eac3to - audio conversion tool

[nautilus7]

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?Maybe i didn't get it, but i asked about renderer like vmr9 or haali, not decoder.

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).Ok, then. I'll bother you again with this when you 've done with all other things. :p

[madshi]

Maybe i didn't get it, but i asked about renderer like vmr9 or haali, not decoder.
Ah, right.

[madshi]

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.
Probably those Blu-Ray titles are mostly h264? :)

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

Writing a detailed report on problems for you.
Great - looking forward to that! However, I'm a bit afraid about getting a mixture of reports from various different titles. So here's what I'd like to get from you:

(1) Please post a list of movies which were problematic but work perfect with v2.19 now (if there are any such).
(2) Please report one title per post which still shows problems. So if you have 5 VC-1 movies with problems, please make 5 separate posts, one for each movie. That would help greatly separating the different movies and issues.

Please check out the VC-1 titles first cause VC-1 is where I think eac3to should be very good with now. MPEG2 and h264 handling still has room to improve in future builds. So if you ask me you can leave out MPEG2 and h264 titles completely for now. Let's care about them at a later time. VC-1 is my focus right now. Once VC-1 handling is as good as I can get it I'll begin on improving MPEG2 and h264.

Should I use both 2.19 and 2.14? Or is it enough to count on 2.19?
(1) Please test with v2.19 first.
(2) If the result seems to be perfect, please add an audio track with mkvtoolnix and retest.
(3) If the result still seems to be perfect, that's all you need to do and you can stop here at step (3).
---------
(4) If you find problems, please rerip the movie to make sure that the rip is clean.
(5) Demux the E-AC3/AC3 audio tracks. During demuxing v2.19 will automatically check whether there's any corruption in the audio data. That's very effective as a corruption test because E(-AC3) is protected by a CRC. If audio demuxing runs through without any complaints, there's a good chance that the rip is (relatively) clean.
(6) If the rip is clean, but you still run into any problems, as a first step please post the v2.19 processing log (see "log.txt" in the "eac3to.exe" folder) together with a detailed problem description and with a summary of your playback system (OS, decoders, renderer) here in this thread. You don't necessarily need to do any further test in the first step. I'll check your report and maybe ask you about further tests then.

Thanks much!! :)

[rickardk]

Yes I mentioned thoose titles because Babel HD DVD is also H264. Maybe mkvmerge have problems with a couple of H264 titles.


Ok, I will start with VC-1 titles!

[madshi]

Ok, I will start with VC-1 titles!
Roll on, mate! :D

Just to be safe: Please do NOT rewrite timestamps with mkvtoolnix.

[Kumo]

i'm converting a lpcm 5.1 ntsc bluray track to a dts pal track using ea3to2.17.a strange issue happened to me.if i convert directly lpcm->dts, it works fine even with "-speedup" command applyed(the resulting dts file is ok).if i convert to wavs and then manually to dts via surcodedvd, the resulting dts file is double-sized,the lenght is twice the original, and it plays 2 times faster(it doesn't happen if speedup command isn't applyied).i tryed many times, and using raw imputs demuxed in different ways. there's any issue with the speedup feature?how does it exactly work?

[madshi]

i'm converting a lpcm 5.1 ntsc bluray track to a dts pal track using ea3to2.17.a strange issue happened to me.if i convert directly lpcm->dts, it works fine even with "-speedup" command applyed(the resulting dts file is ok).if i convert to wavs and then manually to dts via surcodedvd, the resulting dts file is double-sized,the lenght is twice the original, and it plays 2 times faster(it doesn't happen if speedup command isn't applyied).i tryed many times, and using raw imputs demuxed in different ways. there's any issue with the speedup feature?how does it exactly work?
Please run that manually created dts file through eac3to ("eac3to source.dts dest.dts"). There's probably padding in it. eac3to will remove that for you. Maybe that fixes the problem. Or not.

[bertox]

Yes, i'm using Flac to store my good music actually...
Can you add the minimum downmix option then?? Sum all with 9db attenuate in the mix, avoiding clipping.

