EVOB De/Multiplexers [Archive] - Page 12

edo1080

16th February 2007, 00:05

many compliments and thanks to pelican9 for his efforts.

I would like to know if is there any way to convert the .mpa audio track (DD+) obtained from the demuxing process into 6 mono .wav, even downgrading quality but keeping channel separation. Or at least to convert into 2 mono .wav

Thank you for all the suggestions

juhu

16th February 2007, 01:15

many compliments and thanks to pelican9 for his efforts.

I would like to know if is there any way to convert the .mpa audio track (DD+) obtained from the demuxing process into 6 mono .wav,

as all of us since 15-20 pages and even more days :D

woah!

16th February 2007, 01:27

why if demux shows a 29.97 framerate does it give you a total framecount at 23.976 ?

i am starting to lose the plot i think...

tomos

16th February 2007, 01:34

?? That's very odd, you should have it if you have the Sonic HD Nav filter. It's part of Scenarist 4.1.

Junooni, which version of Win DVD do you have installed? Maybe that's the magic ingredient.

never mind. i was being a gimp. found it. hadnt registered it for some reason

Pelican9

16th February 2007, 09:42

why if demux shows a 29.97 framerate does it give you a total framecount at 23.976 ?

i am starting to lose the plot i think...

because this:
Pulldown = 1

3ngel

16th February 2007, 11:53

Where do i read this "Pulldown = 1"? And when you demux it with EvoDemux you get only the frames without the pulldown?

kolak

16th February 2007, 12:08

Where do i read this "Pulldown = 1"? And when you demux it with EvoDemux you get only the frames without the pulldown?

Yes it would be nice:)
Pelican9 any solution?

Thx

Clown shoes

16th February 2007, 12:39

3ngel

16th February 2007, 12:45

I've tested right now, that the Pulldown flag is shown only on VC1 streams. On H264 it doesn't show all the informations, but only

H.264 video stream 0 found!
First PTS = 00001274
0 frames before first I-frame
High Profile
Level = 41
First Video PTS = 00001274

So a fix for this would be good.

Moreover i have to think that the demuxed from EvoDemux is without the Pulldown?
I ask this because it's not so obvious if the demuxing is "Pulldown respectful" or "Pulldown Ignoring".

I think that a switch would be necessary 'cause it's not so obvious that the "Pulldown ignored" frames will be the right (and the total) solution, giving the right 23.976 frames.
We can take for example all the experience from DgIndex where doing "Ignore Pulldown" can lead to a mess.
So what is the situation?

EDIT:
I can suggest an algorithm that compare the "Pulldown ignored" calculated duration (at 23.976), with the "Pulldown respectful" calculated duration (at 29.976).
If the duration is the same, then a "Pulldown ignored" demuxing can be done safetly.

Pelican9

16th February 2007, 15:07

3ngel: Did you read this?
Here's the VC-1 sequence header from KK:

SeqHeader: 1
+->Offset: 0
+->Size: 22
+->PROFILE: 3 (Advanced)
+->LEVEL: 3 (AP@L3)
+->COLORDIFF_FORMAT: 1 (4:2:0)
+->FRMRTQ_POSTPROC: 7 (30 frm/s)
+->BITRTQ_POSTPROC: 31 (2016 kbps)
+->POSTPROCFLAG: 0 (FALSE)
+->MAX_CODED_WIDTH: 959 (1920)
+->MAX_CODED_HEIGHT: 539 (1080)
+->PULLDOWN: 1 (TRUE)
+->INTERLACE: 1 (TRUE)
+->TFCNTRFLAG: 0 (FALSE)
+->FINTERPFLAG: 0 (FALSE)
+->RESERVED: 1 (TRUE)
+->PSF: 0 (FALSE)
+->DISPLAY_EXT: 1 (TRUE)
+->DISP_HORIZ_SIZE: 1919 (1920)
+->DISP_VERT_SIZE: 1079 (1080)
+->ASPECT_RATIO_FLAG: 1 (TRUE)
+->ASPECT_RATIO: 1 (1:1)
+->FRAMERATE_FLAG: 1 (TRUE)
+->FRAMERATEIND: 0 (FALSE)
+->FRAMERATENR: 3 (30*1000)
+->FRAMERATEDR: 2 (1001)
+->COLOR_FORMAT_FLAG: 0 (FALSE)
+->HRD_PARAM_FLAG: 1 (TRUE)
+->HRD_PARAM
| +->HRD_NUM_LEAKY_BUCKETS: 1
| +->BIT_RATE_EXPONENT: 3 (9)
| +->BUFFER_SIZE_EXPONENT: 4 (8)
| +->HRD_RATE: 45897
| +->HRD_BUFFER: 57599

