eac3to - audio conversion tool

[madshi]

Madshi, what's that error?
[libav] Lossless check failed - expected 0, calculated 69
Looks like a bug in the libav TrueHD decoder. Could you provide me with a little sample which reproduces this bug?

I'm having the same problem on 300 (Blu-ray).

eac3to 300.thd 300.ac3

This should extract ac3 from TrueHD, right? Extracted ac3 has really low sound level (usually my receiver volume is set at -40dB, with this ac3 I have to tweak it atleast to -20dB).
How are you transporting the AC3 to your receiver? Via SPDIF? As AC3 bitstream or decoded as PCM? I've been told that some cheaper receivers cannot fully handle 640kbps AC3 streams. DVDs are limited to 448kbps. Please try decoding the AC3 stream on your PC. Does it sound ok there?

[madshi]

eac3to v2.18 released

http://madshi.net/eac3to.zip

* fixed bug which stopped eac3to v2.15-17 from working on some PCs
* fixed h264 bitstream parsing bug (Sum of all Fears)
* fps value is added to MKV header now
* relaxed VC-1 gap detection once more
* TrueHD decoding to stdout fixed (always output as 24 bit now)

[BlackJack1]

Never mind :) Now all is ok.

[nautilus7]

I think a sample would help.If i cut the evo with a hex editor, eac3to (haali media splitter) doesn't accept it. So i will upload the evo as it is (~70 MB). EDIT: Sample is up here! (http://www.sendspace.com/file/ksotyc)

Which EVO splitter are you using that gives you smooth playback?I really couldn't find that info.:eek: But i can give you the decoder's (sonic) video in pin info:

CLSID: {D7D50E8D-DD72-43C2-8587-A0C197D837D2}
Filter: Sonic Cinemaster® Video Decoder 4.3.0

- Connected to:

CLSID: {1365BE7A-C86A-473C-9A41-C0A6E82C9FA3}
Filter: C:\Tools\Trailer.EVO
Pin: Video

- Connection media type:

Video: MPEG2 Video 720x480 (16:9) 29.97fps 9800Kbps

AM_MEDIA_TYPE:
majortype: MEDIATYPE_Video {73646976-0000-0010-8000-00AA00389B71}
subtype: MEDIASUBTYPE_MPEG2_VIDEO {E06D8026-DB46-11CF-B4D1-00805F6CBBEA}
formattype: FORMAT_MPEG2_VIDEO {E06D80E3-DB46-11CF-B4D1-00805F6CBBEA}
bFixedSizeSamples: 1
bTemporalCompression: 0
lSampleSize: 1
cbFormat: 282

VIDEOINFOHEADER:
rcSource: (0,0)-(0,0)
rcTarget: (0,0)-(0,0)
dwBitRate: 9800000
dwBitErrorRate: 0
AvgTimePerFrame: 333666

VIDEOINFOHEADER2:
dwInterlaceFlags: 0x00000000
dwCopyProtectFlags: 0x00000000
dwPictAspectRatioX: 16
dwPictAspectRatioY: 9
dwControlFlags: 0x00000000
dwReserved2: 0x00000000

MPEG2VIDEOINFO:
dwStartTimeCode: 0
cbSequenceHeader: 150
dwProfile: 0x00000004
dwLevel: 0x00000008
dwFlags: 0x00000000

BITMAPINFOHEADER:
biSize: 40
biWidth: 720
biHeight: 480
biPlanes: 0
biBitCount: 0
biCompression: 0
biSizeImage: 0
biXPelsPerMeter: 0
biYPelsPerMeter: 0
biClrUsed: 0
biClrImportant: 0