Please, please,please,please,please,please!!!. jejeje. Sorry

It's for people who are not using receivers, but only sound cards in pcs.
Thanks.

[nautilus7]

Yes, i'm using Flac to store my good music actually...
Can you add the minimum downmix option then?? Sum all with 9db attenuate in the mix, avoiding clipping.

Please, please,please,please,please,please!!!. jejeje. Sorry

It's for people who are not using receivers, but only sound cards in pcs.
Thanks.
Don't you have a 5.1 sound card?

[bertox]

No. A card with REAL 24 bits 96khz 5.1 is very expensive to me and for the mostly pc users. My cheapest Chaintech AV-710 SoundCard can manage 24/96-192 but only in stereo mode with she's Wolfson fabulous dac. And....i have not a multispeaker system, only an stereo amplifier with a pair of speakers.
------------------
The programming of a different mixing matrix is not a big problem.
You say that...this is what i want
Bye.

[Thunderbolt8]

so which movies (HD DVD) are regarded as having problems when being remuxed so far (without timestamps rewriting)?

the searchers? more, were all others due to bad rips?

[madshi]

Can you add the minimum downmix option then?? Sum all with 9db attenuate in the mix, avoiding clipping.
Maybe. But not very soon. Need to get video remuxing done properly first.

[madshi]

so which movies (HD DVD) are regarded as having problems when being remuxed so far (without timestamps rewriting)?
We don't know yet for sure. Some movies used to have problems but should work now (I hope POTO USA will be one of those). Other movies had problems with stuttering. I hope that the latest eac3to version will sort that out once and for all (rickardk is testing that as we speak). Some movies have problems due to either bad rips or to authoring faults. I rather guess that eac3to will never handle those well. Right now I'm not aware of any specific title which is proven to be a clean rip (by repeating the ripping etc) and which still makes problems. But let's wait and see.

The above is about VC-1 HD DVD movies only. I'll deal with MPEG2 and h264 HD DVD movies in the next step. Let's not even talk about Blu-Ray right now. Although: For VC-1 Blu-Ray movies you could demux audio and video by using xport and then mux the VC-1 track to MKV by using eac3to v2.19. This should work well and give perfect results, as long as there are no gaps/overlaps in the Blu-Ray video stream (which shouldn't really happen, anyway).

[Thunderbolt8]

I just muxed the starship troopers blu-ray with putting the vc-1 .m2ts into gdsmux and remuxed only the video to .mkv, because I wanted to avoid the demuxing stage of the video, as you prefer it. so do you think it will be different (and better) when I use xport to demux the vc-1 video and use eac3to to mux it to .mkv?

btw. is a .mkv file with video only in it affected regarding its playback structure, when I put it into mkvmerge to add audio, even when I dont rewrite the timestamps, so could it happen it has no problems as .mkv with video stream only, but then shows errors of any kind after the addition of audio?

[rickardk]

It takes some time....I'll have a list up tomorrow I hope.

Can someone please try Million Dollar Baby again (someone tried that one a couple of pages ago).

I'm pretty sure I get a clean rip. But not 100% sure.
Please try to use eac3to and try the result without using mkvmerge to add audio.

Someone should also try Unforgiven HD DVD. That one should suffer from the same problem as Million Dollar Baby.

I just have the Blu-ray version of Unforgiven (that I did in some way remux without problem before) so can't test Unforgiven HD DVD.

[bertox]

Ok. I wait. Thanks for fast reply.:D

[madshi]

I just muxed the starship troopers blu-ray with putting the vc-1 .m2ts into gdsmux and remuxed only the video to .mkv, because I wanted to avoid the demuxing stage of the video, as you prefer it. so do you think it will be different (and better) when I use xport to demux the vc-1 video and use eac3to to mux it to .mkv?
Normally a direct remux is better because doing it that way when there are gaps/overlaps in the video stream these are not lost, so you keep perfect audio sync. However, if the video stream doesn't have any gaps/overlaps (which should theoretically be the normal case), demuxing and remuxing should work just fine, too. At least with MPEG2 and VC-1. I'm somewhat sceptical about doing this with h264. Had bad experiences with that in the past. Didn't have a problem with demuxing/remuxing MPEG2 and VC-1 yet. But still, if you do that, make sure that you double check audio sync both in the beginning and end of the movie.

btw. is a .mkv file with video only in it affected regarding its playback structure, when I put it into mkvmerge to add audio, even when I dont rewrite the timestamps, so could it happen it has no problems as .mkv with video stream only, but then shows errors of any kind after the addition of audio?
I don't really see how adding audio could affect the video, but I can't say for sure. Of course there could also be a bug in mkvtoolnix or something like that.