30*1000/1001 = 29.97 with the pulldown flags describing a 23.976 encoded frame rate

Ron

3ngel

16th February 2007, 15:15

What i have to read? This?

30*1000/1001 = 29.97 with the pulldown flags describing a 23.976 encoded frame rate

That's not enough because there can be irregular pulldowns, in which one or more pulldowned (fake) frames became an integral part of the timeline, so removing them you lose synching (and you get wrong duration).

drmpeg

16th February 2007, 15:38

What i have to read? This?
That's not enough because there can be irregular pulldowns, in which one or more pulldowned (fake) frames became an integral part of the timeline, so removing them you lose synching (and you get wrong duration).
In the general case, you're correct. There can be irregular cadence of the pulldown flags. The overall rate of the stream can be anywhere between 23.976 fps and 29.97 fps.

In the particular case of film on HD-DVD (and especially with VC-1 HD-DVD), the pulldown flags are perfect and the stream really is 23.976 fps.

Ron

3ngel

16th February 2007, 16:09

Ok, i see and i can agree with you.
Mine was a suggestion in order to do a "perfectly compliant" routine. Because it's absolutely not guarantee that with future HDDVD there will be a perfect pulldown cadence.

Anyway a little suggestion for EvoDemux. Add a "Clear" button in order to clean up the report window.

Pelican9

16th February 2007, 17:09

Anyway a little suggestion for EvoDemux. Add a "Clear" button in order to clean up the report window.

You can clear that window manually. :)

3ngel

16th February 2007, 17:19

Ops... :)
Naa, but that's not professional :p
A professional one, is a readonly with a clear button :)

Pelican9

16th February 2007, 18:33

Ops... :)
Naa, but that's not professional :p
A professional one, is a readonly with a clear button :)
Done. You'll see in v0.617

chros

16th February 2007, 21:06

@Pelican9: can you add an option to demux/recreate chapters into ogg or mkv xml chapter format from *.xpl?

Thanks

3ngel

16th February 2007, 21:14

Done. You'll see in v0.617
:cool:

Can you add a little more informations on H264 streams (like VC1)?

Pelican9

16th February 2007, 21:19

Are you a mind reader? :)
I'm working on it...

Pelican9

16th February 2007, 21:21

@Pelican9: can you add an option to demux/recreate chapters into ogg or mkv xml chapter format from *.xpl?

If you specify the format of this, why not?

3ngel

16th February 2007, 21:43

Are you a mind reader? :)
I'm working on it...

Uhuhu :D

chros

16th February 2007, 22:12

If you specify the format of this, why not?
Ogg: is too easy (only problem is the milliseconds, of course anything can stand after the ChapterxxName):
CHAPTER01=00:00:00.000
CHAPTER01NAME=01. Costello Credo
CHAPTER02=00:05:03.300
CHAPTER02NAME=02. School's Out
CHAPTER03=00:08:52.000
CHAPTER03NAME=03. You're No Cop

Matroska chapters xml:
http://www.bunkus.org/videotools/mkvtoolnix/doc/mkvmerge.html#CHAPTERS
Or you can see the examples in the MKVToolnix distribution (C:\Program Files\MKVtoolnix\examples\)

Ogg is very simple, and since mkvmerge is able to import it, it will good enough for now...