pbFormat:
0000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
0010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
0020: 40 89 95 00 00 00 00 00 62 17 05 00 00 00 00 00 @‰•.....b.......
0030: 00 00 00 00 00 00 00 00 10 00 00 00 09 00 00 00 ................
0040: 00 00 00 00 00 00 00 00 28 00 00 00 d0 02 00 00 ........(...Π...
0050: e0 01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ΰ...............
0060: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
0070: 00 00 00 00 96 00 00 00 04 00 00 00 08 00 00 00 ....–...........
0080: 00 00 00 00|00 00 01 b3 2d 01 e0 34 17 ed 23 82 .......³-.ΰ4.ν#‚
0090: 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 ................
00a0: 10 10 10 10 10 10 10 10 10 10 10 10 12 12 12 10 ................
00b0: 10 10 10 10 10 10 10 10 12 12 12 14 12 12 12 12 ................
00c0: 12 14 14 14 14 14 18 18 18 18 1c 1c 1e 22 22 2b .............""+
00d0: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
00e0: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
00f0: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
0100: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
0110: 00 00 01 b5 14 82 00 01 00 00 ...µ.‚....



Error renaming MKV file.
MKV file was successfully created, but the timecodes were not rewritten.
I've tried and cannot reproduce this problem at all, except if I abort timestamp rewriting and don't delete the "trailer.old.mkv" file. In that case exactly the problem you reported occurs. Please make sure that there's no file named "trailer.old.mkv" and then retry. It should work just fine then.Nevermind about this. I found out. It's haali media splitter shell extension which tries to access the file to show its info in the explorer (yes i had the folder where the file was saved opened all the time :rolleyes: )

eac3to v2.18 released

* fixed bug which stopped eac3to v2.15-17 from working on some PCs

Thanks for new version.

Is this the bug that the eac3to -test was not working? What was wrong?

[Thunderbolt8]

Can't you use the MS VC-1 decoder and then still use ffdshow for sharpening?
you can only choose between libavcoded or wmv9 for VC-1 decoding in ffdshow, so I guess its not the same as the VC-1 decoder does.

[nautilus7]

I believe madshi was thinking of using ffdshow as image processor after the MS VC-1 decoder.

EDIT: It should be easy to do it if you enable ffdshow for raw video from the codecs settings.

[Thunderbolt8]

a friend of mine has problems with remuxing lucky number slevin, all newer eac3to versions (2.15-2.18) crash for him. since he cant post here yet, im posting the bugreport for him:
http://www.sendspace.com/file/qkea81

just taking a guess, maybe its because of this: 2: h264/AVC, 1080p13500 ?

[Thunderbolt8]

btw. I asked in the ffdshow thread to add VC-1 libavcodec support for your muxes, but they'd rather see you implementing ffdshow support on your end :S

[madshi]

Nevermind about this. I found out. It's haali media splitter shell extension which tries to access the file to show its info in the explorer (yes i had the folder where the file was saved open all the time :rolleyes: )
Ah well, I'm always disabling this shell extension because it drove me nuts.

Is this the bug that the eac3to -test was not working? What was wrong?
A bug in some internal hooking code.

[madshi]

a friend of mine has problems with remuxing lucky number slevin, all newer eac3to versions (2.15-2.18) crash for him. since he cant post here yet, im posting the bugreport for him:
http://www.sendspace.com/file/qkea81

just taking a guess, maybe its because of this: 2: h264/AVC, 1080p13500 ?
Yeah, that 1080p13500 looks very fishy to me. The crash occurs in a thread created by the Haali Media Splitter. I guess that the Haali Media Splitter was confused by that framerate, too. Maybe a bad rip? Ask your friend to rerip. If the problem doesn't go away then, I'd be happy about a sample. A small one (30MB) should suffice in this case.

[madshi]

btw. I asked in the ffdshow thread to add VC-1 libavcodec support for your muxes, but they'd rather see you implementing ffdshow support on your end :S
Heh, those lazy bastards (just joking)... :)

Let's first find out why it doesn't work then we can still decide who has to fix the problem. Anway, I cannot reproduce the problem?! I've tried with your Hot Fuzz sample and it doesn't matter whether I mux it with gdsmux or with eac3to v2.18. It also doesn't matter whether I ask ffdshow to decode with libavcodec or with wmv, it always works just fine for me (MPC HC, XPSP2, VMR9, ATI).

