs 6.1 channels

i suspect that if compare byte by byte the first 5.1 channels from 2 and 3 they will be different, then compare the first 5.1 channels between 1 and 2 and then between 1 and 3.

if the first 5.1 channels of 1 are byte by byte identical to the first 5.1 channels of 2 then Arcsoft output channels should be set to 5.1.

if the first 5.1 channels of 1 are byte by byte identical to the first 5.1 channels of 3 then Arcsoft output channels should be set to 6.1

[himan2001]

@madshi:

There are a lot more "old" BD Releases which contains seemless branching and "splitted" Releases that produce Audio-Spikes
and loud noise at the cutting/fixxing Points after timecode-rerun :(

Something is borked when using DTS, no matter if it´s Arcsoft or Sonic. With AC3-Track splitting/joining is ok.

As i can remember, the same BD´s using an older eac3to version < 2.58 doesn´t produce this errors.

A good try will be Conair/German BD. Try to assemble
the GERMAN DTS Track. On the cut point there is a loud
spike on the decoded file (destination format can be WAV or AC3 - result is the same)

Please check your audio/DTS Handling. Since the > 5.1 "fixxing" a lot of previous good working, older things, are screwed up more and more :(

The destorted Audio ONLY comes up when correcting overlaps!

On 2.67 Version a quickhelp was, not to use the HD-Part (export the -core and join). But since 2.69 this is no longer working. Now Core and HD/MA-Part produces drops&spikes.

When No other Track (AC3/TrueHD) is available I´am screwed on this overlapping BD´s...

[lchiu7]

I am trying to compress files captured from our terrestrial HD broadcasts which are h.264/ac-3.

The files play fine but when the video is compressed and then muxed back with the audio the various tools (e.g. txmuxer and mkvmerge) complain about crc errors in the AC-3 stream and suggest re-sync.

Tried to pass the AC-3 streams through delaycut in order to fix these. Tried all options including ignore, silence, repair but while it identifies the errors and completes okay, when this file is muxed back with the video, it goes out of sync; presumably after the point where the first CRC error had occurred.

The original files get rejected by eac3to saying run delaycut first. I would have thought silence or ignore would re-sync the streams but in my experience it doesn't.

Any advice would be appreciated here.

Thanks

[evdberg]

Have added that to my to do list. Don't know when I will get to that, though...

Thanks! Looking forward to it. Hope you can start with the bitdepth statistics per channel (interested to examine those Universal releases).

[tebasuna51]

I am trying to compress files captured from our terrestrial HD broadcasts which are h.264/ac-3.

The files play fine but when the video is compressed and then muxed back with the audio the various tools (e.g. txmuxer and mkvmerge) complain about crc errors in the AC-3 stream and suggest re-sync.

Tried to pass the AC-3 streams through delaycut in order to fix these. Tried all options including ignore, silence, repair but while it identifies the errors and completes okay, when this file is muxed back with the video, it goes out of sync; presumably after the point where the first CRC error had occurred.

The original files get rejected by eac3to saying run delaycut first. I would have thought silence or ignore would re-sync the streams but in my experience it doesn't.

Any advice would be appreciated here.

Thanks

The ac3 streams captured from TV can have 2.0 content (commercials), 5.1 (movie), 2.0 (more commercials), ...

These streams are invalid and I don't know the software to manage easily the problem.
You can use DelayCut to see the problems: 'Some basic parameters changed between Frame # and this frame'
and cut the stream to separate the contents 2.0 and 5.1 (most the times).

Warning also with videotools to cancel commercial sequences because don't take care with audio frames.

[Chumbo]

Since your captured streams are h.264, you'd have to use tools like TSPE, found at http://www.bitstreamtools.com, to cut the file, i.e., remove the commercials and leave only the 2.0 or 5.1 programming which will help keep things in sync.

[rebkell]

The ac3 streams captured from TV can have 2.0 content (commercials), 5.1 (movie), 2.0 (more commercials), ...

These streams are invalid and I don't know the software to manage easily the problem.
You can use DelayCut to see the problems: 'Some basic parameters changed between Frame # and this frame'
and cut the stream to separate the contents 2.0 and 5.1 (most the times).

Warning also with videotools to cancel commercial sequences because don't take care with audio frames.

these streams are invalid because they switch from 2.0 to 5.1 and back and forth?