[nautilus7]

Updated version of ffdshow is out. It now plays VC-1 mkv files created by eac3to!

[rickardk]

Sum of all Fears is now working without problems!

[IanD]

The above is about VC-1 HD DVD movies only.
Thanks for such a useful utility.

Have used the current version from your link (2.19?) for the first time in an attempt to strip pulldown flags from VC-1 to see if that improves playback with WMV9 decoder.

However, when I attempted the following (source and dest files are on different drives), I just got a "can not perform this video function" error:

eac3to source.evo 1: dest.vid -stripPulldown

I then tried the following and it seemed to work, but did not honour the destination drive (just put it in the source folder):

eac3to source.evo 1: dest.vid -stripPulldown -demux

Finally, I attempted the following, and it did honour the dest drive but it didn't mention stripping pulldown flags:

eac3to source.evo 1: dest.mkv -stripPulldown

Are there limitations to eac3to, or am I just misunderstanding the syntax?

I also tried using the linked GUI, but it didn't seem to have an option to select the required video track. Adding the video track identifier in the destination field just included it in the quotes for the destination folder, so that didn't work either.

I'm guessing eac3to has some inbuilt assumptions that are not fully explained: for example, an .mkv suffix is interpreted as a requirement to create a matroska output, even though no explicit mkv option is specified. So, I'm wondering if there are others, such as always having to name a video file as .vc1 otherwise it gets rejected in the syntax.

[nautilus7]

What is .vid you use as output file extension? VC-1 video extension is .vc1. Therefore you have to type:

eac3to source.evo dest.vc1 -stripPulldown

to dumux the video only, or

eac3to source.evo dest.mkv -stripPulldown

to mux the video to matroska.

If no message like "removing pulldown flags" is displayed, i would say that no such flags are present in the original video.

[rickardk]

Goodfellas HD DVD (and Blu-ray)

HD DVD:
No problem
fps value added to header


Smooth playback!!!

Blu-ray:

1. Demuxing video with xport
2. Muxing VC-1 with eac3to

eac3to creates a mkv that does not contain fps or default duration in the header. This time ffdshow says Frame duration is 0.0001ms and 25 fps. With eac3to 2.17 it says Frame duration is 39.9999ms and 25 fps.

mkvinfo:
Muxing application: Haali DirectShow Matroska Muxer
Writing application: eac3to at 179
No info about default duration (fps)


eac3to forgets to add fps value to header when muxing raw VC-1 to mkv.

[rickardk]

The Last Samurai

No problems!
Smooth playback..

[rickardk]

Million Dollar Baby

Still getting rainbow frames!
I will try to get another disc tomorrow to rule out bad rip

[Thunderbolt8]

Goodfellas HD DVD (and Blu-ray)

HD DVD:
No problem
fps value added to header


Smooth playback!!!

Blu-ray:

1. Demuxing video with xport
2. Muxing VC-1 with eac3to

eac3to creates a mkv that does not contain fps or default duration in the header. This time ffdshow says Frame duration is 0.0001ms and 25 fps. With eac3to 2.17 it says Frame duration is 39.9999ms and 25 fps.

mkvinfo:
Muxing application: Haali DirectShow Matroska Muxer
Writing application: eac3to at 179
No info about default duration (fps)


eac3to forgets to add fps value to header when muxing raw VC-1 to mkv.
try to put the .m2ts into gdsmux and only remux the video stream to .mkv and see how/this changes the result

[rickardk]

Sum of all Fears is now working without problems!

This is a H264 title and CoreAVC->ffdshow reports 25fps
23.976fps in mkvinfo.


