eac3to - audio conversion tool

[madshi]

Originally Posted by Bluestraw

Are you absolutely sure about this? For example, in the Cars BD I was talking about, there are 3 language versions. All the playlists shown by eac3to are identical, yet different languages do result in a different required join order. The same thing happens in Open Season - there are different orders for the different languages, simply in order to show the movie title on screen in the right language, and all this in the single mpls file.

If you load that mpls file I sent above into the latest version of BDEdit, you'll see that it contains 3 'angles' - I presume these correspond to the 3 languages.

Hmmmm... You're right, I forgot about angles. eac3to always uses the first angle, IIRC. Do you need access to the other ones, too? I don't think the order of the parts changes, though. Instead different languages are using different parts, I believe.

[Bluestraw]

Originally Posted by madshi

Hmmmm... You're right, I forgot about angles. eac3to always uses the first angle, IIRC. Do you need access to the other ones, too? I don't think the order of the parts changes, though. Instead different languages are using different parts, I believe.

That's right I think - e.g. you could have

0000
0001
0004
0005
0008

as the English version,

0000
0002
0004
0006
0008

as the German one

0000
0003
0004
0007
0008

as the French one.

where for example part 0001 has some on-screen English text, 0002 has German and 0003 has French. From what I've seen so far, each one has all the audio tracks, so for example you could watch the 'French version' but keep the English soundtrack.

From my pov, I always want the English one, so it depends for me whether or not which angle that is. My guess would be that I'd typically be lucky and get the first one, but others (like you!) may not find their first choice available!

[madshi]

Originally Posted by Bluestraw

others (like you!) may not find their first choice available!

Actually I want the English one, too, which is usually the first. But there may be other people who want their local version. So I guess angle support would make sense...

[Bluestraw]

Originally Posted by madshi

Actually I want the English one, too, which is usually the first. But there may be other people who want their local version. So I guess angle support would make sense...

Sorry didn't mean to offend, I thought I remember you saying you weren't English. Not that I could tell from what you write - impeccable English!

[nautilus7]

Surf's Up consists of 3 m2ts files. The 2nd is language specific. It might uses angles too.

[Thunderbolt8]

Originally Posted by madshi

Anyway, I hear you. I can lower the threshold to something lower than 40ms. But there's one thing you should know: The timestamps are not 100% reliable. If I lower the threshold too much, we'll get false alarms. I mean then eac3to will report audio gaps/overlaps for movies which don't really have any gaps/overlaps. We don't want that, do we? So it's kind of difficult to find the right threshold. But maybe 40ms is really a bit too much. I chose this value because an AC3 frame is usually 32ms long and I wanted to allow one AC3 frame "off" without reporting a gap/overlap.

false alarms would indeed be really bad, you are right that we dont want that. hm is there maybe a way to combine the gap/overlap checking with the audio tracks used? for example if I only decide to use a pcm or truehd/dts-hd and no ac3 track for my muxing at all would it be possible then to have that checking only apply to those tracks, which have quite shorter frames as ac3?
or would this be too complicated?

[bmnot]

Originally Posted by madshi

Hmmmm... That's quite strange. Could you send me the last 20MB of 00000.m2ts and the first 20MB of 00001.m2ts?

Ok, I will, but I don't understand how to use this HxD program.

[nautilus7]

Originally Posted by bmnot

Ok, I will, but I don't understand how to use this HxD program.

You load the file. Then go to edit--> select block and set the size (20MB) in bytes. Make sure dec is chosen. To select the start of the file you put 0 to start-offset and 20971520 in end-offset. To select the end, click the mouse to the end of the file and then put in the start-offset (end-offset - 20971520). Final step is to copy-paste the selected part to a new file. Be careful not to modify the original file.

[Snowknight26]

Or use clip.exe.

usage: clip <infile> <outfile> <start offset> <length>

[Rectal Prolapse]

madshi, Ratatouille also makes use of multi-angles for different languages.

Is there a way to lower the overlap/gap threshold only for m2ts files that are to be joined from a single playlist, and only at the endpoints of the m2ts files? That may give better results overall - for multiple m2ts and single m2ts files.

Anyways - thanks for the great tool - again.

[jchappo]

howdy, been using your nice program here for awhile. I recently ran into a problem decoding DTS-HD to WAV using Sonic Decoder 4.3. When remuxing the WAV file back with the video in M2TS container, TsMuxer reports the WAV has no channel info.