What program are you using to extract the audio with? I've had plenty of sync problems and those are a pain to deal with, but try dgavcdec to extract the ac3 stream with, if it can be extracted I wouldn't expect any problems muxing it back in.

Sorry madshi, I know this is off topic.

[flyingernst]

I think I asked that a few weeks aga but I think I didn´t got an answer

EVO, 2 video tracks, 4 audio tracks, 8 subtitle tracks, 2:20:02
"Feature Presentation"
1: Joined EVO file
2: Chapters, 40 chapters with names
3: VC-1, 1080p24 /1.001 (16:9) with pulldown flags
4: VC-1, 480p30 /1.001 (3:2), 100ms
5: E-AC3, English, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB, 100ms
6: E-AC3, German, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB, 100ms
7: TrueHD, German, 5.1 channels, 48khz, dialnorm: -27dB, 200ms
8: E-AC3, 2.0 channels, 192kbps, 48khz, dialnorm: -27dB, 100ms
9: Subtitle, English
10: Subtitle, English, "SDH"
11: Subtitle, German
12: Subtitle, German, "SDH"
13: Subtitle, German, "narrative"
14: Subtitle
15: Subtitle
16: Subtitle
[a05] Extracting audio track number 5...
[a05] Removing E-AC3 dialog normalization...
[a05] Applying (E-)AC3 delay...
[a05] Creating file "E:\english.eac3"...
Video track 3 contains 201466 frames.
Video track 4 contains 251744 frames.
eac3to processing took 6 minutes, 4 seconds.
Done.

will I have to insert a 100ms delay while muxing to MKV or M2TS?!

Thanks, Greetings Michael

whats here, here Delay while muxing?!

EVO, 2 video tracks, 4 audio tracks, 8 subtitle tracks, 2:20:02
"Feature Presentation"
1: Joined EVO file
2: Chapters, 40 chapters with names
3: VC-1, 1080p24 /1.001 (16:9) with pulldown flags
4: VC-1, 480p30 /1.001 (3:2), 100ms
5: E-AC3, English, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB, 100ms
6: E-AC3, German, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB, 100ms
7: TrueHD, German, 5.1 channels, 48khz, dialnorm: -27dB, 200ms
8: E-AC3, 2.0 channels, 192kbps, 48khz, dialnorm: -27dB, 100ms
9: Subtitle, English
10: Subtitle, English, "SDH"
11: Subtitle, German
12: Subtitle, German, "SDH"
13: Subtitle, German, "narrative"
14: Subtitle
15: Subtitle
16: Subtitle
[a07] Extracting audio track number 7...
[a07] Removing TrueHD dialog normalization...
[a07] Creating file "E:\deutsch.thd"...
Video track 3 contains 201466 frames.
Video track 4 contains 251744 frames.
eac3to processing took 6 minutes, 12 seconds.
Done.


and perhaps here not right: Why can´t I mux the *-thd (also renamed to *.ac3) mux with TSmuxer to a m2ts file, the EAC3 File works with the VC1 Videofile

[flyingernst]

perhaps we can forget my quetion when you tell me how eac3to can directly change the EVO container in a m2ts container (wiki says nothing about).
ic can work with full disk Structures an the 2 mainEvos...but what comes after

"F:\Download\Brennen\eac3to\eac3to.exe" "P:" 1)

(1) ist the mainmovie
since TSmuxer has problems with importing thd files I want to change container and throw out not wanted streams with tsremuxer


Thanks

EDIT: have it: joining the files to one EVO with eac3to, then change container with TSremux

[Blue_MiSfit]

http://www.mediafire.com/?sharekey=32d600f0ab3cb846d2db6fb9a8902bda

Here's a strange little stream - eac3to, mkvtoolnix, and several other programs absolutely hate it, but VLC and Media Player Classic (with ac3filter) will play it perfectly.

It's 640kbps AC3 5.1ch.

Help???

~MiSfit

[PHD_1976]

I was using 2.68 version to split DD 5.1 EX file from DVD to 6 mono waves. I don't have nero 7.xx installed on my machine so the track was processed by ffmpeg decoder. The result was not good - too many peak level overflows so the audible crack noise is heard when the overflow happens.
I've never had problems with DD tracks+ffmpeg decoder.
Maybe this is connected with EX or I'm just unaware of some option to normalize peaks?

I've tried the very same procedure on other machine where Nero 7.10 is installed - the result was fine: clear 6 channels with no overflow at all.

Madshi, if you want a sample i'll provide one.