If video is demuxed and then muxed with mkvmerge instead of using eac3to (gdsmux). Then CoreAVC->ffdshow reports 29.97fps

So the way H264 is muxed affects the way CoreAVC->ffdshow reports the framerate

[rickardk]

try to put the .m2ts into gdsmux and only remux the video stream to .mkv and see how/this changes the result

Already tried!
That gave stutters at thoose places as before.

And if I try to mux the raw VC-1 with gdsmux it hangs after 0.4%.

But good news is that remuxing evo with eac3to now produce a perfect result and the blu-ray version can be demuxed with xport and muxed with eac3to to give a perfect result (have to rewrite timecode though as eac3to forgets to add fps value to header when muxing raw VC-1)

[Thunderbolt8]

I tried to apply a delay of -15ms to an ac3 file, but the length stayed the same. I guess its because of those 32ms ac3 frame length, right? it can only change in those steps. I tried to put it into .mka then and apply the delay with mkvmerge, but then it got delayed -32ms (according to mediainfo).

I got that value for audio delay from xport, when I demuxed the stream from the .ts
its a 25fps stream and I wanted to remux to .mkv and using a 23.976 audio track. so I though Ill pulldown the speed of that original 25fps file to 23.976 with eac3to and then compare at which time the audio of these tracks begin so I should have the fitting delay for the NTSC track.
but before pulldown I somehow need to calculate in that -15ms delay, because this value would also be altered slightly by that.
so whats the best solution here?

[intomed]

[QUOTE=madshi;1092258]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?


Please excuse my ignorance, as I am new to this whole world of creating/converting video files. I do not understand what you mean by "which renderer". I try to play the file in MPC. If I go to the "filters" option and click on "renderer", under the Pin Info tab, the first box says something to the effect of "VMR input" or something close to that. Is that what you are asking for?

Once again, I apologize for my ignorance. As Tom Cruise said in the movie 'Days of Thunder': " I would like to help, but I can't. I'm stupid. I don't know the language." LOL :stupid:

[nautilus7]

Yes that's it. You can easily see what renderer is being used by clicking on options --> output in mpc.

[IanD]

What is .vid you use as output file extension? VC-1 video extension is .vc1. Therefore you have to type:

eac3to source.evo dest.vc1 -stripPulldown

to dumux the video only, or

eac3to source.evo dest.mkv -stripPulldown

to mux the video to matroska.

If no message like "removing pulldown flags" is displayed, i would say that no such flags are present in the original video.

It's a little confusing when some utilities ignore the file extension and others rely on a specific extension to tell it how to handle the process: that's the difference between using options to control the process and implicit file extensions to do the same thing; unless you know which is being used, it can be difficult to use a new utility.

There wasn't an example of:

eac3to source.evo dest.vc1 -stripPulldown

so I assumed file extension didn't matter as it isn't transcoding video to another type and so I used .vid. I guess I was wrong. How was I supposed to know?

When I used:

eac3to source.evo dest.??? -stripPulldown -demux

I received the "removing pulldown flags" message, so there are definitely flags in the source stream.

[IanD]

Seeking clarification on the use of eac3to 2.19:

1. The 1st post on page 1 suggests default decoder for TrueHD is libav, yet when I attempt to transcode a TrueHD track, eac3to says that Nero is missing (I don't have it installed). So why isn't eac3to using libav by default?

2. eac3to includes an avcodec.dll file: does this include the libav/ffmpeg routines to handle TrueHD? I think madshi suggests libav TrueHD is good.

3. When using eac3to commandline, there are options to specify the trackno for audio and video streams. How are the trackno referenced? If I run eac3to in -test mode, it will list the video and audio tracks, but the first entry is "combined evo", making the first video track, number 2. In the following simulated list, what would be the trackno for the primary video and TrueHD audio tracks?

1. Combined evo
2. Video 1080p 24/1001
3. Video 480p 30/1001
4. E-AC3
5. TrueHD
6. E-AC3
7. E-AC3

I think the first page of eac3to could do with a bit more explanation. It seems to reference 2.18, yet I understand 2.19 is the latest release.