When playing the movie back over a popcorn hour A-100, the audio plays fine for about 20-30% of the movie, then it goes all screetchy and is basically garbage. I can't determine what the problem is. How to I test if Sonic is not causing the problem? It's weird because the audio plays fine for a little while, then goes bad for some reason.

[itsancho]

hi all
strange problem with Basic Instinct, here is the log:

eac3to v2.44
command line: eac3to244\eac3to "K:\Basic Instinct (1992) BluRay AVC dts-HD Hi-Res" 1) 2: i:\video.mkv
------------------------------------------------------------------------------
M2TS, 1 video track, 4 audio tracks, 2:08:15
1: Chapters, 17 chapters
2: h264/AVC, 1080p24 /1.001 (16:9)
3: DTS Hi-Res, English, 6.1 channels, 24 bits, 3093kbit/s, 48khz, dialnorm: -4dB
4: AC3, English, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
5: AC3, English, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
6: AC3, English, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
[v02] Extracting video track number 2...
[v02] 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 142065664.

i've tried to remux raw h264 stream (using 3 different programs) and then mux it with mkvtollnix, but... :-( runtime was 2:09:44
big 10x in advance 4 helping with this one!

Yraen, nope, for sorry that is not the problem, with or without any " the result is the same...

[Yraen]

Looks like you missed an " here:
i:\video.mkv"

That might be the problem. Try it and if it fails post that log.

[bmnot]

Originally Posted by madshi

Hmmmm... That's quite strange. Could you send me the last 20MB of 00000.m2ts and the first 20MB of 00001.m2ts?

Ok here you go: http://www.mediafire.com/?g499bligrz0

end of first m2ts doesn't play, i hope it did right

BTW the end of the file is at 30684420096, while the error is at 30684829696, BUT the file converts to mkv fine when there is no joining.

[gregt]

Hi,
I seem to have run into another bug as follows:
M:\to convert\SciFi\bluray>eac3to UNDERWORLD
1) 00106.mpls, 2:13:39
[93+94+95+96+97+98+99+100].m2ts
- MPEG2, 480i30 /1.001 (4:3)
- AC3, English, stereo, 48khz

2) 00011.mpls, 00011.m2ts, 2:13:37
- h264/AVC, 1080p24 /1.001 (16:9)
- RAW/PCM, English, multi-channel, 48khz
- AC3, English, multi-channel, 48khz
- AC3, French, multi-channel, 48khz
- RAW/PCM, Italian, multi-channel, 48khz
- AC3, Italian, multi-channel, 48khz
- AC3, English, stereo, 48khz

3) 00095.mpls, 00093.m2ts, 0:47:18
- MPEG2, 480i30 /1.001 (4:3)
- AC3, English, stereo, 48khz

4) 00101.mpls, 00099.m2ts, 0:19:13
- MPEG2, 480i30 /1.001 (4:3)
- AC3, English, stereo, 48khz

M:\to convert\SciFi\bluray>eac3to UNDERWORLD UNDERWORLD.mkv
M2TS, 1 video track, 1 audio track, 2:13:38
1: Chapters, 13 chapters
2: MPEG2, 480i60 /1.001 (4:3)
3: AC3, English, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Creating file "UNDERWORLD - Chapters.txt"...
[a03] Extracting audio track number 3...
[v02] Extracting video track number 2...
[v02] Muxing video to Matroska...
[a03] Removing dialog normalization...
[a03] Creating file "UNDERWORLD - 3 - AC3, English, 2.0 channels, 192kbps, 48khz
.ac3"...
[v02] Video overlaps for 1 frames at playtime 0:47:18.
[v02] Video overlaps for 1 frames at playtime 1:00:21.
[v02] Video overlaps for 1 frames at playtime 1:10:18.
[v02] Video overlaps for 1 frames at playtime 1:22:48.
[v02] Video overlaps for 1 frames at playtime 1:34:31.
[v02] Video overlaps for 1 frames at playtime 1:45:18.
[v02] Video overlaps for 1 frames at playtime 2:04:31.
[v02] The MKV file was created without making use of the gap/overlap information
.
[v02] Please check whether audio is in sync. If it is in sync everything is fine
.
[v02] Otherwise you can ask eac3to to repeat the muxing. It will then automatica
lly
[v02] make use of the detailed gap/overlap information.
[a03] Audio overlaps for 31ms at playtime 0:47:18.
[a03] Audio overlaps for 42ms at playtime 1:00:21.
[a03] Audio overlaps for 27ms at playtime 1:10:18.
[a03] Audio overlaps for 24ms at playtime 1:22:48.
[a03] Audio overlaps for 46ms at playtime 1:34:31.
[a03] Audio overlaps for 39ms at playtime 1:45:18.
[a03] Audio overlaps for 25ms at playtime 2:04:31.
[a03] The audio file was demuxed without making use of the gap/overlap informati
on.
[a03] Please rerun the same eac3to command line. That will correct the gaps/over
laps.
Added fps value to MKV header.
Video track 2 contains 240318 frames.
eac3to processing took 8 minutes, 15 seconds.
Done.

