DGMPGDec 1.6.0 Development thread [Archive]

burfadel

9th June 2008, 16:48

An updated mpeg2 decoder. I believe the current one is based on an old decoder, and the decoding part is the same? I'm assuming there are many optimisations that can be done to it to make it noticeably faster for HD use (maybe even multithread it). I realise that this part is not done by yourself, but I'm sure there would be some keen people out there with programming knowledge that would be keen to update the decoder. Its not that the current decoder is slow, just that due to its threading etc. etc. it would have a noticeable impact on the final encoding speed.

Tima

9th June 2008, 18:22

A multithreaded indexer would be nice. Sometimes, the indexing speed is the bottleneck, not the speed of hard drive. E.g. when VOB is in windows file cache and you have no swap, so it's in RAM. :)

Another request is to write information in documentation about the offsets in Location value -- as for me, I don't understand what units are they in :(

Also repeating bug from 1.5.0 thread: http://forum.doom9.org/showthread.php?p=1146938#post1146938

vlada

9th June 2008, 18:41

I have two requests for command line:

1) Allow to display video output during indexing even if it is started from command line.

2) Add command line switch to index only a part of video.

Both things are possible from GUI so this shouldn't be a lot of work to implement (at least I hope so).

=A=RGOS

9th June 2008, 18:59

Hi,

For this thread, I propose a port of DGMPGDec to Linux :)
But, if you separe the program into a two part (GUI and command line) it's possible.

Bye,

Sharc

9th June 2008, 19:32

Would it be possible to include - as an option - auto-correction of fishy audio delay values (such as 87650ms for example)?

GrofLuigi

10th June 2008, 06:49

A "close" menu item.

GL

Zep

10th June 2008, 08:52

sunbeam

10th June 2008, 18:11

Compatibility with W2K as continued in 1.5.x would be great :)

fbgd

10th June 2008, 20:23

Hi,

For this thread, I propose a port of DGMPGDec to Linux :)
But, if you separe the program into a two part (GUI and command line) it's possible.

Bye,
I've never had a problem running it with wine. Plus wouldn't that require a native linux version of avisynth?

Guest

10th June 2008, 23:49

Adding this request in PM from morphinapg:

Thank you and I'm happy to tell you I've been using the

__aud__

to get the first audio track to use for my ASXGui and it works great. But someone just requested a feature to include support for the tracks other than the first. So my question is, keeping with the same idea that the track order is decided by the track ID, could it be possible to add things such as

__aud2__
__aud3__
__aud4__

etc up to maybe 8 tracks or so, and if there weren't that many tracks then they would simply not return anything.

It's not a huge priority just maybe something I would like to see down the road as it would be helpful for my program, and other people may benefit from it as well.

=A=RGOS

11th June 2008, 07:22

I've never had a problem running it with wine. Plus wouldn't that require a native linux version of avisynth?

Yes it work fine on wine :), but a native avisynth version is under developpement (v3.0).
And a DGMPGDec on Linux with native build is more appropiate for certain use.

bye bye

hajj_3

11th June 2008, 19:08

just wondering if something changed in final build of dgmpgdec, when i open a .mpg and choose save project normally it creates a .mpa file as i changed the option in dgindex.ini from .mp2 but it doesnt create any by default, i have to choose "demux all tracks" instead of the default "demux track". I dont remember having to change this in any previous builds, did this change in the final build?

everything is running nicely with this build btw, has done since RC5.

Guest

11th June 2008, 19:23

The Select Tracks dialog defaults to empty. If you choose Demux Tracks, then you have to specify one or more tracks in the Select Tracks dialog.

smok3

11th June 2008, 19:51

a. if i understand correctly, wav files decoded from ac3 are allready delay corrected? Same request/wish for mp2 audio (decoder + delay correction if neccesary)

(I'am using it from cli as well)

b, another thing, basic info about the stream processed (video type, resolution, ipd, audio type, #of channels, # of tracks) dump to __info__ or similar var to be used with avisynth template. So my template could be, example;
a=mpeg2source("__vid__")
b=wavsource("__aud__")
audiodub(a,b)

__END__

__info__

guada2

11th June 2008, 22:06

Hello neuron2,

Just 2 things: :)
- MLP (audio)
- VC-1 (video)

Underground78

12th June 2008, 20:40

Hello,

This is not a serious bug but sometimes when I navigate in a TS file using the "<" button to go the previous I-Frame, I get stuck at a frame and I can't go back in the file until I use the navigation bar. If I go to the next I-Frame using the ">" button when I'm stuck, it works but if I retry to go back I get stuck at the same frame. When I have such a frame in a file, the bug can always be reproduced. I've never seen this bug when I navigate next I-Frame to next I-Frame.

I've tried to give you a sample but when I demux the video with DGIndex, I can't reproduce the bug in the resulting file. If there is another way to cut the file, I would be happy to try ! :)

Guest

12th June 2008, 20:48

There have always been issues with stepping backwards by GOP. I'm aware of them but it's cosmetic so likely I won't do anything about it as it won't displace higher priority things. It's hard to give me a stream because if you cut from the beginning it changes things such that the problem may not be triggered.

Underground78

12th June 2008, 22:08

Guest

12th June 2008, 22:56

No plans yet. That's what this thread is for.

Sharc

15th June 2008, 08:46

Don't know if this is a bug or feature of DGIndex.
I have a VTS with 2 PGCs: PGC1=Movie, PGC2=studio intro (separate VOB ID with 1 cell of 12 seconds duration).