[Thunderbolt8]

with v2.18 there is minimal improvement, meaning the video of the hot fuzz sample plays now, but only to ~0.5 seconds and then automatically jumps back to the beginning of the stream and it begins again and repeat :S

could it maybe a problem of haali renderer, haali splitter, ffdshow or somehow of this combination?

[Thunderbolt8]

btw. ive got another confirmation for rainbow frames on the Babel remux from someone else (wedding scenes in mexico). can anyone else confirm this?

[madshi]

with v2.18 there is minimal improvement, meaning the video of the hot fuzz sample plays now, but only to ~0.5 seconds and then automatically jumps back to the beginning of the stream and it begins again and repeat :S
Ah, my fault. I thought I'd be testing with ffdshow, but actually ffdshow totally refused the connection to the Haali Splitter, so MPC was again using the MS VC-1 decoder... :rolleyes:

Anyway, I think I know why it's failing to work. I'm reporting the VC-1 track slightly different than gdsmux muxed VC-1 MKVs do. This is necessary to make some VC-1 movies work without garbage (e.g. POTO USA). But ffdshow doesn't seem to like this slightly different reporting. I'll post in the ffdshow thread about this. It's definitely their task to fix, not mine.

[madshi]

btw. ive got another confirmation for rainbow frames on the Babel remux from someone else (wedding scenes in mexico). can anyone else confirm this?
You're VERY short on details. Is it a HD DVD or Blu-Ray? Is it VC-1 or h264 or mpeg2? Which tools did you use for remuxing? Which decoders did you use? Etc etc...

[madshi]

Here's a problem i have (v2.17)

eac3to trailer.evo trailer.mkv
EVO, 1 video track, 1 audio track, 0:01:54
1: MPEG2, 480i60 /1.001
2: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB

The image in resulting mkv stutters a lot. Why? How mpeg2 video is treated by eac3to?
Thanks for the sample. It's a good one. It shows the stuttering quite clearly in the scenes with slow camera pans.

The problem is caused by the Haali Media Splitter during the EVO->MKV remuxing process, as far as I can say. The problem should be fixed as soon as eac3to replaces the Haali Media Splitter with its own internal splitter for MPEG2 (same thing I've already done for VC-1). For now you can work around the problem by demuxing the MPEG2 track with eac3to and then throwing the resulting "video.m2v" track into mkvtoolnix.

I don't really understand why rewriting the timestamps with mkvtoolnix doesn't solve the problem, though! In theory it should! But on the other hand this situation confirms that my decision to replace Haali's Media Splitter with my own for remuxing was a good decision after all.

[nautilus7]

Ah, my fault. I thought I'd be testing with ffdshow, but actually ffdshow totally refused the connection to the Haali Splitter, so MPC was again using the MS VC-1 decoder... :rolleyes:

Anyway, I think I know why it's failing to work. I'm reporting the VC-1 track slightly different than gdsmux muxed VC-1 MKVs do. This is necessary to make some VC-1 movies work without garbage (e.g. POTO USA). But ffdshow doesn't seem to like this slightly different reporting. I'll post in the ffdshow thread about this. It's definitely their task to fix, not mine.
I was about to report this.

While all vc-1 remuxes i made with prior versions of eac3to (haali splitter) decode fine with ffdshow, the remuxes made with new versions (your splitter) can't decode with anything except ms vc-1 decoder.


Btw. I have hotfuzz. Would you like me testing?

[nautilus7]

Thanks for the sample. It's a good one. It shows the stuttering quite clearly in the scenes with slow camera pans.

The problem is caused by the Haali Media Splitter during the EVO->MKV remuxing process, as far as I can say. The problem should be fixed as soon as eac3to replaces the Haali Media Splitter with its own internal splitter for MPEG2 (same thing I've already done for VC-1). For now you can work around the problem by demuxing the MPEG2 track with eac3to and then throwing the resulting "video.m2v" track into mkvtoolnix.

I don't really understand why rewriting the timestamps with mkvtoolnix doesn't solve the problem, though! In theory it should! But on the other hand this situation confirms that my decision to replace Haali's Media Splitter with my own for remuxing was a good decision after all.