For some reason the main (listed under 2) movie is not converted by default. The main movie file is 38GB so may be that is the reason. When I went into the stream directory and converted the 00011.m2ts file directly it converted correctly.

[madshi]

Originally Posted by Bluestraw

Sorry didn't mean to offend, I thought I remember you saying you weren't English. Not that I could tell from what you write - impeccable English!

You didn't offend me at all, don't worry. You are right in that my native language is not English. So it makes sense that you thought I'd prefer having my native language muxed. But it's just not that way. And thanks for the compliment on my English capabilities... I guess reading/writing English every day helps!

Originally Posted by Thunderbolt8

false alarms would indeed be really bad, you are right that we dont want that. hm is there maybe a way to combine the gap/overlap checking with the audio tracks used? for example if I only decide to use a pcm or truehd/dts-hd and no ac3 track for my muxing at all would it be possible then to have that checking only apply to those tracks, which have quite shorter frames as ac3?

Not sure if I've understood you correctly. eac3to already checks only those tracks you are demuxing.

Originally Posted by Rectal Prolapse

Is there a way to lower the overlap/gap threshold only for m2ts files that are to be joined from a single playlist, and only at the endpoints of the m2ts files? That may give better results overall - for multiple m2ts and single m2ts files.

That sounds like a good idea in theory. The problem is that I've two different thresholds: (1) For one gap/overlap it's 5ms. (2) For all gaps/overlaps combined it's 40ms. The small threshold for one gap/overlap is not problematic because even if the timestamps are unstable, they are always fluctuating around the right value. And thanks to 2 pass processing eac3to can clear out most short time fluctuations.

I guess I'll try lowering the threshold in the next build. Maybe to 20ms. And maybe I can also offer a switch to lower it even further. But I'll depend on you guys reporting your results with the thresholds, so that we can find the optimal default value (which would be the lowest possible value with which there are no false alarms with 99.9% of the movies).

Originally Posted by jchappo

I recently ran into a problem decoding DTS-HD to WAV using Sonic Decoder 4.3. When remuxing the WAV file back with the video in M2TS container, TsMuxer reports the WAV has no channel info.

There are two different types of WAV headers. Some programs prefer one type, others prefer the other type. eac3to by default writes the simpler type. It can also write the more complicated one, but currently there's no option for that. Will add that in the next build. It seems that TsMuxer wants to have the more complicated type. That most likely doesn't have anything to do with the garbage problems you're having, though...

Originally Posted by jchappo

When playing the movie back over a popcorn hour A-100, the audio plays fine for about 20-30% of the movie, then it goes all screetchy and is basically garbage. I can't determine what the problem is. How to I test if Sonic is not causing the problem?

By playing the WAV file itself (without the video/movie) on your HPTC or on your popcorn hour.

Originally Posted by itsancho

strange problem with Basic Instinct, here is the log:

Unfortunately the Haali Muxer cannot handle this source file.
It doesn't contain enough seek/recovery points.

There's a bug in the Haali Matroska Muxer which results in a total freeze when muxing some h264 movies. eac3to detects this situation and aborts processing directly before the Haali Matroska Muxer freezes. Currently there's no way to properly mux such a movie with eac3to respectively with Haali's Matroska Muxer. You can use the option "-seekToIFrames" to mux the movie in a different way. This should succeed. However, this comes at a cost: Sometimes when seeking such a movie there are very noticable image artifacts for a short time (max a few seconds). Haali knows about this problem but hasn't fixed it yet, sadly.

Originally Posted by bmnot

Ok here you go: http://www.mediafire.com/?g499bligrz0

end of first m2ts doesn't play, i hope it did right

BTW the end of the file is at 30684420096, while the error is at 30684829696, BUT the file converts to mkv fine when there is no joining.

