eac3to - audio conversion tool

[bmnot]

Can anyone confirm that Nero Directshow AAC Decoder filter produce the same output as neroAacDec from NeroDigitalAudio.zip (http://www.nero.com/eng/release-notes-nerodigital-nero-aac-codec.html)?

[tebasuna51]

Can anyone confirm that Nero Directshow AAC Decoder filter produce the same output as neroAacDec from NeroDigitalAudio.zip (http://www.nero.com/eng/release-notes-nerodigital-nero-aac-codec.html)?

The (1) DirectShow decoder output 24 bits but (2) NeroAacDec only 16 bits.

Using a unique test file, the differences between the (1) downsampled to 16 bits and (2) are only in the last significant bit (-90.31 dB).

[Greif]

Since NeroAAcEnc now encodes 7.1, does eac3to support creating 7.1 channel AAC from DTS-HD/TrueHD/EAC3?

[saint-francis]

I am working with a BD that is several TV episodes. When I use -check it seems to only check one episode. Is this intentional? What if there is an error somewhere else in the rip?

[magic144]

Just been looking into some AC3 time-stretching using ffmpeg/sox/aften for the decode/stretch/encode steps (basically to shorten a movie's AC3 soundtrack by 160ms).

Just wondering if there are any hidden problems in using this toolchain/workflow.
I notice that the output .ac3 has the dialnorm gone (according to eac3to), presumably because when ffmpeg transforms the input .ac3 to multichannel .wav, it gets 'lost' (discarded?).

Not sure if ffmpeg is ignoring the dialnorm in doing this conversion, or if it would be worth a preliminary .ac3->.ac3 pass with eac3to to remove dialnorm?? What do those in the know think? Would a preliminary pass (.ac3->.ac3) with eac3to have any other side-effects (don't know too much about bit-depth or any other factors - would anything else be adjusted?)

Cheers in advance,
m

[DrNein]

I found that NeEacDec.dll (from a previous install or otherwise extracted) can be added to the DSFilter folder after installing Nero Micro (all options), thus resulting in Nero decode for eac3to and of course Nero Burning ROM while taking less space than installing only ShowTime (which is not much use anyway).

[DrNein]

magic144, I cannot imagine why it would be desirable to time-stretch a movie like that. More likely a delay/cut or edit is appropriate.

[rebkell]

magic144, I cannot imagine why it would be desirable to time-stretch a movie like that. More likely a delay/cut or edit is appropriate.

I was wondering what the purpose of stretching was myself, of course I'm only looking at it from my perspective, which usually involves dropped audio frames in my captures, which involves figuring out where it happens and adding in the right spots.

[magic144]

@DrNein - I cannot imagine why a movie rip would be so broken in this way, but I've done this on 2 films now, where the audio is gradually seen to go ~160ms out of sync by the end of the feature

I do know that rips I've done with eac3to do not suffer from this issue :-)

anyway, it is certainly possible that there may be 1 or only a handful of distinct points at which the audio might jump in sync (I can't pinpoint them), but from experience, applying a linear progressive delay in this way makes the entire film totally watchable and whatever sync errors might exist throughout are significantly harder to perceive with this approach

I didn't invent this by the way - I found it here:-
http://www.nabble.com/How-to-time-stretch-(speed-correction)-AC3-or-MPEG2-files-under-Linux--td12474612.html

But disregarding the reasons, can anyone answer my specific tech. queries wrt dialnorm?!

[sundansx]

Madshi,
Small request: Next time you are in the code, could you make it so the progressnumbers feature prints the progress percentage on one line instead of printing a new line each time. As you likely know, you can do this by printf-ing a few backspace (\b) escape sequences.

also I saw this tonight for the first time (it looked like the whole process was aborted when this happened):
eac3to v2.87
command line: eac3to.exe F:\VideoStore\DVD\Futurama_BendersGame_HD\FUTURAMA3_BENDERSGAME 1) 1: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\chapters.txt" 2: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\out.mkv" 4: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\out.dts" -arcsoft -core 7: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\subtitle.sup" -2pass
------------------------------------------------------------------------------
M2TS, 2 video tracks, 3 audio tracks, 6 subtitle tracks, 1:27:27
1: Chapters, 13 chapters
2: h264/AVC, 1080p24 /1.001 (16:9)
3: h264/AVC, 480p24 /1.001 (20:11)
4: DTS Master Audio, English, 5.1 channels, 24 bits, 48khz
(core: DTS, 5.1 channels, 24 bits, 1509kbps, 48khz)
5: AC3, English, 2.0 channels, 224kbps, 48khz, dialnorm: -27dB
6: DTS Express, English, 2.0 channels, 24 bits, 192kbps, 48khz
7: Subtitle (PGS), English
8: Subtitle (PGS), Spanish
9: Subtitle (PGS), French
10: Subtitle (PGS), Chinese
11: Subtitle (PGS), English
12: Subtitle (PGS), English
Creating file "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\chapters.txt"...
[v02] Extracting video track number 2...
[a04] Extracting audio track number 4...
[s07] Extracting subtitle track number 7...
[a04] Extracting DTS core...
[v02] Muxing video to Matroska...
[a04] Creating file "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\out.dts"...
[s07] Creating file "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\subtitle.sup"...
[a04] The last DTS frame is incomplete and thus gets skipped.
The temp file could not be interpreted correctly.

