eac3to - audio conversion tool

[LigH]

Depending on your purpose, you may want to support '.wavs' too (demultiplexes/decodes and de-interleaves a multi-channel audio stream to separate mono .wav files per audio channel).

VobSub consists of .sub (RLE-encoded DVD subpicture stream) + .idx (timecode and stream information text file).

[tebasuna51]

It's strange, works same :confused:

MPEG2 video (http://ul.to/78qidjok) -> .m2v and .mpeg2

Why ?
Because the extension is not mandatory for some formats.
The format of the file is defined in the file header.

For instance you can use .dts for all of this:
** DTS = .dts
** DTS Master Audio = .dts
** DTS Hi-Res = .dts
** DTS Express = .dts
The use of other extension is optional and only to remind you the internal format, but there are soft than only accept .dts

To extract the core of a dtshd you need add the -core parameter, if you put only the extension .dts, without -core, you obtain a dtshd.

BTW, DTS-HD (no core) don't exist, a dtshd must have always a core.

- .pcm and .raw have the same audio data but with different byte and channel order. A .raw is exact to a .wav/.w64/.rf64 without header.

- Graphic subtitles.
Only (PGS) = .sup is usefull

With (DVD), subs from VOB container, the .sup extracted is not supported by mkvmerge to mux or subrip/suprip/subtitleedit to convert to text subtitles.
If you want convert to text subs better use subrip/subtitleedit directly from VOB files.
BTW the .sup extracted by eac3to (or PgcDemux) can be used, with the .IFO file from DVD structure, by BDsup2sub to obtain the graphic subtitles sub/idx accepted by MkvMerge. But there are other tools to extract sub/idx directly from VOB's

The (VobSub), subs sub/idx in mkv container, aren't supported by eac3to to extract. Use MkvExtract instead.

EDIT:
** DTSWAV = .dtswav
** AC3WAV = .ac3wav
Aren't supported like output by eac3to.
Some times (not always) are supported like input to extract the .dts/.ac3 included in the fake wav container.

[filler56789]

...

BTW, DTS-HD (no core) don't exist, a dtshd must have always a core.

...

Only if one wants/needs compliance with the Blu-Ray specifications. In the so-called "Digital Delivery" mode, the Master Audio Encoder Suite can output a pure lossless stream, but only for certain channel layouts (2.0, 5.1 and 7.1).

[nevcairiel]

Even on Blu-ray you have DTS-HD streams without core, the so called DTS Express uses a low-bitrate HD stream without core to be able to produce a VBR stream with low bitrate for mostly voice (commentary tracks)

[frencher]

OK Thank's all :)
If update or error(s) send me update ;)

## Video
** VC-1 = .vc1
** h264/AVC = .h264
** MPEG1 = .mpeg1
** MPEG2 = .m2v
** MS/VFW/FOURCC/XVID = .xvid
## Audio
** AC3 = .ac3
** E-AC3 = .eac3
** AC3 Surround = .ac3
** TrueHD = .thd
** TrueHD/AC3 = .thd+ac3
** DTS = .dts
** DTS Master Audio = .dts
** DTS Hi-Res = .dts
** DTS-HD (no core) = .dts
** DTS Express = .dts
** FLAC = .flac
** RAW/PCM = .pcm
** MLP = .mlp
** WAV = .wav
** W64 = .w64
** RF64 = .rf64
** DTSWAV = .dtswav
** AC3WAV = .ac3wav
** AAC = .aac
** MP1 = .mp1
** MP2 = .mp2
** MP3 = .mp3
## Subtitles
** (PGS) = .sup
** (DVD) = .sup
** (SRT) = .srt
** (ASS) = .ass
** (SSA) = .ssa
** (VobSub) = .sup

[tebasuna51]

Only if one wants/needs compliance with the Blu-Ray specifications. In the so-called "Digital Delivery" mode, the Master Audio Encoder Suite can output a pure lossless stream, but only for certain channel layouts (2.0, 5.1 and 7.1).
Like I never see this kind of files please confirm:
- If this stream can be muxed in any container and extracted by eac3to.
- If eac3to can decode this stream.

