eac3to - audio conversion tool

[Kumo]

Yep, sorry. I just read your original post.

I remember others (including me) have discontinuity errors with xport, but the demuxed pcm file was just fine. Did you try the xport demuxed track with eac3to despite these discontinuities?


If all fails can you upload a sample (50 MB is enough) of the .ts file?

xport keep flashing errors messages ,it creates the .mpa file but it remains at 0k size(no output).
i checked my system.the wavs files can't play cause wave parser can't connect to direct sound device,mpc try to open them using mpeg1 decoder(that's why the weird sound).on others system they play just fine.probably it depends on my sound device(analog devices soundmax).anyway, i don't care about this issue as long as i know the wavs file are ok and i can use them in surcode-dvd.i just want to be sure i'm converting the lpcm track do wavs files in the right way.
supposing all tracks i get(demuxing audio via xmuxer,h264tsto,or mkvextract) are fine,wich command should i use in eac3to?should i specify big/little endian(xmuxer obtained track is recognised as big and swapping endian is applyied during process, others tracks are recognised as little, no swapping is applyied ).

[nautilus7]

As i told you multiple times, wrong endian results in garbage output file (noise). Just tested to be sure! So, NO you do not need to specify the endian.

To encode the pcm file to dts you can do:

eac3to input.pcm output.dts (-speedup) (-768/1536) (+/-delay in ms)

[crazydane]

The bitrate you see is from the decoded 6ch. wav file (the one transmitted through HDMI), not the flac track. Flac is variable bitrate btw, while wav is 16 bits*48 khz*6ch = 4608 kbps
Thanks for addressing all my questions nautilus!

I guess my only remaining thought is this:

Length----Track-----Size------Type
2:23:27---DD+------1.53GB----eac3
2:02:03---TrueHD---1.41GB----flac

From the above, it would appear that the lossless 6 channel flac file ends up being about the same size per hour as a 6 channel lossly eac3 file. How is that possible?

[nautilus7]

Thanks for addressing all my questions nautilus!Happy to help. :)

I guess my only remaining thought is this:

Length----Track-----Size------Type
2:23:27---DD+------1.53GB----eac3
2:02:03---TrueHD---1.41GB----flacBe careful! Truehd is one thing/algorithm and flac is another. It's not like Dolby Digital Plus (name) which uses Enhanced-AC-3 algorithm or Dolby Digital which uses AC-3 algorithm. Truehd and flac are both names and algorithms (completely different ones).

From the above, it would appear that the lossless 6 channel flac file ends up being about the same size per hour as a 6 channel lossly eac3 file. How is that possible?DD+ is a (lossy) format which uses constant bitrates (From 192 kbps - 3072 kbps, i think i got the limits right). Multiply the bitrate by the track's duration and you get the filesize.

Dolby truehd is a (lossless) format with a variable bitrate. Completely different thing.

So, such comparisons don't lead to any useful conclusions.

[madshi]

@madshi,
Do you want a sample from this? I demuxed Fantastic Four 2 and here are the results from 2.16 and 2.17 respectively
Ouch, that's crazy!!!

Yes, a sample would be greatly welcome. Preferably of course a sample with which the problem can be reproduced... :)

[madshi]

with the WMVideo Decoder DMO it works fine, even without rewriting timestamps with mkvmerge. but with libavcodec it wont work even with rewriting the timestamps. im dependent on libavcodec :(
Why are you dependent on libavcodec?

btw. les make a (hopefully small) list of all movies, that cause problems at remuxing so far.
Good idea! But let's limit the list to HD DVD movies.

for example we could list here equilibrium, unless madshi can be sure thats everything is 100% fine and there are no "maybe everything works correctly now". what about POTO, is everything fine there now?
Agreed, Equilibrium should be on the list. POTO (only the USA release) was on the list with v2.14. It should not be on the list for v2.17, anymore, but nobody has confirmed that yet.

dont list movies, which have glitches caused either by a bad rip
Agreed. That would be Unforgiven, Million Dollar Baby and maybe also The Searchers. These 3 should only get to the list once we get confirmation from somebody who is able to rerip them cleanly and confirm.

[madshi]

Rewriting the timestamps will just write correct fps for VC-1. Will not work with AVC unfortunately.
I don't think you're right (see further below).

Just saw that all my HD DVD remuxes with AVC (done with eac3to) still reports 25fps after I had the timestamps rewritten with mkvmerge.
The remuxes don't report anything on their own. It's the filters you're using which are reporting something. Please try this command line "mkvinfo movie.mkv". And check for a field "Default Duration". This is the *REAL* fps value stored in the MKV header. I don't believe that this will indicate 25fps. Or am I wrong?

The frame duration is however right.
Where do you get the frame duration from?

So with AVC we are stuck with the fps written as 25fps or 29.97fps.
I've just checked Transformers which I've remuxed a while ago. It has the proper 23.976fps value in the MKV header.

Don't like to rewrite timestamps as it introduces stutters with some titles. The Last Samurai [HD DVD], Goodfellas [HD DVD], Goodfellas [Blu-ray], The Island [Blu-ray] for example.
The stutters are introduced if the evos (m2ts) are remuxed (with eac3to/gdsmux) then put through mkvmerge for rewriting timestamps.
That's the first time I hear that. Can you provide an EVO sample which reproduces that?

I think it would be best to not rewrite timestamps and instead have pulldown removed and fps written to header. That would be the ultimate solution!
No need to have pulldown removed. On my PC everything works great without that.