is this a bug or a bad disc?
thanks.

[tebasuna51]

Just been looking into some AC3 time-stretching using ffmpeg/sox/aften for the decode/stretch/encode steps (basically to shorten a movie's AC3 soundtrack by 160ms).

Just wondering if there are any hidden problems in using this toolchain/workflow.
I notice that the output .ac3 has the dialnorm gone (according to eac3to), presumably because when ffmpeg transforms the input .ac3 to multichannel .wav, it gets 'lost' (discarded?).

Not sure if ffmpeg is ignoring the dialnorm in doing this conversion, or if it would be worth a preliminary .ac3->.ac3 pass with eac3to to remove dialnorm?? What do those in the know think? Would a preliminary pass (.ac3->.ac3) with eac3to have any other side-effects (don't know too much about bit-depth or any other factors - would anything else be adjusted?)

There are many ffmpeg versions, forks, ...
I can't know what your version do, maybe apply the DialNorm and the DRC.

Use eac3to instead ffmpeg to decode the ac3.

[idbirch2]

I can't seem to extract the video and dtshd audio from the King Kong BluRay usig 2.87. Here's the log:
eac3to v2.87
command line: eac3to "I:\0Stuff\King Kong" 1) 2:F:\kk.mkv 5:kk.dtshd 10:F:\kk1.sup
------------------------------------------------------------------------------
M2TS, 3 video tracks, 5 audio tracks, 5 subtitle tracks, 3:20:07
1: Chapters, 54 chapters
2: VC-1, 1080p24 /1.001 (16:9)
3: VC-1, 480p24 /1.001 (3:2)
4: VC-1, 480p24 /1.001 (3:2)
5: DTS Master Audio, English, 5.1 channels, 24 bits, 48khz
(core: DTS, 5.1 channels, 24 bits, 1509kbps, 48khz)
6: DTS, Spanish, 5.1 channels, 24 bits, 768kbps, 48khz
7: DTS, French, 5.1 channels, 24 bits, 768kbps, 48khz
8: AC3 Surround, English, 2.0 channels, 192kbps, 48khz, dialnorm: -27dB
9: DTS Express, English, 2.0 channels, 24 bits, 192kbps, 48khz
10: Subtitle (PGS), English
11: Subtitle (PGS), Spanish
12: Subtitle (PGS), French
13: Subtitle (PGS), Spanish
14: Subtitle (PGS), French
[a05] Extracting audio track number 5...
[v02] Extracting video track number 2...
[s10] Extracting subtitle track number 10...
[v02] Muxing video to Matroska...
[a05] Creating file "kk.dtshd"...
[s10] Creating file "F:\kk1.sup"...
[a05] The last DTS frame is incomplete and thus gets skipped.
[a05] Audio overlaps for 11ms at playtime 1:14:09.
[a05] Audio overlaps for 7ms at playtime 1:16:59.
[a05] Audio overlaps for 6ms at playtime 1:38:53.
[a05] Audio overlaps for 9ms at playtime 2:09:19.
[a05] Audio overlaps for 6ms at playtime 2:46:16.
[a05] Audio overlaps for 8ms at playtime 2:46:38.
[a05] Audio overlaps for 11ms at playtime 2:53:16.
[a05] Audio overlaps for 10ms at playtime 3:14:19.
[a05] Starting 2nd pass...
[a05] Realizing DTS gaps...
[a05] Creating file "kk.dtshd"...
[a05] Writing the destination file failed.
Aborted at file position 5681233920.
edit: I'm getting the same error with another disc, no seamless branching on this one though, used -demux switch instead and got everything I needed.

[73ChargerFan]

magic144,

160ms is very short... that seems to be a sync problem, not time-stretching.

eac3to can remove time at various places, like 30ms at 20%, 40%, 60%, 80% and 40ms at the end.

Or apply a negative delay?

Edit: sorry for getting to this discussion late...