Thanks, and greets from the same country ... :)

2439rf8hasjdert4

16th February 2007, 23:13

http://isochroma.com/Testfiles/Misc/doom9/EVOB-graphedit-1-sm.png (http://isochroma.com/Testfiles/Misc/doom9/EVOB-graphedit-1.png)
Click image for full size

I do not have Sonic Scenarist.
Is there any alternative?:confused:

Pelican9

16th February 2007, 23:42

Change for v0.617

- Fixed bug(?) with .sup export (decrease PTS with value of the first video PTS)
- Clear button on Status page

EVOdemux v0.617 (http://pel.hu/down/EVOdemux.exe)

woah!

17th February 2007, 02:11

The pulldown is simply a flag that tells the decoder how to handle the stream. You can see it in Chros's sample code from EVODemux

Code:

Opening file MAININTRO.EVO
File size: 130 Mbytes.
It contains 1 video stream and 1 audio stream.
VC-1 video stream 0 found!
First PTS = 00000D75
Sequence Header found
0 frames before first I-frame
Advanced Profile
Level = 3
Chroma Format = 4:2:0
Size = 1920x1080
Pulldown = 1
Interlace = 1
Display Horizontal size = 1920
Display Vertical size = 1080
Aspect ratio = 1:1 (square samples)
Frame Rate = 29,97002997003 (30000/1001)
DD+ audio stream 0 found!
First PTS = 00000D75
Stream 0 is Dolby Digital Plus
frame size = 1536, number of blocks per frame = 3
bitrate = 768000, sampling rate = 48000
mode = 3/2, bsid = 16, lfe on =1
Done.

1541 frames (0:01:04 s) count in video stream 0.

The true frame count should not change as there is no more real data present.

ok but what i am finding is if i enter the resulting framecount into my script, i get a (eg: 1:20:02 timeline when it should be 1:42:45)

if i calc the amount of frames at 29.97 and add it to my script then it runs at the correct timeline.

i would like the option to pick / switch on or off the pulldown option and dump all the frames at 29.97. i can use tfm(order=0).tdecimate() afterwards if needed anyways.

i think its because of how sonic's decoder is looking for 29.97 source not an already pulldowned evob. and as i am using sonic filter to pas to my avs script i think that is where the problem is coming from.

edtchie

17th February 2007, 02:32

Change for v0.617

- Fixed bug(?) with .sup export (decrease PTS with value of the first video PTS)
- Clear button on Status page

EVOdemux v0.617 (http://pel.hu/down/EVOdemux.exe)

I am not sure I missed something but EVOdemux is not working properly on Canal released HD DVD title: Ramblo, First blood and Total Recall. It looks the structure is a little different, and it stoped when finished first evo file, not continued on the second one.

orbitlee

17th February 2007, 05:03

I am not sure I missed something but EVOdemux is not working properly on Canal released HD DVD title: Ramblo, First blood and Total Recall. It looks the structure is a little different, and it stoped when finished first evo file, not continued on the second one.

Additional information, the EVO files are named as
L0_mainMovie.EVO
L1_mainMovie.EVO

instead of
FEATURE_1.EVO
FEATURE_2.EVO

Clown shoes

17th February 2007, 12:07

I'll go out on a limb and guess that renaming them will probably fix the problem then.

MichalHabart

17th February 2007, 17:23

I'll go out on a limb and guess that renaming them will probably fix the problem then.

Yes, exactly. Renaming to feature_x.evo is solving the problem. But still with Rambo there is problem with english dts-hd. german and french ones have 1005mb each (it means 93 minutes) but english one is demuxed to 1220mb (it is 115 minutes).

Clown shoes

17th February 2007, 18:40

Is it possible the English DTS-HD is a higher bitrate, hence the larger size?