Even on Blu-ray you have DTS-HD streams without core, the so called DTS Express uses a low-bitrate HD stream without core to be able to produce a VBR stream with low bitrate for mostly voice (commentary tracks)
OK, if we can call DTS Express like DTS High Definition.
BTW, DTS Express was already listed.

OK Thank's all :)
If update or error(s) send me update ;)

** MPEG1 = .mpeg1
I don't know, never tested.

** MS/VFW/FOURCC/XVID =
eac3to can't open avi container and can't extract this stream if present in mkv container.

** AC3 Surround = .ac3
Is a particular case of standard AC3, if you list this item separately seems you suggest than eac3to can apply a DPL decoder and is not true.

** DTS-HD (no core) = .dts
I don't know, never tested.

** RAW/PCM = .pcm/.raw
Both formats useless, must be extracted like .wav/.w64/.rf64

** DTSWAV = .dts
eac3to can't output .dtswav, only can convert DTSWAV to .dts

** AC3WAV = .ac3
eac3to can't output .ac3wav, only can convert AC3WAV to .ac3

** MP1 = .mp1
I don't know, never tested.

** (VobSub) =
eac3to can't extract this kind of sub.

[filler56789]

Like I never see this kind of files please confirm:
- If this stream can be muxed in any container and extracted by eac3to.
No.
- If eac3to can decode this stream.
No, and I serious doubt any DirectShow decoder available a.t.m. can do that as well.

BTW, neither mkvmerge nor LAV Splitter support A_DTS/LOSSLESS thus far.

Besides,

http://matroska.org/technical/specs/codecid/index.html says:

Please, when reading through this list, always keep in mind that the intention behind it is NOT to list all existing audio and video codecs, but more to list those codecs that are currently supported in matroska (or will be supported soon),

so the problem of course, is what they mean by "soon" :)

http://lists.matroska.org/pipermail/matroska-devel/2011-June/004051.html

[IanD]

I recently tried converting an HD-DVD with TrueHD audio to Bluray format using first ClownBD 0.81 and then eac3to 3.27, but each software insists on decoding and re-encoding the AC3 core of the TrueHD track.

Does anyone know why this might be occurring and how to stop it happening?

[DarkSpace]

I recently tried converting an HD-DVD with TrueHD audio to Bluray format using first ClownBD 0.81 and then eac3to 3.27, but each software insists on decoding and re-encoding the AC3 core of the TrueHD track.

Does anyone know why this might be occurring and how to stop it happening?

From what I remember, TrueHD tracks on HD-DVD don't have an AC3 core, so eac3to will have to generate one from the decoded TrueHD track for the BluRay version. There is no "re-encoding" of the AC3 core, eac3to is just generating data that wasn't there previously.

[IanD]

From what I remember, TrueHD tracks on HD-DVD don't have an AC3 core, so eac3to will have to generate one from the decoded TrueHD track for the BluRay version. There is no "re-encoding" of the AC3 core, eac3to is just generating data that wasn't there previously.

Thanks for the clarification: I didn't realise that was the case.

A number of HD-DVD have better video and/or audio than the equivalent Bluray, still, and since I no longer have an HD-DVD player it's good to be able to get them into a playable format again.

[yesgrey]

I wanted to start using flaccl instead of flac inside eac3to, because it will save me a lot of time, but I want to be sure that it gives accurate results, hence why I wanted to use eac3to to convert a flaccl generated file to a flac eac3to flac, to compare with the original one created with eac3to, which I consider as the reference.

The tests I did using graphstudio and madflac to output pcm files were not very good to flaccl, because not only the files are different than the eac3to generated one, but they are also different between them if I compare two different compression levels. It seems flaccl is messing the data in someway...
Just to let you know that the latest flaccl version fixed the problems with the multiple channels files. Now eac3to can read the files and I have checked with eac3to to decode and then re-encode the flaccl created files and they were exactly the same as the ones created directly with eac3to.

Now I can remux my movies much faster... :)

[rapscallion]

Video frames have a duration of 40 ms (25 fps) then delays of +/- 20 ms are absolutely undetectable.
From an old post, however I couldn't find the answer to this:
What is the video frame duration @23.976 fps?

Next question. I extracted a Dolby THD track to wavs, where the track has a video delay of -32ms.