[mbcd]

Hi

I think its a small bug:

eac3to crashes while extracting the core of an eac3-File

eac3-File is 640kbps and i didn`t wanted to reencode to get normal ac3.

so i tried to extract the core, but eac3to crashed using libav ...

eac3to-v2.87>eac3to.exe "XXXXX\bb - 6 - E-AC3, German, 5.1 channels, 640kbps, 48khz.eac3"
E-AC3, 5.1 channels, 2:20:03, 640kbps, 48khz, dialnorm: -27dB

eac3to-v2.87>eac3to.exe "XXXXXXX\bb - 6 - E-AC3, German, 5.1 channels, 640kbps, 48khz.eac3" g:\BB\ger.ac3 -core

E-AC3, 5.1 channels, 2:20:03, 640kbps, 48khz, dialnorm: -27dB
The Nero decoder doesn't seem to work, will use libav instead.
eac3to crashed...


Possible to get the core (as normal ac3) without reencoding ? That "should" normaly not be any problem, or ?

Regards

[madshi]

i'm sorry if this has been asked before but i was wondering if its poss to output celltimes.txt or not?
I don't even know what that is right now. What is it exactly and what purpose does it have?

The writer told me that he could use eac3to if eac3to could be able to handle mkv or pipes as input/output.
Due to how eac3to is designed, it doesn't support input pipes, so that won't work, unfortunately...

And second problem - "sound forge 8" can't open stereo wav-files, generated by "eac3to".
That's most probably a bug/problem with sound forge 8 and not with eac3to. There are 2 different kinds of WAV files and sound forge 8 seems to only like the older one (which is outdated). Use the "-simple" switch to create WAV files in the older format with eac3to.

Is there special reason that "eac3to" creates different header for stereo file?
IIRC eac3to creates "simple" WAV files for mono files if you use the "something.wavs" output. The reason for that is that outputting "wavs" is usually useful for Surcode DTS encoding and Surcode only understands the "simple" WAV headers, too. Furthermore, with mono WAV files the channel mask information is not so important as with multichannel WAV files (the channel mask cannot be stored in the "simple" WAV format).

I forgot in time the switch, with which I can get float-wav. Can somebody please remind me?
"-full"

When capturing output from "eac3to" like this: "eac3to>help.txt", file is filled with 08h characters, but they must have not be there.
That's caused by the drawing of the progress bar. I don't think I can get rid of that.

Could you add a command that forces dialog normalization...something like -forceDialnorm

Or even a command that would me me specify what level of dialog normalization an ac3 file has? Either way would work. Yet another option would be if eac3to detected a -0 dialog normalization level as having dialog normalization so that when eac3to extracted the file it would change it to -31.
No option for forcing dialnorm right now. But at least the next version will automatically patch "-0db" dialnorm AC3 tracks to "-31db".

I really don't like that the second pass is needed to eliminate clipping.
There's no way to detect and fix clipping in one pass. However, you can manually lower volume by adding e.g. "-3db" to the command line. In most cases that should make sure that no clipping occurs, so no 2nd pass will be necessary.

Since NeroAAcEnc now encodes 7.1, does eac3to support creating 7.1 channel AAC from DTS-HD/TrueHD/EAC3?
To be honest, I'm not sure myself right now. Why don't you simply give it a try and report back if it doesn't work.

Next time you are in the code, could you make it so the progressnumbers feature prints the progress percentage on one line instead of printing a new line each time. As you likely know, you can do this by printf-ing a few backspace (\b) escape sequences.
The progressnumbers feature is not meant for end users, it's only meant to be used by other programs which want to "remote control" eac3to. For such programs it's better if I don't use backspace.

I can't seem to extract the video and dtshd audio from the King Kong BluRay usig 2.87. Here's the log:
[a05] Writing the destination file failed.
Aborted at file position 5681233920.

Probably your harddisk is full. Or the network connection got lost. Or your harddisk is broken? I don't see many other reasons for why writing to a file might fail. This doesn't look like an eac3to related error to me.

[madshi]

also I saw this tonight for the first time (it looked like the whole process was aborted when this happened):

eac3to v2.87
command line: eac3to.exe F:\VideoStore\DVD\Futurama_BendersGame_HD\FUTURAMA3_BENDERSGAME 1) 1: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\chapters.txt" 2: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\out.mkv" 4: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\out.dts" -arcsoft -core 7: "F:\VideoStore\DVD\test\Futurama_BendersGame_HD\subtitle.sup" -2pass
You're asking eac3to to process the subtitle in 2pass mode. That's not supported. The 2pass mode must be activated for every track you want it be active for. You should only use it for audio tracks.