I've done hours and hours of testing with diffrent graphic cards, drivers and power strip and my Pioneer Kuro that takes 23.976 Hz and 24.0 Hz.
Stutters is small and very hard to see if feeding the display with a 59.94 Hz signal (3:2 pulldown masks the tiny stutters).
Nobody of us wants to use 59.94Hz, of course... ;)

You're writing a lot about how almost everything seems to stutter for you and how you've spent hours and days on testing that. But what would REALLY help would be if you could provide me with a sample. Or at least concentrate on one specific movie which demonstrates the stuttering and walk with me through some testing steps. If you want to have this solved nicely you need to use a scientific approach. Right now you've mentioned so many new things in this one post alone that I'm heavily confused about anything and everything. Please take one HD DVD VC-1 movie where you experience stuttering and then please do all your testing only with this one movie and report how the different remuxing variations behave. This is the one and only way you can really help with finding the cause of the trouble.

Only way I could make Goodfellas play without micro stutter at some places was to demux the vc-1 stream and mux it with eac3to. The resulting mkv will be 25fps
I can't believe that. Does eac3to report 1080p25? Please don't confuse what ffdshow shows with the REAL thing.

this time its not just the wrong frame rate written in the header.
eac3to v2.17 does not even write a frame rate into the header with VC-1 movies!? :confused:

The frame duration is wrong too: 39.9999999ms
Where do you get this number from?

But after running it through mkvmerge to rewrite timestamps to 23.976xxxx it plays ok. BUT if I remux it into an mkv instead it will contain micro stutters.
I don't understand the last sentence. I thought we were talking about remuxing the movie to mkv all the time?

But when demuxed (raw) VC-1 is muxed with eac3to and then using mkvmerge to rewrite timestamps I have no problems at those places
Which VC-1 decoder are you using?

I thought eac3to used gdsmux for doing raw vc-1 into mkv.
No. For VC-1 movies eac3to is using it's own splitter which feeds the Haali Matroska Muxer. gdsmux internally using Haali's Splitter combined with Haali's Matroska Muxer.

[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

The image in resulting mkv stutters a lot. Why? How mpeg2 video is treated by eac3to?
mpeg2 video is treated the "old" way by using Haali Media Splitter -> Haali Matroska Muxer and then rewriting timestamps with mkvtoolnix. Can you please check the runtime of the final MKV file. Is it 0:01:54 or something else?

[Chumbo]

Ouch, that's crazy!!!

Yes, a sample would be greatly welcome. Preferably of course a sample with which the problem can be reproduced... :)
I think it's a bad rip to begin with. Playing the evo directly with WMVideo Decoder DMO results in total green garbage all over the place. The same if playing the .vc1 file after demuxing it with graphedt or eac3to 2.14. The ffdshow decoder plays it clean up to certain scenes and then you get "rainbow" garbage for a few seconds or more and then back okay.

In both cases of .evo and .vc1, the Sonic 4.3 video decoder plays everything just fine and clean.

After I demuxed the vc1 stream to a .vc1 file with 2.16 and 2.17, both now play like ffdshow (in WMVideo decoder dmo), i.e., all the green garbage is gone, however, where ffdshow plays rainbow-like images, you get blocky garbe at about the same place.

It's early on that this happens. I'll let you know when it's available somewhere.

[Richum]

Extracting primary video track...
Muxing video to Matroska...
Extracting audio track number 3...
Removing dialog normalization...
Applying (E-)AC3 delay...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...
DirectShow reports 5.1 channels, 24 bits, 48khz
Encoding AC3...
Creating/writing file "D:\MKV\audio.ac3"...
Video overlaps for 2 frames at playtime 0:04:47.
Video has a gap of 2 frames at playtime 0:04:50.


This comes from 12 Monkeys, which I previously converted using ffmpeg to Mpeg2.ts I had audio sync problems then, could it have been caused by the referenced video overlaps above?

Also, for those who have successfully converted Transformers using eac3to, I would be interested in knowing where (Europe, Asia) your HD DVD was purchased. Mine is a US copy, which plays fine in my Toshiba XA2, but has consistently (3 attempts) produced out of sync audio and stutter in the video after muxing, video and audio alone are fine. I may borrow a copy from a friend to see if there is a problem with mine, or perhaps all US sold discs.

[crazydane]

To go from DD+ to flac, is this the command line to use?

eac3to feature_1.evo+feature_2.evo 1: soundtrack.flac movie.mkv

And to go from DD+ to DTS, would this be it?

eac3to feature_1.evo+feature_2.evo 1: soundtrack.dts 1536 movie.mkv

[nautilus7]

To go from DD+ to flac, is this the command line to use?

eac3to feature_1.evo+feature_2.evo 1: soundtrack.flac movie.mkv

And to go from DD+ to DTS, would this be it?

eac3to feature_1.evo+feature_2.evo 1: soundtrack.dts 1536 movie.mkvCan you please spent some time reading/understanding the 1st post of this thread?

:thanks:

[rickardk]

I don't think you're right (see further below).


The remuxes don't report anything on their own. It's the filters you're using which are reporting something. Please try this command line "mkvinfo movie.mkv". And check for a field "Default Duration". This is the *REAL* fps value stored in the MKV header. I don't believe that this will indicate 25fps. Or am I wrong?


I'm sorry that my posts make no sense and that I keep telling you about all testing with stutters and stuff. I hope I did not whined to much about all titles stuttering. There is diffrent kinds of stutters I think. Performance related, frame rate/refresh rate missmatch related and thoose described later in this post. Never had performance related stutters but I did wrote a couple of PMs about the frame rate/refresh rate missmatch because I did some testings with diffrent graphic cards and how it affected playback of 23.976fps material on a 23.976fps capable television. Then the question was if eac3to should write 24/1.001 or 23.976.

I guess it's the english language and the fact that I always done all testing in the middle of the night that makes my post confusing (not that good in english).
I'm just so anxious about having my Blu-rays and HD DVDs backed up in a perfect way.



To sum up from the two previous posts:
I know that if the frame duration is about 41.708ms then it is 23.976fps. I think I also wrote that the mkvs plays back as 23.976fps. But the filter (ffdshow) reports them as something else.