Thanx for your work.

[tebasuna51]

http://www.mediafire.com/?sharekey=32d600f0ab3cb846d2db6fb9a8902bda

Here's a strange little stream - eac3to, mkvtoolnix, and several other programs absolutely hate it, but VLC and Media Player Classic (with ac3filter) will play it perfectly.

It's 640kbps AC3 5.1ch.

Help???

~MiSfit

You can use DelayCut to see how many corrupt is this file.

The best 'player' for this file is the 'Recycle Bin', just drag and drop ... :p

[Blue_MiSfit]

DelayCut does indeed vomit out a waterfall of errors, but it plays perfectly in MPC or VLC ???

~MiSfit

[madshi]

I don't know very much about 6.1. Only the samples I make with DTS-ES encoder:

If I call FL,FR,FC,LF,BL,BR,BC my source 6.1 channels to encode, and FL,FR,FC,LF,SL,SR the 5.1 channels after decode with ArcSoft or libav (1) seems:

SL = j x BL + 0.707 x BC
SR = -j x BR + 0.707 x BC

where j is +90º phase shift and -j is -90º phase shift.
Hmmmm, that's interesting. So would you say that 7.1 downmixed to 5.1 would likely be:

SL = j x SL + BL
SR = -j x SR + BR

?

Maybe we can know something about the algorithms if we ask ArcSoft to output 6.1 using the samples from the DTS-ES encoder previously mentioned.
I'll add a switch which will allow you to ask ArcSoft to output any number of channels you want. That should open the door to check what ArcSoft is doing for 7.1 -> 6.1 and 7.1 -> 5.1, too.

Only know the FooBar2000 plugins:

- foo_dsp_downmix (http://www.hydrogenaudio.org/forums/index.php?showtopic=52887): make the dpl II downmix with or without phase shift.

- foo_dsp_fsurround (http://www.hydrogenaudio.org/forums/index.php?showtopic=52235): recover the 5.1 channels from a stereo with dpl II

Both have the sources availables but need the libfftw3f-3.dll (Fast Fourier Transform library)
Looks nice. Unfortunately they're both GPL. Since eac3to is closed source, I can't use that code. Would need LGPL... :(

Okay, that seems to pacify it. But there are some scary remarks about potential artifacts if I use this option. Under what conditions can this occur?
As long as you don't seek there shouldn't be any problems. If you seek, there may be artifacts visible for a short time (probably not much longer than a few seconds) - or not. I'd suggest to simply give it a try...

Don't know if someone as already reported on that, but doesn't seem so... Using latest versions of eac3to (2.6*), I encounter a problem when using the -down16 option on 24 bits soundtracks (WAV and DTS-HD, so far, not tested on trueHD or PCM).

downmixing to 16 bits is OK, but the audio is broken (weird sounds) as if the header was wrong. In fact, just passing the audio through wavewizard without any modifications fixes the problem...

Edit: just checked again with wavewizard.. the corrupt wav file says 16 bits per sample and 24 value bits per sample in its header. That may be the explanation...
Can you upload a little sample? What you report sounds a bit strange to me. It should be 24 bits per sample data structure and 16 valid bits per sample. Maybe wavewizard just names the fields in a strange way. But yes, newer eac3to versions write different wav files compared to older eac3to versions and that may confuse some applications or DirectShow filters. But as far as I can see, eac3to's WAV files are perfectly fine. It's not my fault if other software can't handle them. You can ask eac3to to create stupid simple WAV files by using the "-simple" option. Also if your WAV file is 16/24, you can use "eac3to source.wav dest.wav". That will then automatically strip the zero bytes and give you a smaller WAV file which is straight 16bit.

There are a lot more "old" BD Releases which contains seemless branching and "splitted" Releases that produce Audio-Spikes and loud noise at the cutting/fixxing Points after timecode-rerun :(

Something is borked when using DTS, no matter if it´s Arcsoft or Sonic. With AC3-Track splitting/joining is ok.

As i can remember, the same BD´s using an older eac3to version < 2.58 doesn´t produce this errors.

A good try will be Conair/German BD. Try to assemble
the GERMAN DTS Track. On the cut point there is a loud
spike on the decoded file (destination format can be WAV or AC3 - result is the same)

Please check your audio/DTS Handling. Since the > 5.1 "fixxing" a lot of previous good working, older things, are screwed up more and more :(

The destorted Audio ONLY comes up when correcting overlaps!