[asarian]

@DrNein - I cannot imagine why a movie rip would be so broken in this way, but I've done this on 2 films now, where the audio is gradually seen to go ~160ms out of sync by the end of the feature

I do know that rips I've done with eac3to do not suffer from this issue :-)


Never seen a commercial movie (the last ten years, at least) where the sound gradually goes out of sync, and if it's not caused by eac3to, as you seem to say, then why not just use eac3to?

My guess, in the words of eac3to itself, is that "The audio file was demuxed without making use of the gap/overlap information." (in a second pass).

[Jeff Flowerday]

I'm trying to repackage some of my SDs in to mkv containers. In many cases eac3to reports multiple video and audio tracks on the concatinated vob files. It also indicates 4 subtitle tracks when there aren't any in the file.

I'm not sure how much of a priority vob files are for you, but I've attached a sample.

Ultimately it doesn't matter which tool I use to extract the elementary streams from Pineapple Express the remuxed mkv has the audio slowly getting more and more out of sync as the movie progresses.

VOB, 2 video tracks, 2 audio tracks, 4 subtitle tracks, 0:04:59
1: MPEG2, 480p24 /1.001 (16:9) with pulldown flags, -173381ms
2: MPEG2, 480p24 /1.001 (16:9) with pulldown flags, -41ms
3: AC3, 5.1 channels, 448kbps, 48khz, dialnorm: -29dB, -173381ms
4: AC3, 5.1 channels, 448kbps, 48khz, dialnorm: -29dB, -41ms
5: Subtitle
6: Subtitle
7: Subtitle
8: Subtitle

Here is the clip:

http://www.flowerdayconsulting.com/Misc/Clip.zip

[setarip_old]

@Jeff Flowerday

For conversion of standard DVDs to MKVs, I'd suggest you try the recently introduced "MakeMKV" which, if need be will first very quickly "rip"/decrypt your DVD and then (also very quickly) convert them to MKV...

[madshi]

Hi

I think its a small bug:

eac3to crashes while extracting the core of an eac3-File

eac3-File is 640kbps and i didn`t wanted to reencode to get normal ac3.

so i tried to extract the core, but eac3to crashed using libav ...

eac3to-v2.87>eac3to.exe "XXXXX\bb - 6 - E-AC3, German, 5.1 channels, 640kbps, 48khz.eac3"
E-AC3, 5.1 channels, 2:20:03, 640kbps, 48khz, dialnorm: -27dB

eac3to-v2.87>eac3to.exe "XXXXXXX\bb - 6 - E-AC3, German, 5.1 channels, 640kbps, 48khz.eac3" g:\BB\ger.ac3 -core

E-AC3, 5.1 channels, 2:20:03, 640kbps, 48khz, dialnorm: -27dB
The Nero decoder doesn't seem to work, will use libav instead.
eac3to crashed...


Possible to get the core (as normal ac3) without reencoding ? That "should" normaly not be any problem, or ?
Please try again with the next build, once it's out (later today). If the crash still occurs, please send me the crash report via email (dear (at) madshi (dot) net).

Extracting the core doesn't really work since HD DVD style E-AC3 tracks don't have a core. But still there should be no crash, of course...

I'm trying to repackage some of my SDs in to mkv containers. In many cases eac3to reports multiple video and audio tracks on the concatinated vob files. It also indicates 4 subtitle tracks when there aren't any in the file.

I'm not sure how much of a priority vob files are for you, but I've attached a sample.

Ultimately it doesn't matter which tool I use to extract the elementary streams from Pineapple Express the remuxed mkv has the audio slowly getting more and more out of sync as the movie progresses.

VOB, 2 video tracks, 2 audio tracks, 4 subtitle tracks, 0:04:59
1: MPEG2, 480p24 /1.001 (16:9) with pulldown flags, -173381ms
2: MPEG2, 480p24 /1.001 (16:9) with pulldown flags, -41ms
3: AC3, 5.1 channels, 448kbps, 48khz, dialnorm: -29dB, -173381ms
4: AC3, 5.1 channels, 448kbps, 48khz, dialnorm: -29dB, -41ms
5: Subtitle
6: Subtitle
7: Subtitle
8: Subtitle
Does that VOB come from a DVD? If so, it would probably be wise to use a tool like DVDDecrypter, because it makes use of the IFO files to properly extract the tracks. eac3to works fine for "simple" VOB files. But complicated DVDs (e.g. multi angle or seamless branching) are not yet properly supported. I'll have to add full IFO parsing etc in order to make that work...

However, with the next version the vob sample you uploaded is only listed with one video and audio track. So maybe the problem will already be fixed then.

The subtitles are shown because they're listed in the VOB information record. eac3to doesn't check whether there's actual subtitle data because sometimes with forced subtitles the first data comes in the middle of the movie.

Mashi, found a bug in v2.87 with a BD, wrong play duration is displayed.