I think it was something like this:

VC-1 from HD DVD without having the timestamps rewritten:
= 25fps (BUT frame duration is 41.708ms)

VC-1 from HD DVD having the timestamps rewritten (24/1.001):
= 23.976fps (Frame duration is 41.708ms)

AVC from HD DVD without having the timestamps rewritten:
= 25fps (Frame duration is 41.708ms)

AVC from HD DVD having the timestamps rewritten (24/1.001):
= still 25fps (Frame duration is 41.708ms)

AVC stream from HD DVD demuxed and muxed with mkvmerge setting fps to 24/1.001:
= 29.97fps (Frame duration is still 41.708ms)

Based on the fact that AVC streams demuxed from Blu-ray and muxed with same settings with mkvmerge gave:
= 23.976fps (Frame duration 41.708ms)
I wrote that I thought the reported 29.97fps from HD DVD AVC had something to do with pulldown.


I know that all the above plays at 23.976fps BUT filter like ffdshow tells me diffrent fps depending on muxing method. And it would be good if ffdshow could see them all as 23.976fps if you have features or macros that reads the frame rate from filters like ffdshow.

Maybe this is a ffdshow problem...



Then we had titles with tiny tiny stutters:

Using the usual eac3to feature_1.ev+feature_2.evo title.mkv and then rewriting the timestamps give tiny tiny stutter at some places in a couple of movies.

Now I tried to first demux the VC-1 stream and mux it with the new VC-1 muxing feature of eac3to. ffdshow now tells me that the resulting mkv is 25fps. But this time it also tells me the frame duration is 39.999999 (usually 41.708ms).

After rewriting the timestamps with mkvmerge I get a file that ffdshow reports as 23.976fps and frame duration is 41.708ms.

And now I'm suprised to see that it plays back perfectly without thoose tiny stutters at thoose places where it used to be small stutters.





I can't explain why diffrent approaches gives diffrent results. And if the filter ffdshow reads the frame duration in the header of the mkv, why does it reports diffrent fps under the above described situations? I first thought it was the decoder that told ffdshow the correct fps and that ffdshow recieved the frame duration from the splitter.

And the tiny stutter problems at just some places might indicate that the some single frames are read or written in the wrong order or something. I really have no clue BUT the stutters always happens at the same places. Hard to understand why it plays back perfectly when first demuxing and then muxing through eac3to?

I really have no understanding about how the frames are handled.


And I do understand that it would help alot if I could provide more samples. I would if my connection was faster. Having 5kB/s upstreams right now due to some problems with our 20/3 ADSL connection.
To upload that Sum of all Fears sample (that made eac3to crash) took forever. But I will soon get 100/100 so I will provide more samples then.


Sorry for taking your time with this...
I will be quite now!

[BlackJack1]

Tryed to convert flac file to dts.
Downloaded flac file inside *.mka container. Used MKVextractGUI for extract flac file. All passed OK but this file can not play. I've got error: invalid flac file. Tryded eac3to to convert - it does not started...
How to propertly extract flac from mka file?

[nautilus7]

Mkvextract gui outputs flac files in a ogg container. You have to use the cmd of this tool and use a special option for the flac files.

[madshi]

Video overlaps for 2 frames at playtime 0:04:47.
Video has a gap of 2 frames at playtime 0:04:50.
Hmmmm... That's bad. Is that the only gap/overlap complaint for the whole movie? Or are there more such complaints?

This comes from 12 Monkeys, which I previously converted using ffmpeg to Mpeg2.ts I had audio sync problems then, could it have been caused by the referenced video overlaps above?
No, if there's a "gap" and an "overlap" with the same frame number, audio sync should not be affected.

Also, for those who have successfully converted Transformers using eac3to, I would be interested in knowing where (Europe, Asia) your HD DVD was purchased. Mine is a US copy, which plays fine in my Toshiba XA2, but has consistently (3 attempts) produced out of sync audio and stutter in the video after muxing, video and audio alone are fine. I may borrow a copy from a friend to see if there is a problem with mine, or perhaps all US sold discs.
USA. Not a problem with audio sync at all. Which format did you convert the audio to? Did you mux the audio into the MKV? Did you reencode the video or did you leave it as it it (just remuxed into MKV)?

[BlackJack1]

Mkvextract gui outputs flac files in a ogg container. You have to use the cmd of this tool and use a special option for the flac files.

I'm newbie in this subject. Can you explain more clearly please and give me exact phrase to put in cmd?

[madshi]

I guess it's the english language and the fact that I always done all testing in the middle of the night that makes my post confusing (not that good in english).
I'm just so anxious about having my Blu-rays and HD DVDs backed up in a perfect way.
Don't worry, I'm neither angry nor impatient nor anything else. I'm just trying to say that I need the information in a somewhat organized form. We need to handle one problem at a time. If you throw 5 problems at once at me, maybe even mixed up in some way, I don't know where to start... ;)

To sum up from the two previous posts:
I know that if the frame duration is about 41.708ms then it is 23.976fps. I think I also wrote that the mkvs plays back as 23.976fps. But the filter (ffdshow) reports them as something else.

I think it was something like this:

VC-1 from HD DVD without having the timestamps rewritten:
= 25fps (BUT frame duration is 41.708ms)

VC-1 from HD DVD having the timestamps rewritten (24/1.001):
= 23.976fps (Frame duration is 41.708ms)

AVC from HD DVD without having the timestamps rewritten:
= 25fps (Frame duration is 41.708ms)

AVC from HD DVD having the timestamps rewritten (24/1.001):
= still 25fps (Frame duration is 41.708ms)

AVC stream from HD DVD demuxed and muxed with mkvmerge setting fps to 24/1.001:
= 29.97fps (Frame duration is still 41.708ms)

Based on the fact that AVC streams demuxed from Blu-ray and muxed with same settings with mkvmerge gave:
= 23.976fps (Frame duration 41.708ms)
I wrote that I thought the reported 29.97fps from HD DVD AVC had something to do with pulldown.