On 2.67 Version a quickhelp was, not to use the HD-Part (export the -core and join). But since 2.69 this is no longer working. Now Core and HD/MA-Part produces drops&spikes.

When No other Track (AC3/TrueHD) is available I´am screwed on this overlapping BD´s...
I guess you haven't compared v2.67 with v2.69 with the same title, have you? I don't think I've changed anything. I think the latest version should behave similar to much older versions.

I'll check out Con Air, anyway.

Tried to pass the AC-3 streams through delaycut in order to fix these. Tried all options including ignore, silence, repair but while it identifies the errors and completes okay, when this file is muxed back with the video, it goes out of sync; presumably after the point where the first CRC error had occurred.
delaycut can only do so much. If the audio stream is broken too much, audio sync may be lost. There's nothing we can do about that.

I think I asked that a few weeks aga but I think I didn´t got an answer
That's probably because the very same question has already been asked & answered about a million times in this thread... ;)

As long as you use eac3to to demux or remux or transcode the audio, you don't have to care about audio delays. eac3to will do everything automatically for you.

I was using 2.68 version to split DD 5.1 EX file from DVD to 6 mono waves. I don't have nero 7.xx installed on my machine so the track was processed by ffmpeg decoder. The result was not good - too many peak level overflows so the audible crack noise is heard when the overflow happens.
I've never had problems with DD tracks+ffmpeg decoder.
Maybe this is connected with EX or I'm just unaware of some option to normalize peaks?

I've tried the very same procedure on other machine where Nero 7.10 is installed - the result was fine: clear 6 channels with no overflow at all.

Madshi, if you want a sample i'll provide one.
Yes, a sample would be welcome. I'd forward it to the libav AC3 decoder maintainer.

DelayCut does indeed vomit out a waterfall of errors, but it plays perfectly in MPC or VLC ???
MPC and VLC just skip over broken data. That's why they're able to play these files. If you let delaycut run over the track, the end result should be similar to what MPC and VLC are doing...

[xkodi]

don't know if this is valuable information, just found it while looking with hex editor in DTSHDDec.dll from WinDVD9:

-dn/m instructs the decoder to perform the down-mix to the desired channel configuration e.g,:

-d1 - down-mix to mono
-d2 or -d2/o - down-mix to Lo/Ro
-d2/t - down-mix to Lt/Rt
-d2/1 - down-mix to L-R-Cs
-d2/1/.1 - down-mix to L-R-Cs-LFE
-d3 - down-mix to L-R-C
-d3/0/.1 - down-mix to L-R-C-LFE
-d2/2 - down-mix to L-R-Ls-Rs

-lx defines the mixing of the LFE data into the Front Channels of the Downmix where x is the flag 0 or 1 st.
0 -> LFE data is ignored; 1 -> LFE is mixed with 10dB boost.

It has effect only in the cases where -dm/n/.0 indicates no LFE output channel:
- for the case -d1 (-d1/0/.0) -l0 the Co=Co
- for the case -d1 (-d1/0/.0) -l1 the Co=0.5*(Co+3.16*LFE)
- for the case -d2 (-d2/0/.0) -l1 the Lo=0.5(*Lo+0.7078*3.16*LFE) and Ro=0.5*(Ro+0.7078*3.16*LFE)
- for the case -d3 (-d3/0/.0) -l1 the Lo=0.5*(Lo+0.7078*3.16*LFE), Ro=0.5*(Ro+0.7078*3.16*LFE) and Co=0.5*Co
- for the case -d2/2 (-d2/2/.0) -l1 the Lo=0.5*(Lo+0.7078*3.16*LFE), Ro=0.5*(Ro+0.7078*3.16*LFE), Lso=0.5*Lso and Rso=0.5*Rso

[PHD_1976]

Madshi,
Here's the 5 minutes sample of DD 5.1 EX with peak overflow when processed with ffmpeg to 6 mono wavs.
sample (http://www.sendspace.com/file/pn8z40)

Thanx.

[madshi]

don't know if this is valuable information, just found it while looking with hex editor in DTSHDDec.dll from WinDVD9:
It is interesting in that way that it shows that the LFE channel is boosted by 10db for downmixing. eac3to currently doesn't do that. I guess I should add the 10db boost?

[madshi]

Here's the 5 minutes sample of DD 5.1 EX with peak overflow when processed with ffmpeg to 6 mono wavs.
Hmmmm... In Audacity I don't see that much of a difference between libav and Nero. Both have very high peaks. It might be that with libav the peak overflows over so slightly and with Nero not. But the difference must be really small.