O:\bd\2>eac3to h:
1) 00021.mpls, 00005.m2ts+00006.m2ts, 1:53:44
- VC-1, 1080p24 /1.001 (16:9)
- DTS, German, multi-channel, 48khz
- DTS, English, multi-channel, 48khz
- RAW/PCM, German, multi-channel, 48khz
- RAW/PCM, English, multi-channel, 48khz

2) 00006.mpls, 00006.m2ts, 1:53:39
- VC-1, 1080p24 /1.001 (16:9)
- DTS Master Audio, German, multi-channel, 48khz
- DTS Master Audio, English, multi-channel, 48khz
- RAW/PCM, German, multi-channel, 48khz
- RAW/PCM, English, multi-channel, 48khz

O:\bd\2>eac3to h: 1)
M2TS, 1 video track, 4 audio tracks, 2 subtitle tracks, 0:00:05
1: Chapters, 22 chapters
2: VC-1, 1080p24 /1.001 (16:9)
3: DTS, German, 5.1 channels, 24 bits, 768kbps, 48khz
4: DTS, English, 5.1 channels, 24 bits, 768kbps, 48khz
5: RAW/PCM, German, 5.1 channels, 16 bits, 48khz
6: RAW/PCM, English, 5.1 channels, 16 bits, 48khz

Here the duration is now just 5 sec - seems this is the time for 00005.m2ts file which just displays "Senator"?

O:\bd\2>eac3to h: 2)
M2TS, 1 video track, 4 audio tracks, 2 subtitle tracks, 1:53:39
1: Chapters, 22 chapters
2: VC-1, 1080p24 /1.001 (16:9)
3: DTS Master Audio, German, 5.1 channels, 24 bits, 48khz
(core: DTS, 5.1 channels, 24 bits, 768kbps, 48khz)
4: DTS Master Audio, English, 5.1 channels, 24 bits, 48khz
(core: DTS, 5.1 channels, 24 bits, 768kbps, 48khz)
5: RAW/PCM, German, 5.1 channels, 16 bits, 48khz
6: RAW/PCM, English, 5.1 channels, 16 bits, 48khz

Here the time is correct...
It seems that the two m2ts parts differ slightly. E.g. 00005.m2ts seems to only have DTS in it, while 00006.m2ts has full Master Audio. eac3to doesn't like that. An audio track should always stay in the same format.

I'd still like to find out why the runtime is only 5 seconds for the playlist 1), but unfortunately for that I'd need the full 00005.m2ts file and the first 50MB of the 00006.m2ts. But in the end all I may find is that the runtime is wrong because the audio track properties differ between the two m2ts parts. So it's your choice whether uploading samples is worth it for you or not.

Had finally time to make a sample to reproduce the issue from post #7225.
(issue: cannot extract EAC3 5.1 audio track from EVO, got endless 13ms audio overlap warnings, preventing eac3to from completing)

However, while testing the sample, eac3to doesn't like incomplete EAC3 packets anymore seems. Got crash.
Will be fixed in the next build.

Another problem - "eac3to" crashes with this "m2ts"-file (http://rapidshare.de/files/41302230/00001.m2ts.html).
VC-1 stream was incorporated from WMV-file.
That VC-1 stream does not have any framerate information in it. That made eac3to crash. The next build will handle this gracefully. But it will "guess" the framerate to be 23.976.

eac3to seems not to strip zero bytes from truehd->flac anymore
Thanks for the note. Will be fixed in the next build.

When I convert a DTS Hi-Res stream (from Predator 2) to LPCM, eac3to doesn't seem to list the final bit-depth.

Normally, it gives me a message saying the final output file had 16/24 bit depth, etc. This time not. Is that a small bug?
It seems to work with the next build. Don't know if I accidently fixed it?

[madshi]

Now "eac3to" creates file "j5e.wav" in current folder of current disk, but v.2.80 (I've upgraded from this version to 2.87) has created file "5e.wav" in current folder of disk "j", as it's supposed to be. Can you please fix it back?!
Will be fixed in the next build.

[zeropc]

question...

how do i use the "fix gaps" feature in cmd? currently i use the gui and check "fix gaps" but i'd like to know how to use it in cmd.

[madshi]

how do i use the "fix gaps" feature in cmd? currently i use the gui and check "fix gaps" but i'd like to know how to use it in cmd.
There's nothing you need to do. The current eac3to build will fix audio gaps automatically. Things are different with video gaps, but they are extremely rare. The check box in the GUI is probably intended for audio gaps and it's not needed, anymore. Some of the eac3to GUIs are kind of outdated, sadly...

[kurt]

small bug when using -test. eac3to reports a new version of haali matroska muxer (but there is none).