I know that all the above plays at 23.976fps BUT filter like ffdshow tells me diffrent fps depending on muxing method. And it would be good if ffdshow could see them all as 23.976fps if you have features or macros that reads the frame rate from filters like ffdshow.

Maybe this is a ffdshow problem...
Those comparisons above are helpful and interesting. But I think there's one set of information missing: You're saying "VC-1 from HD DVD". But you're not differing between eac3to v2.14 (or older) and v2.15 (or newer). But those two versions behave quite different for VC-1 movies. So in your tests you need to differ between those 2 versions.

Furthermore could you please try "mkvinfo movie.mkv" for all those situations posted above to find out what is really stored in the MKV file in each case? This is just to find out whether ffdshow is wrong or right with what it reports. BTW, why do you use ffdshow in the first place? The MS VC-1 decoder should be superior. I do not see a need for ffdshow for VC-1 playback on a Windows PC.

Then we had titles with tiny tiny stutters:

Using the usual eac3to feature_1.ev+feature_2.evo title.mkv and then rewriting the timestamps give tiny tiny stutter at some places in a couple of movies.

Now I tried to first demux the VC-1 stream and mux it with the new VC-1 muxing feature of eac3to. ffdshow now tells me that the resulting mkv is 25fps. But this time it also tells me the frame duration is 39.999999 (usually 41.708ms).

After rewriting the timestamps with mkvmerge I get a file that ffdshow reports as 23.976fps and frame duration is 41.708ms.

And now I'm suprised to see that it plays back perfectly without thoose tiny stutters at thoose places where it used to be small stutters.
Which eac3to version did you use for this test? v2.17? Did you use the "-stripPulldown" option?

When using v2.17 *without* the "-stripPulldown" option doing the following 2 things should be 100% identical:

(1) remuxing VC-1 EVO directly to MKV (with eac3to).
(2) demuxing VC-1 and then muxing it to MKV (both with eac3to).

The only situations where there can be a difference between those 2 actions are if you use the "-stripPulldown" flag during demuxing of the VC-1 track, or when eac3to complains about gaps/overlaps. If you don't use the "-stripPulldown" flag and if eac3to doesn't complain about gaps/overlaps during remuxing then both of the listed actions should yield 100% identical results.

And I do understand that it would help alot if I could provide more samples.
It would help. But you can also help a lot if you do exactly those tests I'm asking you for and if you report the test results without throwing in anything unrelated to that or tests with other movies or anything like that. In that case we might get along without any samples.

I will be quite now!
No, please don't!! Actually I'd *LOVE* to hear more from you. Especially if you answer all the questions I've asked you. Your posts are very helpful in the eac3to development. I just need the information in a bit more organized form. One movie at a time. With all comparisons described in detail. With "mkvinfo". And with differing between eac3to v2.14 (or older) <-> v2.15 (or newer).

[nautilus7]

mpeg2 video is treated the "old" way by using Haali Media Splitter -> Haali Matroska Muxer and then rewriting timestamps with mkvtoolnix. Can you please check the runtime of the final MKV file. Is it 0:01:54 or something else?
Well runtime is 00:01:53.880 which is correct.

Do you want a small sample, or you have a mpeg2 evo?

Btw, there is something else i didn't tell you before. When i first did this conversion i got:

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
Muxing video to Matroska...
Extracting audio track number 2...
Removing dialog normalization...
Creating/writing file "audiotrack 1.ac3"...
Error renaming MKV file.
MKV file was successfully created, but the timecodes were not rewritten.

And i repeated it again till i finally managed to get it work. Actually 9/10 times didn't work. I think it's because the files processed are very small and the procedure finishes almost instantly, so eac3to is still using the output file.

Of course, the stuttering is present even if the timestamps are re written. It's the same actually. :p

[rickardk]

Ok, I will do more testing tomorrow and follow your guide lines.

I actually do use the MS VC-1 decoder but I need to have ffdshow in the graph (ffdshow set to raw video) as the essential Media Control (Vista Media Center plugin... GREAT APP) is depending on ffdshow.

[nautilus7]

I'm newbie in this subject. Can you explain more clearly please and give me exact phrase to put in cmd?

Use the search function to find what you need. It's all here (not in this thead) in the forum.

This thread is already big enough... Let's try not to post things that don't belong here, please.

:thanks:


EDIT: Because i am a nice person: mkvextract tracks input.mka --no-ogg 1:audio.flac :p

[madshi]

Well runtime is 00:01:53.880 which is correct.

Do you want a small sample, or you have a mpeg2 evo?
I do have some mpeg2 480i/p evos but I didn't notice stuttering. But I didn't really look for it. Does it happen for you with every mpeg2 480i/p evo you try or is it only with one specific evo?

[madshi]

Ok, I will do more testing tomorrow and follow your guide lines.
Thanks! Looking forward to that...

I actually do use the MS VC-1 decoder but I need to have ffdshow in the graph (ffdshow set to raw video) as the essential Media Control (Vista Media Center plugin... GREAT APP) is depending on ffdshow.
Ah, I see.

[nautilus7]

I do have some mpeg2 480i/p evos but I didn't notice stuttering. But I didn't really look for it. Does it happen for you with every mpeg2 480i/p evo you try or is it only with one specific evo?
Tried 3. 100% fail. :mad:

Can it be because all were interlaced? PowerDVD plays them fine. What about the re writing timestamps error?

[Richum]

Hmmmm... That's bad. Is that the only gap/overlap complaint for the whole movie? Or are there more such complaints?


No, if there's a "gap" and an "overlap" with the same frame number, audio sync should not be affected.