The log shows (note lines 2 and 7):


[a03] Extracting audio track number 3...
[a03] This track begins with a non-major frame. <WARNING>
[a03] Extracting TrueHD stream...
[a03] Removing TrueHD dialog normalization...
[a03] Decoding with DirectShow (Nero Audio Decoder 2)...
[a03] DirectShow reports 5.1 channels, 24 bits, 48kHz
[a03] Applying RAW/PCM delay...
[a03] Writing WAVs...

I then converted the wavs to DTS-MA.

So, does the "applying RAW/PCM delay..." mean that I don't have to apply any delay when muxing the DTS track back with the video?

[tebasuna51]

What is the video frame duration @23.976 fps?
1000 ms / 23.976 = 41.7 ms

So, does the "applying RAW/PCM delay..." mean that I don't have to apply any delay when muxing the DTS track back with the video?
Correct.

[rapscallion]

@tebasuna51

As always, thank you for the quick and knowledgeable response.

I'm always a bit confused by audio/video delays.
Also, this is the first time I've seen "This track begins with a non-major frame. <WARNING>"

[tebasuna51]

Also, this is the first time I've seen "This track begins with a non-major frame. <WARNING>"
I can't help you with this.
I don't know TrueHD specs.

[rapscallion]

A TrueHD stream is composed of Major sync frames and non-major frames. Apparently tracks are supposed to start with a major frame.

This may have accounted for the -32ms delay and the warning.

[DoctorM]

Love how useful this tool is. I have 2 feature requests:

1) A SoundTouch type Tempo adjustment. The ability to speed up and a slow down a track without any change to the pitch is sometimes preferable.

2) Dolby Pro Logic II/IIx/IIz Upmix. I often want to expand the number of tracks on a stereo source to 5.1 or 6.1. Manually doing this is labor intensive and requires multiple programs and in between conversions. Be nice to just: Eac3to StereoTrack.wav SurroundTrack.ac3 -upDpl6

[filler56789]

Love how useful this tool is. I have 2 feature requests:

1) A SoundTouch type Tempo adjustment. The ability to speed up and a slow down a track without any change to the pitch is sometimes preferable.

2) Dolby Pro Logic II/IIx/IIz Upmix. I often want to expand the number of tracks on a stereo source to 5.1 or 6.1. Manually doing this is labor intensive and requires multiple programs and in between conversions. Be nice to just: Eac3to StereoTrack.wav SurroundTrack.ac3 -upDpl6

Those features are non-trivial to implement I.M.H.O., BUT if you donate some hundreds or thousands of $$$ to madshi,
then possibly he'll do that for you :)

[Nico8583]

Hi :)
Any news about fix for left view / right view with MVC flag ?
This case is open in Mantis : http://bugs.madshi.net/view.php?id=86
Thanks !

[DoctorM]

Thanks, filler56789, float me a loan? ;-)

[frencher]

Like I never see this kind of files please confirm:
- If this stream can be muxed in any container and extracted by eac3to.
- If eac3to can decode this stream.


OK, if we can call DTS Express like DTS High Definition.
BTW, DTS Express was already listed.



** MPEG1 = .mpeg1
I don't know, never tested.

** MS/VFW/FOURCC/XVID =
eac3to can't open avi container and can't extract this stream if present in mkv container.

** AC3 Surround = .ac3
Is a particular case of standard AC3, if you list this item separately seems you suggest than eac3to can apply a DPL decoder and is not true.

** DTS-HD (no core) = .dts
I don't know, never tested.

** RAW/PCM = .pcm/.raw
Both formats useless, must be extracted like .wav/.w64/.rf64

** DTSWAV = .dts
eac3to can't output .dtswav, only can convert DTSWAV to .dts

** AC3WAV = .ac3
eac3to can't output .ac3wav, only can convert AC3WAV to .ac3

** MP1 = .mp1
I don't know, never tested.

** (VobSub) =
eac3to can't extract this kind of sub.

OK thanks tebasuna51 ;)

[frencher]

Hi :)
Any news about fix for left view / right view with MVC flag ?
This case is open in Mantis : http://bugs.madshi.net/view.php?id=86
Thanks !

Temporary While waiting for a update eac3to
Small tool for detect AVC/MVC <=> MVC/AVC order (http://forum.doom9.org/showthread.php?p=1635457#post1635457)

[maxmercy]

EDIT - Careful searching found that for any 6.1 format the remap must be done to accomplish -double7 processing correctly.