Yes it's a good sample! The words/titles go up and down like crazy. :p

Thanks for your alternate method to make this working correctly, but i don't want to do any conversion. Just testing things. :)

[Thunderbolt8]

You're VERY short on details. Is it a HD DVD or Blu-Ray? Is it VC-1 or h264 or mpeg2? Which tools did you use for remuxing? Which decoders did you use? Etc etc...
hd dvd, h264,
ffdshow, libavcodec, used eac3to to remux with the version which was up to date some days after christmas (24th)
what tools/filters the other person used, I dont know, have to ask :S

btw. the lucky number slevin sample: http://www.sendspace.com/file/sj2297

[madshi]

While all vc-1 remuxes i made with prior versions of eac3to (haali splitter) decode fine with ffdshow, the remuxes made with new versions (your splitter) can't decode with anything except ms vc-1 decoder.
The Sonic VC-1 decoder works for me, too.

ffdshow not working is something that ffdshow needs to fix. I've posted the information about what they need to change to their thread. Hopefully they will accept the change.

Btw. I have hotfuzz. Would you like me testing?
That's not necessary, thanks.

[madshi]

hd dvd, h264,
ffdshow, libavcodec, used eac3to to remux with the version which was up to date some days after christmas (24th)
what tools/filters the other person used, I dont know, have to ask :S
Could it be a bad rip again?

[nautilus7]

ffdshow not working is something that ffdshow needs to fix. I've posted the information about what they need to change to their thread. Hopefully they will accept the change.Yes, i belive they should fix. Since the "official" Microsoft vc-1 decoder plays ok, your splitter is ok. Just what a guy doesn't know coding thinks. :)

[Thunderbolt8]

Could it be a bad rip again?
frankly said dont know, but afaik this time I didnt need to use delaycut to fix the audio, as it is often an indicator for a ripping problem.
the other guy used "eac3to 2.14 with newest mkvmerge beta (from 08), ffdshow rev1763 clsid in MPC with libavcodec and haali 1.7.401.3"

btw. there are also rainbow frames at the beginning of the 1st band of brothers episode (also confirmed by others), which are only present with ffdshow and for example not when using coreavc. so maybe this could also be a problem ffdshow since not all h264 features are yet implemented there and no problem of remuxing.

[madshi]

frankly said dont know, but afaik this time I didnt need to use delaycut to fix the audio
You? Why you? I thought someone else reported about Babel and you didn't test it yourself? You're confusing me...

[Richum]

I'm not fully sure about that. I cannot really imagine that a bad disc is causing this. But in the end I'm not really sure...

I have been trying different things to determine if the disc itself can be at fault. I installed the Demo of DVDFab Platinum, it would not even recognize my Transformer HD DVD or my friends copy.

I fed it 18 different HD DVDs and it acknowledged 15 as being present and 3 would not even show up as inserted. I was able to rip 12 monkeys and when I tried it in eac3to, I got the same timing errors as previous. I am putting AnyDVD back on the machine and will try v2.18 and see if I can make Transformers go, as I now have a different disc to try.

[madshi]

a friend of mine has problems with remuxing lucky number slevin, all newer eac3to versions (2.15-2.18) crash for him. since he cant post here yet, im posting the bugreport for him:
http://www.sendspace.com/file/qkea81

just taking a guess, maybe its because of this: 2: h264/AVC, 1080p13500 ?
The sample seems to be alright, no damage/corruption, as far as I can see. The framerate information is stored in a strange format, which confused eac3to. This will be fixed in the next build. For me the sample didn't crash with v2.18, though. The sample converted just fine. The fps value in the MKV header was totally wrong, of course (since eac3to misinterpreted the framerate information). But still the MKV remuxed fine and plays fine. Not sure where the crash comes from on your friend's PC. He should retest with the next eac3to build (maybe released on Sunday, don't know). If there's still a crash he should redownload and reinstall the latest Haali version cause that's what's crashing for him.

[madshi]

I have been trying different things to determine if the disc itself can be at fault. I installed the Demo of DVDFab Platinum, it would not even recognize my Transformer HD DVD or my friends copy.