USA. Not a problem with audio sync at all. Which format did you convert the audio to? Did you mux the audio into the MKV? Did you reencode the video or did you leave it as it it (just remuxed into MKV)?

There were about 10 gap/overlaps noted, so yes there were more than a few gaps and overlaps. I am trying -ignoreGaps just to see what happens and will post the results.


In Transformers I converted the audio to AC3 -640 maybe that screwed it up, but I muxed the video into MKV! Since your US release was problem free I will try only muxing both V and A.

[madshi]

Tried 3. 100% fail. :mad:

Can it be because all were interlaced? PowerDVD plays them fine. What about the re writing timestamps error?
Have you tried playing the EVOs directly in MPC or ZoomPlayer (or whatever player you're normally using)? Do they play fine? If yes then there must be something wrong with eac3to's remuxing. If you have the same stuttering issue when playing the EVO files then your DirectShow filters are probably at guilt.

[nautilus7]

Have you tried playing the EVOs directly in MPC or ZoomPlayer (or whatever player you're normally using)? Do they play fine? If yes then there must be something wrong with eac3to's remuxing. If you have the same stuttering issue when playing the EVO files then your DirectShow filters are probably at guilt.

Just did... Fortunately for my system, but unfortunately for eac3to, the evos play fine in mpc (sonic cinemaster decoder 4.3 for both evos and mkv).

[Thunderbolt8]

Why are you dependent on libavcodec?
because imho libavcodec still has the best picture quality (even if its only a tiny bit better, but still) and it also allows me to apply real time sharpening, which I need to display all that 1080p stuff on my 720p resolution, otherwise everything mostly looks a bit blurred due to that downsizing.

regarding that 480i thing which causes stuttering, I know from another mpeg2 TV source (planet terror, 720p) that haali splitter has problems with 720p files at 59,94 frames and drops lots of frames then. dont know why, but thats the way it is atm. so maybe its quite similar here.

btw. another thing, I experienced two mpeg2 1080i sources now, in which the remuxing with gdsmux resulted in little video stuttering of ~0.5 seconds every few seconds in the movie (this was also the case when I put that file in mkvmerge afterwards, didnt change). but when I demuxed it with xport and then remuxed audio and video with mkvmerge, then it was fine. since then im a bit skeptical what haali and mpeg2 muxing is concerned :S

[darkbeats]

You can do this by using this command:

C:\eac3to>eac3to.exe EVOB002.EVO 1: ddda.mkv 2: audio.ac3

Alternately you can specify different options for each output file, like this:

C:\eac3to>eac3to.exe EVOB002.EVO 1: ddda.mkv -[options for 1:] 2: audio.ac3 -[options for 2:]

hey nautilus7, I tried that a few times, with and without audio options, with and without nero, and everytime i get a mkv with no audio, and a seperate audio.ac3 file...

G:\eac3to>eac3to.exe EVOB002.EVO 1: c:\temp\ddda.mkv 2: c:\temp\audio.ac3 -192
EVO, 1 video track, 2 audio tracks, 2:03:53
1: VC-1, 1280x1080 24p /1.001
2: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
3: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Extracting audio track number 2...
Muxing video to Matroska...
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...
DirectShow reports 2.0 channels, 24 bits, 48khz
Encoding AC3...
Creating/writing file "c:\temp\audio.ac3"...
Video track 1 contains 178220 frames.
eac3to processing took 2 hours, 58 minutes.
Done.

alternatively, i can demux the vc1 and audio cleanly, both play great, but i don't know how to use eac3to (or other) to get them into a mkv container.

I'm just trying to mux track 1 (vc1 video), with track 2 (ac3) while ignoring track 3, since eac3to errors out on track 3 no matter what i do. ( original post (http://forum.doom9.org/showthread.php?p=1090853#post1090853))

btw: i love this thread and the great work madshi and nautilus are doing here. i think i look forward to this thread every morning, more than my first cup of coffee.

[BlackJack1]

mkvextract tracks input.mka 1:audio.flac --no-ogg[/B] :p

You are very nice man :D Thanks a lot :)

[nautilus7]

hey nautilus7, I tried that a few times, with and without audio options, with and without nero, and everytime i get a mkv with no audio, and a seperate audio.ac3 file...

G:\eac3to>eac3to.exe EVOB002.EVO 1: c:\temp\ddda.mkv 2: c:\temp\audio.ac3 -192
EVO, 1 video track, 2 audio tracks, 2:03:53
1: VC-1, 1280x1080 24p /1.001
2: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
3: AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Extracting primary video track...
Extracting audio track number 2...
Muxing video to Matroska...
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...
DirectShow reports 2.0 channels, 24 bits, 48khz
Encoding AC3...
Creating/writing file "c:\temp\audio.ac3"...
Video track 1 contains 178220 frames.
eac3to processing took 2 hours, 58 minutes.
Done.
This is how eac3to works. Muxes video to mkv with correct timestamps and process audio as you like. No problem there! :)

Then you have to use mkvmergegui to mux the audio file(s) to the mkv file created with eac3to.

[saint-francis]

Extracting primary video track...
Muxing video to Matroska...
Extracting audio track number 3...
Removing dialog normalization...
Applying (E-)AC3 delay...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...
DirectShow reports 5.1 channels, 24 bits, 48khz
Encoding AC3...
Creating/writing file "D:\MKV\audio.ac3"...
Video overlaps for 2 frames at playtime 0:04:47.
Video has a gap of 2 frames at playtime 0:04:50.


This comes from 12 Monkeys, which I previously converted using ffmpeg to Mpeg2.ts I had audio sync problems then, could it have been caused by the referenced video overlaps above?


I just tried this with a US copy of 12 monkeys again but using eac3to and I didn't encounter any problems. Try to re rip the disk.