Hey everyone,

I have a 6.1 channel .w64 file with L C R LFE BC SL SR channel mapping, and I need to upmix to a 7.1 .w64 file for proper muxing with TSMuxer (TSM does not like 6.1, and converts it to 7.0, putting my LFE channel in one of my rear surrounds). Can I simply use the -double7 feature, or do I need to remap channels as I have found when searching for how to use the '-double7' feature?

I was wanting to just use:

eac3to input.w64 output.w64 -double7

Will this work?

Thanks in advance,

JSS

[tebasuna51]

After my test seems you need:

eac3to input.w64 output.w64 -0,1,2,3,5,6,4 -double7

[maxmercy]

Thanks! It worked great!

JSS

[Overdrive80]

Would be possible implement option for output wav on 32 bit float??

[Bigmango]

Could we please get a replacement of the FLAC encoder with the one in ffmpeg 2.0 with "-compression_level 12" ?

It now properly supports 24bit. I have tested up to 24bit 7.1 channel FLAC and it plays fine on all the players I have tried (XBMC, MPC-HC, etc...).

"-compression_level 12" allows to shave off another 100 to 200Mb on my MKV backups.

It's a little cumbersome to compress to FLAC with eac3to and then to recompress it again to FLAC with compression level 12 with ffmpeg. It would be nice if we could do it in one go.

Thanks.

[Boulder]

Can't you utilise piping to ffmpeg?

[Bigmango]

Can't you utilise piping to ffmpeg?

I am using the HD-DVD/Blu-Ray Stream Extractor (the MeGUI version as it allows to queue tasks - this way I can set up 20 or more MKV to FLAC taks in the queue and have it run in one go), so I don't think it's possible to setup piping with the gui.

Thanks for the feedback.

[Boulder]

I tested simple piping to ffmpeg and it doesn't seem to work properly :/ The resulting file is only a part of the source, maybe it's due to some filesize restrictions. I couldn't find a way to force ffmpeg to ignore any audio length. The documentation of that application is just not very clear (I've ran into that many times in the past), but some googling told me that ffmpeg should support large files without a problem. There is something funny there anyway, I also tried decoding to wav and then using ffmpeg on the result but it didn't help either.

[gnol009]

Anyone know how to creat qpfile.txt?

[filler56789]

Anyone know how to creat qpfile.txt?

Looks like you have posted in the wrong forum AND in the wrong thread :)

http://forum.doom9.org/forumdisplay.php?f=77

[jpsdr]

I have a little question...
If i do something like : "eac3to 00001.m2ts 3: c:\File.wavs"
Track 3 is DTS-HD MA, either 2.0, 5.1 or 7.1. I have the decoderdtsdll.dll installed.
Will the wave files be created only from the DTS core (so from lossy audio), or will they be created from the full DTS-HD MA information (the lossless audio) ?
Thanks.

[rapscallion]

I Use a GUI with eac3to and do the same kind of extractions and always get the full lossless wavs.

[Sparktank]

It will extract the core if you tell it to with "-core" switch.
Otherwise, if Arcsoft decoder registered, then it will extract the lossless DTS.

[DoctorM]

I oddly find that when I'm decoding AC3 I have to force the switch -Nero. Yes, Nero 7 is properly installed, eac3to -test KNOWS it is installed (and says it works fine) and it seems to work properly when I add -Nero to my line. For some reason it doesn't seem to be the default decoder though.

Any thoughts on this?

[nevcairiel]

I oddly find that when I'm decoding AC3 I have to force the switch -Nero. Yes, Nero 7 is properly installed, eac3to -test KNOWS it is installed (and says it works fine) and it seems to work properly when I add -Nero to my line. For some reason it doesn't seem to be the default decoder though.

Any thoughts on this?

The default decoder for AC3 is ffmpeg/libav, so that behaviour is expected (as documented in the first post in this thread)
I don't think the Nero decoder does a better job at it, so the default choice here makes sense, IMHO.

[DoctorM]

Thank you. I misread that. Nero is best for AAC not AC3. In fact nothing is listed as 'best for ac3". I guess I just assumed.