[zeropc]

does the auto gap fix work for video too or will become a feature in the future?

[madshi]

small bug when using -test. eac3to reports a new version of haali matroska muxer (but there is none).
There *is* a new Haali version.

does the auto gap fix work for video too or will become a feature in the future?
It does not work for video and I don't plan to make it available because (1) video gaps are very rare and (2) there can be false alarms. So if video gaps are reported you need to do some manual work in any case...

[sundansx]

madshi - thanks for the help. thanks for your work on eac3to - excellent program.

[kurt]

There *is* a new Haali version.

hm, I'm quite sure I have the latest haali media splitter (http://haali.cs.msu.ru/mkv/) (1.9.42.1) installed.


eac3to -test
...
Haali Matroska Muxer (2009-01-10) is installed
There's a new version (2009-01-11) available
http://haali.net/mkv
....

[Jeff Flowerday]

hm, I'm quite sure I have the latest haali media splitter (http://haali.cs.msu.ru/mkv/) (1.9.42.1) installed.


I get the same message with the new Haali splitter.

[Thunderbolt8]

madshi, is there anything useful for eac3to with the new haali version? I mean any advantages for you and us? because just with the changelog I cannot do much, not sure how that
# Broken Matroska files with looped SeekHeads could cause a hang in Matroska Parser, the number of SeekHeads is now limited to 10
# Removed the workaround to find tags written by Matroska Shell Extension, this caused excessive file scanning when opening files created by recent MKVToolnix
cold affect remuxing haali files with mkvmerge and such

another question regarding the -bitrate switches for ac3 decoding: when slowing down a 25fps track to 23.976 then the default value is always 640 kbit. assuming the original track has a lower initial bitrate, can I use that switch then to adjust to that bitrate and still would have zero quality loss of the file I slowed down?

[madshi]

hm, I'm quite sure I have the latest haali media splitter (http://haali.cs.msu.ru/mkv/) (1.9.42.1) installed.
Oh well. There's nothing I can do. The homepage says 11th January while the splitter itself says 10th. So eac3to thinks that there's a new version available. No way to fix that, unfortunately...

[Jeff Flowerday]

Does that VOB come from a DVD? If so, it would probably be wise to use a tool like DVDDecrypter, because it makes use of the IFO files to properly extract the tracks. eac3to works fine for "simple" VOB files. But complicated DVDs (e.g. multi angle or seamless branching) are not yet properly supported. I'll have to add full IFO parsing etc in order to make that work...


Yes, it's from a DVD.

I used the MakeMKV and it worked, so I'm fine. Thanks again!

[magic144]

@all
thanks for the comments and responses re: audio stretching
just to say that this technique is merely a band-aid solution for an obviously broken rip - the stretching merely hides the problem(s) as far as general human perception is concerned - as for the tools, I will consider using eac3to to convert ac3 to wav in future, for a measure of certainty (for whatever the desired application/result)

the real solution is of course to go back to the source and rip the original using the latest and greatest version of eac3to!

cheers

[asarian]

another question regarding the -bitrate switches for ac3 decoding: when slowing down a 25fps track to 23.976 then the default value is always 640 kbit. assuming the original track has a lower initial bitrate, can I use that switch then to adjust to that bitrate and still would have zero quality loss of the file I slowed down?

I distinctly recall Madshi stating, earlier, that stretching always causes quality degradation at the same bitrate. I take it that's why it will use 640Kbps anyway (even though the source may have been at, say, 448Kbps).

[setarip_old]

@Jeff Flowerday I used the MakeMKV and it worked, so I'm fine. Thanks again!I'm glad to hear that my suggestion resolved your dilemma ;>}

[madshi]

First release in the year 2009. Since I didn't like the idea of reaching "v2.100" sometime this year, I decided to use the opportunity to jump to "v3.00". That's also the reason why there was no new build last week. I've "collected" some things, so that v3.00 really is a worthy upgrade. Enjoy!