[darkbeats]

This is how eac3to works. Muxes video to mkv with correct timestamps and process audio as you like. No problem there! :)

Then you have to use mkvmergegui to mux the audio file(s) to the mkv file created with eac3to.

wha?! oh, snap! how did i miss that... trying that right now. *shakes head* ... no wonder, and here i thought i was going crazy....

[Tegeril]

Audio encoded fine on my 300 issue from the original EVO. Must be something in the EVOs that EVOdemux can't do right.

[Tegeril]

My happiness was premature. Audio goes in and out of sync when muxed now. Some parts are perfect, others are a second off.

Sigh.

Hm, might be a decoding problem. Parts that were out of sync are not always out of sync...
Edit: Decoding problem, watched from the beginning to one of the out of sync spots and it was fine. Doesn't seek very well >.<

[Richum]

I just tried this with a US copy of 12 monkeys again but using eac3to and I didn't encounter any problems. Try to re rip the disk.

I am pretty sure that my copy of 12 Monkeys and Transformers are sub-standard presses. I have noticed from personal experience and from comments made on another forum (AVS) that the quality of the HD DVDs (the disc itself) can vary.

For example you have had no problem with 12 Monkeys and Madshi/Nautilus7 have not had a problem with Transformers and I have not had a problem with Riddick or Knocked Up.

Now add to that all the folks I have read about who had problems playing a HD DVD on their Toshiba Players and returned the disc only to be replaced with one which played flawlessly.

Two out of four worked like a charm, so I am not going to fight to try and convert a stubborn disc any more, it doesn't go in two tries using proven protocol I'll forget it and move to the next one.

I have about 40 HD DVDs counting a couple of boxed sets and over time I intend to see just how many convert. It is well documented that the quality of these discs varies.

[Richum]

wha?! oh, snap! how did i miss that... trying that right now. *shakes head* ... no wonder, and here i thought i was going crazy....

Welcome to the club, I made the same mistake.:rolleyes:

[BlackJack1]

Tryed to make dts from flac (not first time - previously all was ok) but now received strange message:
[img=http://img81.imageshack.us/img81/1664/beztytuuzb5.th.jpg] (http://img81.imageshack.us/my.php?image=beztytuuzb5.jpg)

Anybody knows what happened?

[Yraen]

My happiness was premature. Audio goes in and out of sync when muxed now. Some parts are perfect, others are a second off.

Sigh.

Hm, might be a decoding problem. Parts that were out of sync are not always out of sync...
Edit: Decoding problem, watched from the beginning to one of the out of sync spots and it was fine. Doesn't seek very well >.<

If I use mkvmerge I get files that don't seek well also. Using gdsmux I get files that seek fine.

[Yraen]

madshi,

You seemed interested in the gaps and overlaps generated from evos. I currently have three titles that show this, all three of the Matrix trilogy. I just ran it through eac3to 2.17 using the -stripPulldown and without it. Command line is on top.

-stripPulldown

"D:\editing apps\eac3to\2.17\eac3to.exe" "H:\HD\THE_MATRIX_HDDVD\HVDVD_TS\PEVOB_1.EVO"+"H:\HD\THE_MATRIX_HDDVD\HVDVD_TS\PEVOB_2.EVO" 2: "H:\matrix\strippulldown.mkv" -strippulldown > "H:\matrix\stripped.txt"
EVO, 2 video tracks, 9 audio tracks, 2:16:18
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: VC-1, 480p30 /1.001, 66ms
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
11: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
12: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB, -3ms
Extracting primary video track...
Muxing video to Matroska...
Video overlaps for 2 frames at playtime 0:00:19.
Video has a gap of 2 frames at playtime 0:00:21.
Video overlaps for 2 frames at playtime 0:01:27.
Video has a gap of 2 frames at playtime 0:01:30.
Video overlaps for 2 frames at playtime 0:01:55.
Video has a gap of 2 frames at playtime 0:02:01.
Video overlaps for 2 frames at playtime 0:02:12.
Video has a gap of 2 frames at playtime 0:02:15.
Video overlaps for 2 frames at playtime 0:03:06.
Video has a gap of 2 frames at playtime 0:03:13.
Video overlaps for 2 frames at playtime 0:03:15.
Video has a gap of 2 frames at playtime 0:03:22.
Video overlaps for 2 frames at playtime 0:05:08.
Video has a gap of 2 frames at playtime 0:05:10.
Video overlaps for 2 frames at playtime 0:05:50.
Video has a gap of 2 frames at playtime 0:05:57.
Video overlaps for 2 frames at playtime 0:09:44.
Video has a gap of 2 frames at playtime 0:10:01.
Video overlaps for 2 frames at playtime 0:23:06.
Video has a gap of 2 frames at playtime 0:23:15.
Video overlaps for 2 frames at playtime 0:24:48.
Video has a gap of 2 frames at playtime 0:25:06.
Video overlaps for 2 frames at playtime 0:32:09.
Video has a gap of 2 frames at playtime 0:32:15.
Video overlaps for 2 frames at playtime 0:34:25.
Video has a gap of 2 frames at playtime 0:34:29.
Video overlaps for 2 frames at playtime 0:44:20.
Video has a gap of 2 frames at playtime 0:44:27.
Video overlaps for 2 frames at playtime 0:53:15.
Video has a gap of 2 frames at playtime 0:53:22.
Video overlaps for 2 frames at playtime 0:59:52.
Video has a gap of 2 frames at playtime 0:59:58.
Video overlaps for 2 frames at playtime 1:20:10.
Video has a gap of 2 frames at playtime 1:20:13.
Video overlaps for 2 frames at playtime 1:21:22.
Video has a gap of 2 frames at playtime 1:21:30.
Video overlaps for 2 frames at playtime 1:23:13.
Video has a gap of 2 frames at playtime 1:23:15.
Video overlaps for 2 frames at playtime 1:23:28.
Video has a gap of 2 frames at playtime 1:23:43.
Video overlaps for 2 frames at playtime 1:54:16.
Video has a gap of 2 frames at playtime 1:54:18.
Video overlaps for 2 frames at playtime 1:59:05.
Video has a gap of 2 frames at playtime 1:59:07.
Video overlaps for 2 frames at playtime 2:00:53.
Video has a gap of 2 frames at playtime 2:00:58.
Video overlaps for 2 frames at playtime 2:01:09.
Video has a gap of 2 frames at playtime 2:01:12.
Video overlaps for 2 frames at playtime 2:02:28.
Video has a gap of 2 frames at playtime 2:02:35.
Video overlaps for 2 frames at playtime 2:03:08.
Video has a gap of 2 frames at playtime 2:03:10.
Video overlaps for 2 frames at playtime 2:06:56.
Video has a gap of 2 frames at playtime 2:07:22.
Video track 2 contains 196071 frames.
Video track 3 contains 244929 frames.
eac3to processing took 1 hour, 21 minutes.
Done.