I'm not sure where in the 152 pages of this thread the answers to my questions might reside (if at all).

[intomed]

Yes that's it. You can easily see what renderer is being used by clicking on options --> output in mpc.

When I check that in MPC, it is marked to use the system defaults. Should I be settting this to something else? All my other mkv's play just fine.

By the way Madshi, I retried doing POTO using version 2.19 and had the same results. I'll try and get you a sample. Just got to go back in this thread and figure out how to do that.

[intomed]

Well I hope this works. Took a 50MB sample of the POTO troublesome mkv. I hope this is what you wanted Madshi and not a sample of the evo file.

http://www.sendspace.com/file/pb6vlm

[kmitalian]

I am loving this MKV with eac3to, but im having problems with VC1 when after i demux it into MKV from eac3to, it stutters?

i dont know if i left something out or am i supposed to add a decoder i dunnow?

[madshi]

Goodfellas HD DVD (and Blu-ray)

HD DVD:
No problem
fps value added to header

Smooth playback!!!
The Last Samurai

No problems!
Smooth playback..
Good results!! :)

BTW, you can combine the "works perfectly" messages into one post to save posts/space. The "one post per movie" rule is only necessary for movies that do not work perfectly yet.

Hmmmmm... Two questions about these two VC-1 titles:

(1) Does ffdshow now correctly show the correct fps and frame duration values?
(2) If you ignore movies with rainbow frames for now (which could be caused by bad rips or by authoring faults), would you say that VC-1 remuxing is now near perfect? Or is there anything which could still be improved?

Will check out why fps is not added when muxing raw VC-1 streams to MKV.

[madshi]

Sum of all Fears is now working without problems!

This is a H264 title and CoreAVC->ffdshow reports 25fps
23.976fps in mkvinfo.

If video is demuxed and then muxed with mkvmerge instead of using eac3to (gdsmux). Then CoreAVC->ffdshow reports 29.97fps

So the way H264 is muxed affects the way CoreAVC->ffdshow reports the framerate
Maybe ffdshow checks the timecode difference between first and 2nd frame or something like that. Don't know. Anyway, as I said, let's look at h264/AVC later. I'm mainly interested in VC-1 for now...

[madshi]

I tried to apply a delay of -15ms to an ac3 file, but the length stayed the same. I guess its because of those 32ms ac3 frame length, right? it can only change in those steps. I tried to put it into .mka then and apply the delay with mkvmerge, but then it got delayed -32ms (according to mediainfo).

I got that value for audio delay from xport, when I demuxed the stream from the .ts
its a 25fps stream and I wanted to remux to .mkv and using a 23.976 audio track. so I though Ill pulldown the speed of that original 25fps file to 23.976 with eac3to and then compare at which time the audio of these tracks begin so I should have the fitting delay for the NTSC track.
but before pulldown I somehow need to calculate in that -15ms delay, because this value would also be altered slightly by that.
so whats the best solution here?
If you want to have the delay as accurate as possible you need to decode and reencode the audio. If you want to slowdown the audio track you'll have to do that anyway, so you should better apply the delay at the same time as you slowdown.

As you correctly guessed, applying a delay on AC3 bitstream works by adding or removing whole AC3 frames, each of which in your case is 32ms long. Consequently eac3to does nothing for -15ms. It would apply a delay for -17ms, though. Don't know about -16ms.

[rickardk]

Good results!! :)

BTW, you can combine the "works perfectly" messages into one post to save posts/space. The "one post per movie" rule is only necessary for movies that do not work perfectly yet.

Hmmmmm... Two questions about these two VC-1 titles:

(1) Does ffdshow now correctly show the correct fps and frame duration values?
(2) If you ignore movies with rainbow frames for now (which could be caused by bad rips or by authoring faults), would you say that VC-1 remuxing is now near perfect? Or is there anything which could still be improved?

Will check out why fps is not added when muxing raw VC-1 streams to MKV.

1) Yes!! fps and default duration is correct.
2) I think so. I just finished to more titles that did not play perfect before:
Mi2 and Dantes Peak

I tested thoose places with micros stutters and they are gone. Plays perfect!