-------------------------------------------

eac3to v3.00 released

http://madshi.net/eac3to.zip

* broken AC3, DTS, AAC and MPx streams are now automatically repaired
* errors in TS/m2ts files are now reported (with runtime) and ignored
* damaged first max 5MB and max 5% of a TS/m2ts file are automatically skipped
* video/audio tracks which can't be parsed, are now demuxed in raw form
* added support for "line 21" closed captions in ATSC/NTSC broadcasts and DVDs
* added reading of movie / network name from "line 21" XDS information
* for gaps, edits & repairs > 1000ms eac3to now inserts silence by default
* for gaps, edits & repairs < 1000ms eac3to now loops audio by default
* option "-silence" forces eac3to to insert silence instead of looping audio
* option "-loop" forces eac3to to loop audio instead of inserting silence
* newly encoded AC3 frame is now used for "silence" instead of file's 1st frame
* increased reading block size (might improve reading performance)
* optimized TS/m2ts demuxing performance
* optimized MPEG2, VC-1 and h264 parsing performance
* command line output is colored now (e.g. errors drawn in red)
* MPEG2 1920x1088 bitstream is now automatically patched/cropped to 1920x1080
* log file now contains "<WARNING>" and "<ERROR>" indicators
* workaround for movie playlists which want the same m2ts file played twice
* added version check for eac3to (doh!)
* when a read error occurs, reading is tried again up to 3 times
* (E-)AC3 frames with -0db dialnorm are now automatically patched to -31db
* updated to newer libAften build -> fixes 44.1khz encoding
* fixed: sometimes "The last DTS frame is incomplete" was a false alarm
* fixed: mkvtoolnix version check didn't work, anymore
* fixed: errors were meant to be output to stderr, but they weren't
* fixed: automatic gap/overlap fixing with AAC targets aborted processing
* fixed: positive edit began a bit too early
* fixed: two ID3 tags after each other made eac3to fail detecting the format
* fixed: some VOB files were not detected properly
Since there have been a lot of changes, some new bugs may appear (hopefully not).

[itsancho]

Thank You, madshi! :thanks:

[Thunderbolt8]

looks huge, thanks!!! :thanks:

edit: can you please explain how that automatic repair thing technically works? I guess this mainly refers to those .ts streams with audio frames errors?

[asarian]

@Madshi,

My, you've been busy! :) :thanks:

[madshi]

can you please explain how that automatic repair thing technically works? I guess this mainly refers to those .ts streams with audio frames errors?
Corrupt frames are simply deleted from the audio streams. If the source is a container, the audio gap detection will automatically detect the resulting gaps in the audio stream and fill them again with either looping or silence (depending on how big the gap is). This way the audio tracks should stay in sync.

With demuxed audio tracks eac3to tries to keep the audio track in sync, too, but it's more difficult because the helpful container timestamps are missing. So keeping audio in sync when repairing demuxed audio tracks can be a lottery.

[Atak_Snajpera]

madshi your improvements are really amazing! Excellent JOB!!!

[itsancho]

oops... unfortunately i found the first bug, sorry...
eac3to v3.00
command line: z\eac3to "H:\Casino Royale (2006) Blu-ray AVC PCM" 1) 1: e:\Chapters.txt 2: e:\Casino.mkv 4: l:\Casino.wav 7: d:\Casino.sup 8: d:\Casino1.sup
------------------------------------------------------------------------------
M2TS, 1 video track, 4 audio tracks, 8 subtitle tracks, 2:24:13
1: Chapters, 16 chapters
2: h264/AVC, 1080p24 /1.001 (16:9)
3: AC3, English, 5.1 channels, 448kbps, 48khz
4: RAW/PCM, English, 5.1 channels, 16 bits, 48khz
5: AC3, French, 5.1 channels, 640kbps, 48khz
6: AC3, Spanish, 5.1 channels, 448kbps, 48khz, dialnorm: -29dB
7: Subtitle (PGS), English
8: Subtitle (PGS), English
9: Subtitle (PGS), French
10: Subtitle (PGS), Spanish
11: Subtitle (PGS), Chinese
12: Subtitle (PGS), Korean
13: Subtitle (PGS), Portuguese
14: Subtitle (PGS), Thai
Creating file "e:\Chapters.txt"...
[v02] Extracting video track number 2...
[a04] Extracting audio track number 4...
[a04] The ts pcm reader received a non-supported data format. <ERROR>
[s08] Extracting subtitle track number 8...
[s07] Extracting subtitle track number 7...
[v02] Muxing video to Matroska...
Aborted at file position 1048576. <ERROR> with 2.87 audio is demuxing just fine.

[nautilus7]

Corrupt frames are simply deleted from the audio streams. If the source is a container, the audio gap detection will automatically detect the resulting gaps in the audio stream and fill them again with either looping or silence (depending on how big the gap is). This way the audio tracks should stay in sync.

With demuxed audio tracks eac3to tries to keep the audio track in sync, too, but it's more difficult because the helpful container timestamps are missing. So keeping audio in sync when repairing demuxed audio tracks can be a lottery.

Was wondering about this, too. Thanks for that nice feature.

[magic144]

and thanks for the looping/silence audio edit options!

[sidekick2]

Thanks madshi!

I'm giving it a whirl right now.

Since you're using colors now, how about changing the "----" as it's converting to some bright color? The dark gray is hard to see with my old eyes!