without -stripPulldown

"D:\editing apps\eac3to\2.17\eac3to.exe" "H:\HD\THE_MATRIX_HDDVD\HVDVD_TS\PEVOB_1.EVO"+"H:\HD\THE_MATRIX_HDDVD\HVDVD_TS\PEVOB_2.EVO" 2: "H:\matrix\non_strippulldown.mkv" > "H:\matrix\non_stripped.txt"
EVO, 2 video tracks, 9 audio tracks, 2:16:18
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: VC-1, 480p30 /1.001, 66ms
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
11: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
12: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB, -3ms
Extracting primary video track...
Muxing video to Matroska...
Video overlaps for 2 frames at playtime 0:00:19.
Video has a gap of 2 frames at playtime 0:00:21.
Video overlaps for 2 frames at playtime 0:01:27.
Video has a gap of 2 frames at playtime 0:01:30.
Video overlaps for 2 frames at playtime 0:01:55.
Video has a gap of 2 frames at playtime 0:02:01.
Video overlaps for 2 frames at playtime 0:02:12.
Video has a gap of 2 frames at playtime 0:02:15.
Video overlaps for 2 frames at playtime 0:03:06.
Video has a gap of 2 frames at playtime 0:03:13.
Video overlaps for 2 frames at playtime 0:03:15.
Video has a gap of 2 frames at playtime 0:03:22.
Video overlaps for 2 frames at playtime 0:05:08.
Video has a gap of 2 frames at playtime 0:05:10.
Video overlaps for 2 frames at playtime 0:05:50.
Video has a gap of 2 frames at playtime 0:05:57.
Video overlaps for 2 frames at playtime 0:09:44.
Video has a gap of 2 frames at playtime 0:10:01.
Video overlaps for 2 frames at playtime 0:23:06.
Video has a gap of 2 frames at playtime 0:23:15.
Video overlaps for 2 frames at playtime 0:24:48.
Video has a gap of 2 frames at playtime 0:25:06.
Video overlaps for 2 frames at playtime 0:32:09.
Video has a gap of 2 frames at playtime 0:32:15.
Video overlaps for 2 frames at playtime 0:34:25.
Video has a gap of 2 frames at playtime 0:34:29.
Video overlaps for 2 frames at playtime 0:44:20.
Video has a gap of 2 frames at playtime 0:44:27.
Video overlaps for 2 frames at playtime 0:53:15.
Video has a gap of 2 frames at playtime 0:53:22.
Video overlaps for 2 frames at playtime 0:59:52.
Video has a gap of 2 frames at playtime 0:59:58.
Video overlaps for 2 frames at playtime 1:20:10.
Video has a gap of 2 frames at playtime 1:20:13.
Video overlaps for 2 frames at playtime 1:21:22.
Video has a gap of 2 frames at playtime 1:21:30.
Video overlaps for 2 frames at playtime 1:23:13.
Video has a gap of 2 frames at playtime 1:23:15.
Video overlaps for 2 frames at playtime 1:23:28.
Video has a gap of 2 frames at playtime 1:23:43.
Video overlaps for 2 frames at playtime 1:54:16.
Video has a gap of 2 frames at playtime 1:54:18.
Video overlaps for 2 frames at playtime 1:59:05.
Video has a gap of 2 frames at playtime 1:59:07.
Video overlaps for 2 frames at playtime 2:00:53.
Video has a gap of 2 frames at playtime 2:00:58.
Video overlaps for 2 frames at playtime 2:01:09.
Video has a gap of 2 frames at playtime 2:01:12.
Video overlaps for 2 frames at playtime 2:02:28.
Video has a gap of 2 frames at playtime 2:02:35.
Video overlaps for 2 frames at playtime 2:03:08.
Video has a gap of 2 frames at playtime 2:03:10.
Video overlaps for 2 frames at playtime 2:06:56.
Video has a gap of 2 frames at playtime 2:07:22.
Video track 2 contains 196071 frames.
Video track 3 contains 244929 frames.
eac3to processing took 58 minutes, 41 seconds.
Done.


This is common across all three of those movies. From what I've seen of these two files, both play fine. I ripped the disk again and the overlaps/gaps were in the exact positions they are here. I can put up a sample of the evo tomorrow if you want it.

[intomed]

eac3to v2.17 released

http://madshi.net/eac3to.zip

Also, does anybody have the USA HD DVD of Phantom of the Opera? I really would like to have that tested cause it didn't work with v2.14, but should work now with v2.15-17.