Great job!


Thoose rainbow frames may be caused by bad authoring. But no rainbow frames can be found before muxing into mkv.
I will test a new source of Million Dollar Baby later today!

[madshi]

It's a little confusing when some utilities ignore the file extension and others rely on a specific extension to tell it how to handle the process: that's the difference between using options to control the process and implicit file extensions to do the same thing; unless you know which is being used, it can be difficult to use a new utility.
Not really. Why would you want to use a "vid" file extension? Haven't seen that extension being used anywhere *ever*. Basically it never harms to use the right extensions. So why not doing that.

There wasn't an example of [...]
There are so many combinations of what eac3to can do, depending on source file format etc, I simply cannot add an example of every possible action to the help, unless you want to have 100 pages of help text.

so I assumed file extension didn't matter as it isn't transcoding video to another type and so I used .vid. I guess I was wrong. How was I supposed to know?
By thinking about how eac3to could possibly know what you want it to do. eac3to supports 2 actions for video tracks in EVO containers: You can demux or remux to MKV. Obviously you need to tell eac3to somehow which of those 2 actions you want to have executed. eac3to cannot read your mind.

1. The 1st post on page 1 suggests default decoder for TrueHD is libav, yet when I attempt to transcode a TrueHD track, eac3to says that Nero is missing (I don't have it installed). So why isn't eac3to using libav by default?
Can you please post the full eac3to log? See "log.txt" in the folder where "eac3to.exe" is located.

2. eac3to includes an avcodec.dll file: does this include the libav/ffmpeg routines to handle TrueHD?
Yes. That dll contains decoder routines for TrueHD/MLP, E-AC3, AC3 and DTS.

3. When using eac3to commandline, there are options to specify the trackno for audio and video streams. How are the trackno referenced?
Generally when feeding a container (e.g. EVO, which is currently the only container supported by eac3to) you will want to first do "eac3to source.container" to see which numbers eac3to assigns to the different tracks. Then you can go from there and tell eac3to what to do with which tracks.

In the following simulated list, what would be the trackno for the primary video and TrueHD audio tracks?
There is no flag in the EVO container about which track is the "primary" video or audio track. Generally the first video track listed is usually the primary video track. Of course if one video track is 1080p24 and the other is 480p30 you know for sure which is the main video track. About audio tracks: eac3to can only show the audio tracks sorted per index (the index comes from the EVO container) with detailed information about bitrate etc. It's then your best guess which is the track you want. eac3to cannot possibly know. A first good guess is that the TrueHD track is usually the best track to use. Next best track is usually the E-AC3 track with the highest bitrate.

I think the first page of eac3to could do with a bit more explanation.
If I make the post much longer, nobody will have fun to read it, anymore. That's always the problem: A help must not be too long, or nobody will bother to read it.

[madshi]

I am loving this MKV with eac3to, but im having problems with VC1 when after i demux it into MKV from eac3to, it stutters?

i dont know if i left something out or am i supposed to add a decoder i dunnow?
Demuxing VC1 from MKV is kind of problematic cause during the VC1 MKV muxing process some information was stripped from the VC1 stream which will usually not come back during demuxing. Honestly, I'm wondering why you're getting a picture at all!!

[madshi]

Well I hope this works. Took a 50MB sample of the POTO troublesome mkv. I hope this is what you wanted Madshi and not a sample of the evo file.

http://www.sendspace.com/file/pb6vlm
Argh, too bad. This seems to be a different problem to what I thought. This is from a fresh rip from your HD DVD disc? Did you use the AnyDVD ripping tool?

Problem: The MKV sample just helps me seeing the problem. But it doesn't help me at all to figure out how to solve it. I'd really need a sample of the EVO file! Could you upload that for me? 50MB should do the trick. You can use EvoDemux's "rebuild" option and then simply abort the operation when the destination EVO file is about 50MB. Or alternatively you can also cut the original EVO file with a hex editor or any other file splitting tool.

Thanks!