[Sparktank]

Thank you. I misread that. Nero is best for AAC not AC3. In fact nothing is listed as 'best for ac3". I guess I just assumed.

The Wikibooks (http://en.wikibooks.org/wiki/Eac3to/How_to_Use#Audio_Decoders) says that Nero should offer better sound quality over libav/ffmpeg.
As a result Nero7 is the default decoder for AC3 format.
Outdated since version 3.25...
* libav is now the default decoder for AC3 and E-AC3

But shortly before release of version 3.26, someone asked about this and Madshi said there was a problem where Nero didn't always remove DRC for some files.

Curious, why the change from Nero to libav as the preferred DD decoder?Because there were some problems with the Nero decoder. Mostly, disabling DRC didn't work for all tracks.

[Overdrive80]

The Wikibooks (http://en.wikibooks.org/wiki/Eac3to/How_to_Use#Audio_Decoders) says that Nero should offer better sound quality over libav/ffmpeg.

Outdated since version 3.25...


But shortly before release of version 3.26, someone asked about this and Madshi said there was a problem where Nero didn't always remove DRC for some files.

Somebody users reports issues decoding ac3 files with nero decoder, mainly with multitracks files (5.1). For this reason, libav is decoder for default.

This Wikibooks is outdated.

[Mark Regalo]

What files are used by eac3to from Nero to decode aac files? I have an old Nero 7 cd which I bought a long time ago yet I really don't want to install Nero on my computer.

[Snowknight26]

Then why do you want to use Nero to decode AAC?

[Mark Regalo]

Then why do you want to use Nero to decode AAC?

Isn't that what eac3to uses as the "best" method to decode aac? I just read what madshi wrote about eac3to in the OP.

[Mark Regalo]

So, any idea on how to use Nero with eac3to without installing it?

[Sparktank]

So, any idea on how to use Nero with eac3to without installing it?

There was a batch script created a long time ago that would do it for you.

Try a google search and limit to this forum only.
site:forum.doom9.org something something something eac3to

[utenteanonimo64]

Is there any software that can take the information created by eac3to in "chapters.txt" and apply it to a WAV/W64 file to split it into separate files with the appropriate WAV/W64 headers for each chapter?

[Shevek]

Is there any software that can take the information created by eac3to in "chapters.txt" and apply it to a WAV/W64 file to split it into separate files with the appropriate WAV/W64 headers for each chapter?

Perhaps Medieval CUE Splitter?

[utenteanonimo64]

Perhaps Medieval CUE Splitter?

Thanks for the quick reply. I will have to look if there is an easy way to turn "chapters.txt" into a CUE format file...

[kevmitch]

I noticed that decoding the same surround source file with the -downmix option produced different output files each time (the samples weren't the same). I'm pretty sure that this is because there is dithering which is randomly chosen. Sure enough, when I specify "-full" as suggested by madshi (http://forum.doom9.org/showthread.php?p=1401084#post1401084), the output is identical every time.

My question is how exactly is the random noise seeded/generated and why not just use a method (i.e., same seed) that produces identical results each time? I know it would put my mind at ease a little.

[kevmitch]

While waiting to be able to post here for the first time, I posted over at avsvorums (http://www.avsforum.com/t/1482702/oblivion-blu-ray-official-avsforum-review/150#post_23626101) about a problem I noticed trying to decode the DTS HD MA sound from the Obilivion blu ray. In short, the center channel becomes distorted in loud bits. Since that posting, I have realised that specifying '-core' option to eac3to fixes the problem but of course its not "lossless", which my decode of the MA track clearly isn't either in this case.

I have uploaded a sample (https://dl.dropboxusercontent.com/u/60598588/oblivion_distort.zip) .dtshd file from 8:10-8:20 of the film which exhibits the problem. Also included with the sample are my eac3to decodes of the centre channel with and without "-core" as well as the libav decode, which also exhibits the distiortion. And of course all the logs and eac3to -test results.

I get the same results (bit-for-bit) using eac3to with the Arcsoft decoder versions 1.1.0.0, 1.1.0.7 and 1.1.0.8. I also hear the distortion when decoding the .dtshd directly in mplayer, vlc and audacity (which I think use the same code as libav anyway).

EDIT: On second thought, the distortion is there in core track as well. I don't know how I didn't hear that before.