Madshi,

I do own the US version of Phantom. I have the next several days off. Will re-rip it and try to convert to mkv. I'll get back to you soon.:helpful:

[Joniii]

Madshi, what's that error?
E:\2.17>eac3to 300.thd D:\300.dts
TrueHD/AC3, 5.1 channels, 48khz, dialnorm: -27dB
Extracting TrueHD stream...
Removing dialog normalization...
Writing WAVs...
Creating/writing file "d:\300.L.wav"...
Creating/writing file "d:\300.C.wav"...
Creating/writing file "d:\300.SL.wav"...
Creating/writing file "d:\300.R.wav"...
Creating/writing file "d:\300.LFE.wav"...
Creating/writing file "d:\300.SR.wav"...
[libav] Lossless check failed - expected 0, calculated 69
Found Surcode DTS Encoder version 1.0.21.0.
Surcode encoding successfully started. Please wait...
Closing Surcode...
eac3to processing took 15 minutes, 11 seconds.
Surcode encoding took 22 minutes, 44 seconds.
Done.

[Joniii]

I'm having trouble converting 300's TrueHD track to AC3. I've tried with 2.14, updated to 2.17, and same results. The track has very faint noise, and nothing else. I've converted TrueHD from multiple other titles without issue.

Microsoft Windows [Version 6.0.6000]
Copyright (c) 2006 Microsoft Corporation. All rights reserved.

V:\Encoder\Movie>"V:\Encoder\HD-DVD Tools\EAC3toGUI\eac3to.exe" "V:\Encoder\Movi
e\300 TrueHD.EVO" "V:\Encoder\Movie\300 TrueHD.ac3" -640 -24
EVO, 1 audio track, 1:56:33
1: TrueHD, 5.1 channels, 48khz, dialnorm: -27dB, 2628056ms
Extracting audio track number 1...
Removing dialog normalization...
Applying RAW/PCM delay...
Encoding AC3...
Creating/writing file "V:\Encoder\Movie\300 TrueHD.ac3"...
[libav] End of stream indicated
eac3to processing took exactly 5 minutes.
Done.

V:\Encoder\Movie>

Any thoughts? Filesize also seems too low, 363,000ish KB.

Edit: For the heck of it, I re-rebuilt the audio file out of my EVOs, same problem.

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).

[madshi]

I think it's a bad rip to begin with. Playing the evo directly with WMVideo Decoder DMO results in total green garbage all over the place. The same if playing the .vc1 file after demuxing it with graphedt or eac3to 2.14. The ffdshow decoder plays it clean up to certain scenes and then you get "rainbow" garbage for a few seconds or more and then back okay.

In both cases of .evo and .vc1, the Sonic 4.3 video decoder plays everything just fine and clean.

After I demuxed the vc1 stream to a .vc1 file with 2.16 and 2.17, both now play like ffdshow (in WMVideo decoder dmo), i.e., all the green garbage is gone, however, where ffdshow plays rainbow-like images, you get blocky garbe at about the same place.
This is really an interesting sample. For whatever reason I cannot reproduce the strange gap reports. My latest eac3to version (work in progress) doesn't complain about any gaps/overlaps with your version at all. Can you please retry with just the sample? Does your eac3to version really complain with the sample? If so, then please retry with eac3to v2.18.

What is most interesting is that (as you say) a Sonic graph plays the sample just fine. But the MS VC-1 decoder seems to be have trouble with the sample. I'm not sure where the problem comes from. It could be a bad rip or a mastering error which for whatever strange reason Sonic can handle but the MS VC-1 decoder cannot. It could also be a bug in the MS VC-1 decoder. We will not be able to know either way until someone tries to rerip the disk cleanly. Anybody?

[nautilus7]

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).
The same? The other member's problem was the noisy output file. You are talking about lower volume (though judging from volume setting in your receiver isn't the perfect way to identify the problem, if any).

[madshi]

Well runtime is 00:01:53.880 which is correct.

Do you want a small sample, or you have a mpeg2 evo?
I think a sample would help.

I do have a MPEG2 480i60 sample and also a MPEG2 480p24 sample. The 480i60 sample plays very smoothly both in its original VOB form and also in the remuxed MKV. The 480p24 sample is stuttering a lot for me, but both as EVO and as MKV. So a sample would be helpful.

Which EVO splitter are you using that gives you smooth playback?

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.

[madshi]

because imho libavcodec still has the best picture quality (even if its only a tiny bit better, but still)
Compared to the MS VC-1 decoder!? I'm having a hard time believing that. I mean the MS VC-1 decoder was tested by MS against the reference decoder and it should give out practically perfect images. I don't see how any decoder could top that (except in CPU usage, of course).

and it also allows me to apply real time sharpening, which I need to display all that 1080p stuff on my 720p resolution, otherwise everything mostly looks a bit blurred due to that downsizing.
Can't you use the MS VC-1 decoder and then still use ffdshow for sharpening?

btw. another thing, I experienced two mpeg2 1080i sources now, in which the remuxing with gdsmux resulted in little video stuttering of ~0.5 seconds every few seconds in the movie (this was also the case when I put that file in mkvmerge afterwards, didnt change). but when I demuxed it with xport and then remuxed audio and video with mkvmerge, then it was fine. since then im a bit skeptical what haali and mpeg2 muxing is concerned :S
Well, eac3to will sooner or later replace Haali's splitter with my own for MPEG2 content, too. Maybe that will take care of the problem. We'll have to wait and see...

[madshi]

I am pretty sure that my copy of 12 Monkeys and Transformers are sub-standard presses. I have noticed from personal experience and from comments made on another forum (AVS) that the quality of the HD DVDs (the disc itself) can vary.

For example you have had no problem with 12 Monkeys and Madshi/Nautilus7 have not had a problem with Transformers and I have not had a problem with Riddick or Knocked Up.
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...

[madshi]

You seemed interested in the gaps and overlaps generated from evos. I currently have three titles that show this, all three of the Matrix trilogy. I just ran it through eac3to 2.17 using the -stripPulldown and without it. Command line is on top.
Thanks!

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

[madshi]

I do own the US version of Phantom. I have the next several days off. Will re-rip it and try to convert to mkv. I'll get back to you soon.:helpful:
Great - looking forward to your report!