I fed it 18 different HD DVDs and it acknowledged 15 as being present and 3 would not even show up as inserted. I was able to rip 12 monkeys and when I tried it in eac3to, I got the same timing errors as previous. I am putting AnyDVD back on the machine and will try v2.18 and see if I can make Transformers go, as I now have a different disc to try.
Sounds strange. I don't know such problems. Maybe your HD DVD drive is in a bad state? I'd also suggest checking your playback chain. Maybe the remuxes are fine and just the DirectShow filters you're using screw up the audio sync? Can't really say from here...

[rickardk]

Thunderbolt8
Can you give me the timecode for the rainbow frame problem in Babel?

I have that one remuxed and if there is a problem I can give it a new try with latest eac3to.

[Thunderbolt8]

You? Why you? I thought someone else reported about Babel and you didn't test it yourself? You're confusing me...
of course I also did it myself, I just said I am quite sure I didnt have to use delaycut for that one, though I cant remember 100% any more.

[Thunderbolt8]

Thunderbolt8
Can you give me the timecode for the rainbow frame problem in Babel?

I have that one remuxed and if there is a problem I can give it a new try with latest eac3to.
sorry, but I cant access the movie on my HD right now, I just remember it was at least once during that mexican wedding trip from that nanny and in few other situations. though there werent too many of those occasions in that movie.

[Richum]

Sounds strange. I don't know such problems. Maybe your HD DVD drive is in a bad state? I'd also suggest checking your playback chain. Maybe the remuxes are fine and just the DirectShow filters you're using screw up the audio sync? Can't really say from here...

The HD DVD drive I have is new and it has ripped, and also can play Transformers with PowerDVD, so I suspect it is OK. However one never knows.


It could be the filters, I have a free program from River Past called direct show detective, it polls the machine collects a list of what is installed. I could run it and post the results here if it would help. I wouldn't know what I was looking at, sorry.

[rickardk]

sorry, but I cant access the movie on my HD right now, I just remember it was at least once during that mexican wedding trip from that nanny and in few other situations. though there werent too many of those occasions in that movie.


Just watched the whole wedding scene. No problem. Did you rewrite the timestamps with mkvmerge?
mkvmerge introduces rainbow frames on some tiltes

[esoteradactyl]

when converting DTS HD to flac, do the audio settings for the sonic decoder need to be set to anything specific?

[Yraen]

Thanks!

Could you please retry with v2.18? I think those gap/overlap complaints should be solved in that version.

No gaps or overlaps present in The Matrix with v2.18, with or without pulldown. I'll set the other two movies of the trilogy to extracting overnight, but I doubt anything will appear there. I did notice that the -stripPulldown ran 23 minutes faster this time. :thanks:

[Thunderbolt8]

Just watched the whole wedding scene. No problem. Did you rewrite the timestamps with mkvmerge?
mkvmerge introduces rainbow frames on some tiltes
yes, I always do this, because of adding subs and adding info to video and audio stream.
so you mean you watched only that wedding part itself? try again with the whole mexican part, when they arrive there please

[intomed]

Madshi,

Bad news. I just did a fresh re-rip of POTO (US HDDVD version). Followed by EVO-->MKV (via the following command line: eac3to pevob_1+pevob_2 movie.mkv). When completed played the subsequent MKV file in MPC (no sound) and still had major macroblocking/rainbow effect. Is there anything else you would like me to try?:confused:

[bertox]

Hi. Thanks for this app Madshi, first!! Saves me a lot of time.
But...i encountered some problems (to me, particularly).
I'm using eac3to with eac3togui to decode my DVD-Audio MLP files to Wav for two reasons: PowerDVD (the only player who reads DVD-A) is not using KernelStreaming, what i use in Foobar2000 for better sound quality; and the downmixes options of PowerDVD are very bad to me, and i need them since i don't have a multispeaker system, only stereo.