[madshi]

1) Yes!! fps and default duration is correct.
2) I think so. I just finished to more titles that did not play perfect before:
Mi2 and Dantes Peak

I tested thoose places with micros stutters and they are gone. Plays perfect!

Great job!
Thanks. I'm happy to hear about your results! :) No gaps/overlaps reported by eac3to in any of the movies you've tried so far with v2.19, I hope? (except Million Dollar Baby)

Thoose rainbow frames may be caused by bad authoring. But no rainbow frames can be found before muxing into mkv.
I will test a new source of Million Dollar Baby later today!
Is the new source a fresh rip with the AnyDVD ripping tool?

[rickardk]

Thanks. I'm happy to hear about your results! :) No gaps/overlaps reported by eac3to in any of the movies you've tried so far with v2.19, I hope? (except Million Dollar Baby)


Is the new source a fresh rip with the AnyDVD ripping tool?
No gaps/overlaps reported

Yes

[DeepBeepMeep]

This mastroska support is really a great feature! Thanks Madshi!

However, I have noticed that if the EVO file used as an input was created with EVODemux (just removed a few tracks and merged the two EVo files), the mastroska muxing feature of EAC3TO hangs and then nothing happens. Hopefully you can do something about it. Thanks in advance.

[rickardk]

New Million Dollar Baby still have rainbow frames. I'v been uploading VC-1 sample since 40 minutes ago (slow connection) soon ready at:http://www.earselect.se/test.vc1

Looks just like the sample I just downloaded of POTO.
This is not a muxing problem I think. The VC1 file can be dropped at PowerDVD and playback is fine. Also you can build a graph with sonic demuxer and sonic video decoder and it plays perfectly.

But Microsoft VC-1 and libavcodec can't handle this. They don't behave exactly the same. Diffrent errors.

I think it's safe to say that this is not a mux problem but maybe it can be solved by some bitstream modification while muxing. Will have the ffdshow guys have a look at this sample...

[nautilus7]

I believe the demuxed vc1 is not enough. You should upload a sample from the .evo file.

[rickardk]

I believe the demuxed vc1 is not enough. You should upload a sample from the .evo file.

Ok! Will start upload a sample evo now...

[Yraen]

My collection of HD is small, but I'll add it to the list here since I haven't seen most of them mentioned. All titles are US.

Success stories: VC-1
Bladerunner
The Matrix
The Matrix Reloaded
The Matrix Revolutions *
The Thing

Success Stories: AVC/H264
Transformers

* The Matrix Revolutions shows one gap of three frames. This was with both 2.18 and 2.19. This gap falls right at the beginning of the credits. The mkv plays fine, there doesn't appear to be any kind of visual anomaly at that time point.
eac3to v2.19
"D:\editing apps\eac3to\2.19\eac3to.exe" "H:\HD\MATRIX_REVOLUTIONS_HDDVD\HVDVD_TS\PEVOB_1.EVO"+"H:\HD\MATRIX_REVOLUTIONS_HDDVD\HVDVD_TS\PEVOB_2.EVO" 2: "H:\matrix_revolutions\video.mkv" -strippulldown 5: "h:\matrix_revolutions\audio.ac3" -640
EVO, 2 video tracks, 7 audio tracks, 2:09:16
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: VC-1, 480p30 /1.001, 2002ms
4: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
5: TrueHD, 5.1 channels, 48khz, dialnorm: -27dB
6: E-AC3, 5.1 channels, 640kbit/s, 48khz, dialnorm: -27dB
7: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
8: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
9: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
10: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB, 2002ms
Extracting primary video track...
Muxing video to Matroska...
Extracting audio track number 5...
Removing dialog normalization...
Encoding AC3...
Creating/writing file "h:\matrix_revolutions\audio.ac3"...
Video has a gap of 3 frames at playtime 2:00:21.
[libav] End of stream indicated
Added fps value to MKV header.
Video track 2 contains 185956 frames.
Video track 3 contains 232265 frames.
eac3to processing took 45 minutes, 51 seconds.
Done.


If I can find it today, I plan on picking up POTO US. I'll post my results if I do.

[madshi]

I always want EVO/m2Ts samples. Anyway, thanks for uploading that sample!!