I don't know if this a bug, or if I did something wrong with this: (It didn't like .thd+ac3 -- Bourne-Ultimatum hd-dvd)

C:\eac3to-ver3>eac3to.exe g: 1) 3: L:\bourne-ult\bu.vc1 2: L:\bourne-ult\bu.txt 5: L:\bourne-ult\bu.
thd+ac3 10: L:\bourne-ult\bu.sup
EVO, 2 video tracks, 5 audio tracks, 3 subtitle tracks, 1:55:18
"Main Movie"
1: Joined EVO file
2: Chapters, 20 chapters with names
3: VC-1, 1080p24 /1.001 (16:9) with pulldown flags
4: VC-1, 480p24 /1.001 (3:2)
5: TrueHD, English, 5.1 channels, 48khz, dialnorm: -27dB
6: E-AC3, English, 5.1 channels, 384kbps, 48khz, dialnorm: -27dB, -17ms
7: E-AC3, French, 5.1 channels, 384kbps, 48khz, dialnorm: -27dB, -17ms
8: E-AC3 Surround, English, 2.0 channels, 192kbps, 48khz, dialnorm: -27dB, -17ms
"Commentary"
9: E-AC3, 2.0 channels, 192kbps, 48khz, dialnorm: -27dB, -4ms
10: Subtitle (VobSub), English, "SDH"
11: Subtitle (VobSub), French
12: Subtitle (VobSub), French, "Forced"
Creating file "L:\bourne-ult\bu.txt"...
v03 Extracting video track number 3...
v03 There's no valid framerate in this bitstream.
a05 Extracting audio track number 5...
a05 Extracting audio track number 5...
v03 Writing new framerate "24fps /1.001" to bitstream.
a05 Removing TrueHD dialog normalization...
a05 Removing TrueHD dialog normalization...
v03 Writing new framerate "24fps /1.001" to bitstream.
a05 Decoding with libav/ffmpeg...
v03 Removing VC-1 pulldown...
a05 Encoding AC3 <640kbps> with libAften...
a05 TrueHD overflow in the thd ac3 joiner.
Aborted at file position 1048576.

[madshi]

oops... unfortunately i found the first bug, sorry...
with 2.87 audio is demuxing just fine.
I don't know if this a bug, or if I did something wrong with this: (It didn't like .thd+ac3 -- Bourne-Ultimatum hd-dvd)
These two bugs will be fixed in v3.01.

Since you're using colors now, how about changing the "----" as it's converting to some bright color? The dark gray is hard to see with my old eyes!
I find the dark progress bar prettier. But if you don't like it, you can adjust the colors to your heart's content in the command prompt settings (right click on the top left corner of the command prompt's title bar, then choose "settings").

What is everyone's opinion? Is the dark gray progress bar good? Or do you prefer the old normal colored one?

[madshi]

eac3to v3.01 released

http://madshi.net/eac3to.zip

* fixed: m2ts LCPM demuxing didn't work with v3.00
* fixed: TrueHD -> TrueHD+AC3 conversion didn't work with v3.00

[bmnot]

Great update, thanks!

So when's this program going to be renamed to Madshi's Uber HD Tool? :D

[itsancho]

thanx, madshi!

well, about colours... i've ment not to border u with this, but since u've asked. ;-)
i prefer command prompt to look "normal" - white background/black screen text, but with v3.00/3.01 every time i hit "Enter" eac3to automaticly is chaning my custom made settings to black background/grey text...

[nautilus7]

I like the colors as they are...

madshi, what does this mean?

eac3to "F:\bd-sc-j" 1) 3: aaa.flac
M2TS, 1 video track, 4 audio tracks, 2:04:01
1: Chapters, 29 chapters
2: h264/AVC, 1080p24 /1.001 (16:9)
3: TrueHD/AC3, English, 5.1 channels, 48khz, dialnorm: -27dB
(embedded: AC3, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB), -13ms
4: TrueHD/AC3, Japanese, 5.1 channels, 48khz, dialnorm: -27dB
(embedded: AC3, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB), -13ms
5: AC3, English, 2.0 channels, 256kbps, 48khz, dialnorm: -27dB, -13ms
6: AC3, English, 2.0 channels, 256kbps, 48khz, dialnorm: -27dB, -13ms
a03 Extracting audio track number 3...
a03 This track begins with a non-major frame.
a03 Extracting TrueHD stream...
a03 Removing TrueHD dialog normalization...
a03 Decoding with libav/ffmpeg...
a03 Applying RAW/PCM delay...
a03 Encoding FLAC with libFlac...
a03 Creating file "aaa.flac"...

It's Sin City JAP Blu-ray Disc. The procedure seems to continue normally after that.