I like to see downmix option in eac3to..., but...why ProLogic II???? Seems to be horrible for me.
Why not a simple sum of channels?
Could you add a more flexible Downmix option like a independent volume change in every channel? Or minimum a sum of channels and an 8 or 9 db attenuate option (for all channels at same time for saturating mix avoid)?????

Something like this:
FL' = FL + C + LFE + SL
FR' = FR + C + LFE + SR

plus -9db (attenuate) for every single channel...or better...independent control volumes for such one!!! This part is important to avoid mix clipping.

This is what i want. No rare downmixing like prologic who sounds strange to my ears...

I don't want to use SoundForge anymore for downmixing stuff...

Another thing, for the Eac3toGui author. Can you add for source MLP files???
I'm really tired of re-naming my MLP files to THD or DTSHD to be accepted by this Gui!!!!!

Thanks. Víctor, from Argentina (sorry my bad english..):D

[intomed]

Oh...by the way Madshi, the version of eac3to that I used was 2.18.

[madshi]

It could be the filters, I have a free program from River Past called direct show detective, it polls the machine collects a list of what is installed. I could run it and post the results here if it would help. I wouldn't know what I was looking at, sorry.
It's not very important which filters you have installed. It's much more important which filters you're actually using for playback. So that list of installed filters wouldn't really help.

[madshi]

yes, I always do this, because of adding subs and adding info to video and audio stream.
so you mean you watched only that wedding part itself? try again with the whole mexican part, when they arrive there please
You don't need to (and probably shouldn't) rewrite timestamps for VC-1 movies, anymore, when using eac3to v2.18 or newer. Of course you can still add audio tracks and subtitles. But no timestamp rewriting for VC-1, anymore, please. It will most probably not improve things, the chance is bigger than it will screw things up. With eac3to v2.18 the timestamps should be as good as possibly can. Plus the fps is added to the MKV header now, too. So there's no benefit whatsoever that timestamp rewriting could bring, as far as I can see. Of course things are different for MPEG2 and h264 movies right now cause I'm still using the Haali splitter for those which is giving out funny timestamps sometimes...

[madshi]

Bad news. I just did a fresh re-rip of POTO (US HDDVD version). Followed by EVO-->MKV (via the following command line: eac3to pevob_1+pevob_2 movie.mkv). When completed played the subsequent MKV file in MPC (no sound) and still had major macroblocking/rainbow effect. Is there anything else you would like me to try?:confused:
I'm said to hear that. I guess (hope) the image is screwed up right from the start? If so, could you provide me with a very little sample (e.g. 30-50MB)? That would be most helpful. Thanks!

[madshi]

But...i encountered some problems (to me, particularly).
I'm using eac3to with eac3togui to decode my DVD-Audio MLP files to Wav for two reasons: PowerDVD (the only player who reads DVD-A) is not using KernelStreaming, what i use in Foobar2000 for better sound quality; and the downmixes options of PowerDVD are very bad to me, and i need them since i don't have a multispeaker system, only stereo.
Why don't you use FLAC? Foobar2000 has very good FLAC decoding built in and it's lossless. So you will save space without losing even a single bit of audio quality.

I like to see downmix option in eac3to..., but...why ProLogic II????
Because eac3to's main target is home cinema. And ProLogic II is the best choice we have for that purpose.

Why not a simple sum of channels?
Because no receiver on earth can properly reconstruct the channels from such a simple sum of channels.

Could you add a more flexible Downmix option like a independent volume change in every channel? Or minimum a sum of channels and an 8 or 9 db attenuate option (for all channels at same time for saturating mix avoid)?????

Something like this:
FL' = FL + C + LFE + SL
FR' = FR + C + LFE + SR

plus -9db (attenuate) for every single channel...or better...independent control volumes for such one!!! This part is important to avoid mix clipping.

This is what i want. No rare downmixing like prologic who sounds strange to my ears...