Thanks, will check this out. If it's a bug in eac3to, it will be fixed in the next build.

Originally Posted by gregt

I seem to have run into another bug as follows: [...] For some reason the main (listed under 2) movie is not converted by default. The main movie file is 38GB so may be that is the reason. When I went into the stream directory and converted the 00011.m2ts file directly it converted correctly.

There's no easy to find information (I know of) in the Blu-Ray disc structure which tells us which is the main movie. Consequently eac3to can only list all playlists sorted by their runtime. The playlist with the longest runtime is listed first cause that's usually the movie. In your case it's not the movie. So you have to do "eac3to UNDERWORLD 2) UNDERWORLD.mkv".

[sundansx]

Madshi,
first, thanks for a great program.
I am working on Serenity HDDVD and it fails to use libav when I specify it with other streams as shown below. It tries to use nero, which I dont have installed. When I run a pass with the audio separately followed by a separate pass to pick up the rest it works fine and uses libav. As a note, this same thing also happened on Beowulf. I dont think it is title dependent.

F:\VideoStore\DVD\Serenity>c:\utils\HDCruncher\eac3to.exe d:\HVDVD_TS 1) 2: chap
ters.txt 3: out.mkv 4: out.ac3 8: subtitle.sup -libav -down6 -640 -24
EVO, 1 video track, 4 audio tracks, 3 subtitle tracks, 1:59:01
"MainMovie"
1: Joined EVO file
2: Chapters, 20 chapters with names
3: VC-1, 1080p24 /1.001 (16:9)
4: E-AC3, English, 5.1 channels, 1536kbit/s, 48khz, dialnorm: -27dB
5: E-AC3, Spanish, 5.1 channels, 768kbit/s, 48khz, dialnorm: -27dB
6: E-AC3, French, 5.1 channels, 768kbit/s, 48khz, dialnorm: -27dB
7: AC3, English, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
   "Director's Commentary"
8: Subtitle, English, "SDH"
9: Subtitle, Spanish
10: Subtitle, French
Creating file "chapters.txt"...
[a04] The Nero decoder doesn't seem to work, will use libav instead.
[v03] Extracting video track number 3...
[v03] Muxing video to Matroska...
[a04] Extracting audio track number 4...
[a04] Removing dialog normalization...
[a04] Decoding with DirectShow (Nero Audio Decoder 2)...
[a04] Getting "Nero Audio Decoder 2" instance failed.
Aborted at file position 32768.

[Beastie Boy]

The -libav switch should be used in the audio 'group of the command line, so your command lines becomes:

F:\VideoStore\DVD\Serenity>c:\utils\HDCruncher\eac3to.exe d:\HVDVD_TS 1) 2: chap
ters.txt 3: out.mkv 4: out.ac3 -libav 8: subtitle.sup

Note that -640 is not required as that is the default.

Cheers, Beastie

[monohouse]

hi, I was wondering, is there a way to use the libav/ffmpeg TrueHD decoder in real time ? as in decoding the stream as it being played ? not for reconversion purpose ?

I have demuxed the truehd track using evodemux and now wondering how to replace the FLAC I created with the truehd track that is present, is there a program that is used to do this or evodemux can do this also ? I seemed to notice that eac3to appears to be able to mux it as well, but I was unable to find the manual for eac3to.

I have a few questions: why doesn't eac3to tell the original bit-depth of the stream ? only resolution and channels, and can this stream be directly muxed with a video without additional operations, I understand that the use of FLAC in video is new, so I am not entirely sure if any additional procedures are required, are there any guides that explain how to handle evo+truehd ?

[itsancho]

Originally Posted by madshi

There's a bug in the Haali Matroska Muxer which results in a total freeze when muxing some h264 movies. eac3to detects this situation and aborts processing directly before the Haali Matroska Muxer freezes. Currently there's no way to properly mux such a movie with eac3to respectively with Haali's Matroska Muxer. You can use the option "-seekToIFrames" to mux the movie in a different way. This should succeed. However, this comes at a cost: Sometimes when seeking such a movie there are very noticable image artifacts for a short time (max a few seconds). Haali knows about this problem but hasn't fixed it yet, sadly.

BIG, BIG 10x madshi!!! Everything is absolutely OK and btw there is no any "seeking artifacts"!