Pelican9

17th February 2007, 20:26

Additional information, the EVO files are named as
L0_mainMovie.EVO
L1_mainMovie.EVO

instead of
FEATURE_1.EVO
FEATURE_2.EVO

Thanks. I'll fix it.
Question: Every HD DVD released by Canal have these filenames?
Is there any other name pattern?

Schmendrick

17th February 2007, 21:15

Pelican9

17th February 2007, 21:51

Changes for v0.618

- Fixed bug with 'Continue on second EVO' (with Canal's filenames like L0_mainMovie.EVO)
- New filenames if 'Continue on second EVO' checked (FEATURE_1_2 -> FEATURE_1_FEATURE_2)

EVOdemux v0.618 (http://pel.hu/down/EVOdemux.exe)

orbitlee

18th February 2007, 04:16

Thanks. I'll fix it.
Question: Every HD DVD released by Canal have these filenames?
Is there any other name pattern?

Looks like Canal use this name pattern so far. But another name pattern(in "Oldboy") may confuse your algorithm again.
PEVOB_11MAIN_1.EVO
PEVOB_11MAIN_2.EVO

In theory, the name of two EVOs could be anything, foo.EVO and bar.EVO. Also some HD-DVDs have single EVO file only

So I suggest to add an input box for the second EVO. User choose the first EVO, you guess the name of second EVO. If file does not exist, an alert will be raised to ask users to choose the second EVO(or cancel for single EVO) manually.

woah!

18th February 2007, 07:31

well i have found the best way for me to get things working is to forget demuxing and just feed the evob to graphedit and then into my script. seems to keep sync with the audio except for some delay at the start that can be fixed.

until another filter comes along i dont think the sonic one works well with a demuxed file, or something is happening between them for me anyways.

Pelican9

18th February 2007, 07:55

Looks like Canal use this name pattern so far. But another name pattern(in "Oldboy") may confuse your algorithm again.
PEVOB_11MAIN_1.EVO
PEVOB_11MAIN_2.EVO

In theory, the name of two EVOs could be anything, foo.EVO and bar.EVO. Also some HD-DVDs have single EVO file only

So I suggest to add an input box for the second EVO. User choose the first EVO, you guess the name of second EVO. If file does not exist, an alert will be raised to ask users to choose the second EVO(or cancel for single EVO) manually.
The program recognises the following patterns:

*_(0..8)*.evo --> *_(1..9)*.evo
*(0..8)_*.evo --> *(1..9)*.evo

*: any alphanumeric character (except '_')
(0..8): a numeric digit between 0 and 8
(1..9): an incremented numeric digit between 1 and 9

2439rf8hasjdert4

18th February 2007, 09:23

@ 2439rf8hasjdert4:

If you have WinDVD 8 Platinum installed a VC1-video decoder will be installed and you can build the following graph in Graphedit (see attachment below) used as a feed to other programs for reencoding to a different format :

DirectShowSource("D:\Audio_Video\vc1.GRF",audio=false,video=true,fps=171.7)

Strangely the frame rate has to be increased so that all frames are decoded.

Schmendrick

Thanks, but I do not have WinDVD. Is there a way to make it work with Cyberlink's PowerDVD?

Also, my file ends in EVO, not VC1. Did you have to do some kind of demuxing to get VC1? And does that VC1 have audio as well?

:thanks:

Schmendrick

18th February 2007, 10:23

Thanks, but I do not have WinDVD. Is there a way to make it work with Cyberlink's PowerDVD?

I think the right version of PowerDVD which has VC1-decoding capabilities also has a VC1-capable demuxer and decoder.

Also, my file ends in EVO, not VC1. Did you have to do some kind of demuxing to get VC1? And does that VC1 have audio as well?

So far I only have used a VC1-sample I had downloaded and a demuxed VC1-sample from a WMV-clip I had produced using the VC1/WVC1-WIndows Media Advanced Profile encoder, but this thread is about EVO-demuxing so Pelican9's-demuxer should be able to produce a VC1-file from EVO-files.