I don't want to use SoundForge anymore for downmixing stuff...
The programming of a different mixing matrix is not a big problem. But keeping the usage of eac3to as simple as possible is. Now offering a volume control for every single channel would require me to add at least 8 new parameters which would dramatically increase the size of the help text. If eac3to was a GUI it would be much easier to add such options without making things more complicated. But adding another 100 options to a console program really does damage ease of use. Because of that reason I'm trying to avoid adding shitloads of new options, unless it's really very important. So I'm sorry to say, but at this point in time I don't like the idea of adding a multitude of new options for one single user... ;)

[madshi]

when converting DTS HD to flac, do the audio settings for the sonic decoder need to be set to anything specific?
Nope.

[Thunderbolt8]

You don't need to (and probably shouldn't) rewrite timestamps for VC-1 movies, anymore, when using eac3to v2.18 or newer. Of course you can still add audio tracks and subtitles. But no timestamp rewriting for VC-1, anymore, please. It will most probably not improve things, the chance is bigger than it will screw things up. With eac3to v2.18 the timestamps should be as good as possibly can. Plus the fps is added to the MKV header now, too. So there's no benefit whatsoever that timestamp rewriting could bring, as far as I can see. Of course things are different for MPEG2 and h264 movies right now cause I'm still using the Haali splitter for those which is giving out funny timestamps sometimes...
I still will need another remux with mkvmerge, because of adding subs & adding video and audio stream info. so even if before everything should be ok already, as long as theres no other tool which can also do this, my final mux will always be with mkvmerge.

[madshi]

I still will need another remux with mkvmerge, because of adding subs & adding video and audio stream info. so even if before everything should be ok already, as long as theres no other tool which can also do this, my final mux will always be with mkvmerge.
Please reread my post. I clearly said: "Of course you can still add audio tracks and subtitles". But you should not rewrite timestamps. Do you understand the difference? You can add audio tracks and subtitles with mkvtoolnix without rewriting the timestamps, you know?

[madshi]

Madshi,

Bad news. I just did a fresh re-rip of POTO (US HDDVD version). Followed by EVO-->MKV (via the following command line: eac3to pevob_1+pevob_2 movie.mkv). When completed played the subsequent MKV file in MPC (no sound) and still had major macroblocking/rainbow effect. Is there anything else you would like me to try?:confused:
Wait a moment, I think it's my fault. Could you please retry with v2.19? If it still fails with v2.19, getting a sample would be great, though.

[madshi]

eac3to v2.19 released

http://madshi.net/eac3to.zip

* fixed h264 bitstream parsing of framerate information format
* fixed (again) muxing of some rare VC-1 titles like e.g. POTO USA

[nautilus7]

Thanks, again, for the update.

* Regarding POTO, you didn't ask intomed what renderer is using...

* Regarding downmix... It's fine i think. But you could also add upmix with dolby prologic II (any source --> 5.1) or/and dolby prologic IIx (any source --> 7.1).

[madshi]

* Regarding POTO, you didn't ask intomed what renderer is using...
True. But since I know that this movie has always made problems with the MS VC-1 decoder I was quite sure he was using that. intomed?

* Regarding downmix... It's fine i think. But you could also add upmix with dolby prologic II (any source --> 5.1) or/and dolby prologic IIx (any source --> 7.1).
Downmixing is quite easy to do. Upmixing is much more complicated, as far as I know. I don't know how to do this properly, so I better don't even try. Furthermore I've more important things to do like e.g. adding my own splitter for MPEG2 and h264 movies to get perfect timestamps for those, too. And finally adding Blu-Ray support sooner or later (not too soon, though).

[rickardk]

Please reread my post. I clearly said: "Of course you can still add audio tracks and subtitles". But you should not rewrite timestamps. Do you understand the difference? You can add audio tracks and subtitles with mkvtoolnix without rewriting the timestamps, you know?

IIn my experience some titles can be screwed up (rainbow frames) just by running the mkv file through mkvmerge (even without rewriting the timestamps). I have not found thoose rainbow frames in Babel HD DVD (no time to watch the whole file yet) but I have encountered the same problem with some Blu-ray titles.

Today I will redo all titles with problems I talked about before.

Writing a detailed report on problems for you.

Should I use both 2.19 and 2.14? Or is it enough to count on 2.19?