Don't know what to say/do about this now. Maybe adding "gain" functionality to eac3to would make sense...

[PHD_1976]

Well, it's not that I'm complaining of course, just wanted you to know about this issue.
With Nero decoder peaks are also high, you're right, but no audible crackling noise.

I know that besweet decoder performs downmix when overflow happens, maybe you would like to do it the same way?

Anyway, thank you for your time and sample checking :-)

[Zwitterion]

It is interesting in that way that it shows that the LFE channel is boosted by 10db for downmixing. eac3to currently doesn't do that. I guess I should add the 10db boost?
Does eac3to actually downmix the LFE channel using -down2? Dolby recommends to just discard it (phase issues, overdrive, you don't hear it anyway...).

[madshi]

Well, it's not that I'm complaining of course, just wanted you to know about this issue.
With Nero decoder peaks are also high, you're right, but no audible crackling noise.
Thanks for insisting... ;) Cause that made me find a bug in eac3to, which didn't handle peak clipping correctly, when using libav for decoding. This will be fixed in the next build. The crackling problems should be gone then...

Does eac3to actually downmix the LFE channel using -down2? Dolby recommends to just discard it (phase issues, overdrive, you don't hear it anyway...).
The LFE is not mixed in by default, but you can ask eac3to to do that by using the "-mixlfe" option.

[bmnot]

Instead of having to pick a version of a seamless branching movie, and then mux that single version to mkv, would it be possible to make a chapter/playlist for the separate files, convert all to mkv, then then add them all into one mkv and be able to pick the version you want to watch? I know there would be some issues with audio gaps, but it might be worth it instead of having to make a separate mkv for each version of the movie. I've seen it done for some movies.

[madshi]

Instead of having to pick a version of a seamless branching movie, and then mux that single version to mkv, would it be possible to make a chapter/playlist for the separate files, convert all to mkv, then then add them all into one mkv and be able to pick the version you want to watch? I know there would be some issues with audio gaps, but it might be worth it instead of having to make a separate mkv for each version of the movie. I've seen it done for some movies.
I don't think MKV supports such a kind of thing in any reasonable way. Sure, one could mux a monster movie containing all the fragments of all branchings, but then you'd probably have to skip chapters manually. I don't think there's any kind of "playlist" for chapters in MKV. And even if there was, the jumps from one chapter to another would surely not be seamless. So I don't think that this is a good idea. I understand why you want this, but you'll have to keep using the Blu-Ray structure for seamless branching movies, if you want to keep the seamless branching alive.

[kurt]

I have another question about these gaps:

Do I have to rerun the exact same command line for seamless branching titles or is it only necassary for audio parts?

i.e:

1. eac3to 2) 2: video.vc1 4: english.ac3
2. eac3to 2) 4: english.ac3

that could save some time, right?

[madshi]

I have another question about these gaps:

Do I have to rerun the exact same command line for seamless branching titles or is it only necassary for audio parts?

i.e:

1. eac3to 2) 2: video.vc1 4: english.ac3
2. eac3to 2) 4: english.ac3

that could save some time, right?
Correct, you only need to rerun those tracks which you want to have gap/overlap-fixed. It's just easier to rerun the same command line. Also it would be difficult to explain what you need to do exactly to save space. That's why eac3to simply asks to rerun the same command line.

[piratburner]

Is it possible to extract movie and sound of a BD and make a m2ts file of this files ( perhaps another prog.)
I working on Million Dollar Baby 2004 1080p Blu-ray GER AVC DTS-HD HR 7.1
On TsMuxeR I have no sound after I have remuxed with movie and English DTS-HD 7.1 sound

How should I do it with eac3to ?

[lexor]

Is it possible to extract movie and sound of a BD and make a m2ts file of this files ( perhaps another prog.)
I working on Million Dollar Baby 2004 1080p Blu-ray GER AVC DTS-HD HR 7.1
On TsMuxeR I have no sound after I have remuxed with movie and English DTS-HD 7.1 sound

How should I do it with eac3to ?

eac3to does not produce m2ts files. it's a demuxing tool (it only stores video into mkv if you tell it to). So your best bet is to use eac3to to convet that DTS-HD into something TSmuxer can work with. Ask the guys in TSmuxer thread regarding which formats that tool is good for.