Schmendrick

klona

18th February 2007, 10:51

Hi.
Really great tool I just discovered.
Many thanks Pelican9 for this little baby.

One question, regarding audio delay. Is taht possible to know what is the orginal delay in the HD DVD ? Re-finding it manually to synch a re-encoded h264+ac3 is quite boring..

Pelican9

18th February 2007, 10:51

Change for v0.619

- Display DTS HD (DTS) audio stream information

EVOdemux v0.619 (http://pel.hu/down/EVOdemux.exe)

Pelican9

18th February 2007, 10:58

One question, regarding audio delay. Is taht possible to know what is the orginal delay in the HD DVD ? Re-finding it manually to synch a re-encoded h264+ac3 is quite boring..
What does 'original delay' mean?
You can see the first PTS of the streams (divide by 90000 to seconds).
Is it help for you?

klona

18th February 2007, 11:45

(-:
Thx. Did not notice the PTS data !

woah!

19th February 2007, 08:46

well after "alot" for head scratching i have finally got a long winded way to backup my hddvd source to a x264/aac audio encode.

here's a clip of kong at 720x304 : hxxp://file2upload.com/file/27157/KONG-3958-3990-rar.html

i can of course do whatever res i wanted, but am testing upto now. i used 2ch aac sound at 256kb and video was using --crf 23

now i used anydvd to backup the disc, also i only demuxed the audio to a .ddp file, and put the evob into graphedit directly for the video. this gave me good results. i found the audio was exactly 1200ms delayed after some fiddling.

juhu

19th February 2007, 09:06

now i used anydvd to backup the disc, also i only demuxed the audio to a .ddp file, and put the evob into graphedit directly for the video. this gave me good results. i found the audio was exactly 1200ms delayed after some fiddling.

(scratch scratch...)
Do you have some special magic to turn a ddp 6ch audio stream into 2ch aac... ?
Because no one here was able to fully decode ddp (only 2 streams...) till now

tomos

19th February 2007, 14:13

i went over a page or 2 ago how it worked forme too. although there was no delay in mine

idamien

19th February 2007, 19:08

I just bought and installed Vista Ultimate... Tried to do a x264 encode backup of one of my HD-DVD disks using MeGUI but x264 crashes everytime.

Has anyone had any success with the latest MeGUI and x264 under Vista Ultimate 32 Bit???

I was considering buying the MainConcept encoder... until I saw the price: US$600.00 is too much when you have a free (and very good) alternative IMHO.

sillKotscha

19th February 2007, 22:52

here's a clip of kong at 720x304 : hxxp://file2upload.com/file/27157/KONG-3958-3990-rar.html

you should investigate more time into "how to encode using x264/MeGUI"... 'cause your encode could be considered as bad advertisement for MeGUI using x264 :o

I mean, you told us you used a HD video stream as your source and your resulting video clip is 8,14MB of size for 31sec ... c'mon

http://img255.imageshack.us/img255/5568/snapshotmv8.png (http://imageshack.us)

I bet you can do better or you should think about leaving the source untouched ;)

P.S.: of course this is just one frame without any further analysis... but nonetheless the result is insulting for x264

Clown shoes

19th February 2007, 23:46

@sillKotscha

This thread is about demuxing evobs which is exactly what woah! is doing. He clearly mentions in his post that he is testing out different methods to do such. Right now this is not quality discussion and I see no reason to flame someone who is working hard to find a solution we would all benefit from.

hd1080p

19th February 2007, 23:58

That explains why the picture looks blurry.
you should investigate more time into "how to encode using x264/MeGUI"... 'cause your encode could be considered as bad advertisement for MeGUI using x264 :o

I mean, you told us you used a HD video stream as your source and your resulting video clip is 8,14MB of size for 31sec ... c'mon

http://img255.imageshack.us/img255/5568/snapshotmv8.png (http://imageshack.us)

I bet you can do better or you should think about leaving the source untouched ;)

P.S.: of course this is just one frame without any further analysis... but nonetheless the result is insulting for x264