Demuxing with DGIndex produces 2 audio tracks: 0x80(en) and 0x81(de).
The demuxed (0x80 en) stream includes the (en) audio of the movie and the (en) audio of the studio intro. DELAY= 0ms.
The demuxed (0x81 de) stream includes the (de) audio of the movie only, but no audio for the studio intro. DELAY =-16ms.
Apparently the studio logo uses the same audio stream 0x80 for both languages (en and de), as reported by VobBlanker for example.

The problem comes with re-muxing: The (de) audio is then 12 seconds out of sync (ahead of the video) because of the missing audio of the studio intro in the 0x81 stream, or due to the low delay indication of -16 ms instead of +12 seconds. Don't know if and how DGIndex can or should handle such cases.

Added:
Workaround id to step with > button to the first cell of the movie and pressing [, then Save Project, i.e. skipping the intro pgc/vob id. This will produce synchronous audio tracks with same DELAY for the movie.

Malow

15th June 2008, 21:29

perfect sync with broken HDV clips? :p

Sharc

17th June 2008, 06:39

Wish for future release:
- Option for indexing and demultiplexing split according to PGCs, not only on a per file basis (including all PGCs)
- Angle selection

morphinapg

17th June 2008, 06:51

I have some HD recordings that switch between 5.1 AC3 and 2.0 AC3 between commercials. Is it possible to have an option to demux all of the 2.0 AC3 to one file and all of the 5.1 AC3 to another (with empty space filling the rest)? Either that or split on audio format change. I'd prefer the first option, but I could live with the second.

Otherwise I have horrible sync issues that are often not fixable.

Guest

21st June 2008, 02:08

Adding request in PM from LordIntruder:

Some months ago I ripped a movie, PAL and progressive, around 1h50 long. From 1h40 to 1h45 suddenly the sequence was interlaced and then the last 5 minutes progressive like the rest of the movie.

This week I encountered the exact same problem with an old serie. In the middle of the episode, around 10 min, suddenly everything is interlaced.

Could you implement such an option to tell us automatically that the video has both progressive and interlaced content? Because as usual, I always see that AFTER the encoding Grrr...

I thought that maybe when we save the project (F5), at the end in the log file, we could have a report saying:

- Both progressive and interlaced

Or even greater:

Say a movie has 150000 frames.

- frame 1 to 57999: progressive
- frame 58000 to 65000: interlaced
- frame 65001 to end: progressive

Here we could see immediately where in the stream the interlaced content is, go to that area and test a small encode with the best deinterlace settings.
------------

So I don't know if this can be done or not, I have no knowledge about that but if this is possible, it would be very great.

Ranguvar

21st June 2008, 05:57

Definitely what LordIntruder suggested... also, please more info (with frame/time ranges) of what's flagged as telecined, etc.

Zep

21st June 2008, 07:08

Definitely what LordIntruder suggested... also, please more info (with frame/time ranges) of what's flagged as telecined, etc.

yes yes I third that! that would be a very awesome feature since you could use trim and or range to process those areas differently and with hybrid source you could target areas with a slower but much better frame blender/double.

stax76

21st June 2008, 16:04

@neuron2

The -minimize cmdl switch isn't really necessary I think. The window state can be defined launching a external process, even from the cmdl shell using the START command. Currently minimizing does not work launching DGIndex from code as only the main window minimizes but not the preview dialog, minimizing manually the preview dialog hides, minimizing any other method it don't hide. I was thinking maybe there is a way I can start DGIndex from code without the preview to show, first thing I tried starting indexing manually and clicked the preview away, this causes bug #2, DGIndex is freezing.

LoRd_MuldeR

22nd June 2008, 17:02

I got a question: When I load an MEPG PS file (e.g. captured from my DVB-S card), save the project and demux the audio, I get an audio file like "FOOBAR T01 DELAY XXXms.mpa". This delay value is very useful to me, because it helps to synchronize the audio. But I don't need the audio file, nor do I need the D2V file. I do all my work on the MPEG file. All I need is the delay value. So is there a way to simply display that delay value in DGIndex? If not, which I think is the case, could it be added please? Like "Tools > Calculate Audio Delay" or something...

Sharc

28th June 2008, 17:53

I wish under "Options" a 16 x 16 grid overlay be shown. It would help for cropping and estimating how the useful picture fits the 16 x 16 macroblocks.

Revgen

29th June 2008, 06:24

Can you add a BFF/TFF switch?

I cap HDTV sports programs and the field order keeps changing from TFF to BFF. The D2V repair tool works fine to fix that, but it fixes it according to the field order at the beginning of the video. I always want my resulting .d2v file to be TFF. I currently can do that using Stickboy's JDL_ReverseFieldDominance(), but it would be nice if it could be changed within DGIndex.

Thanks.

Ranguvar

29th June 2008, 20:47

And, could you add a line to the log file that reports where (in frames) any corrected field order transitions occurred?

GrofLuigi

30th June 2008, 01:25

Support for Sharc's idea - 16X16 grid (switchable, of course) and Revgen's idea (force the output stream to become TFF/BFF only). :thanks:

GL

mikeytown2

1st July 2008, 21:24

I'll go ahead and request this for SynchronousArts and spyros78. Reading of the SMPTE timecodes (http://forum.doom9.org/showthread.php?t=138139).

Me personally, I also would like the decoder to be multi threaded.

krosswindz

4th July 2008, 18:44

I just upgraded to DGIndex 1.5.0 I am just wondering if this was a problem with DGIndex 1.4.9 or is it a bug in DGIndex 1.5.0

The same DVDs colorimetry information is different. 1.4.9 reports bt709 while 1.5.0 reports the sames bt407

v1.5.0
http://img66.imageshack.us/img66/8110/dgi15zg2.jpg

v1.4.9
http://img354.imageshack.us/img354/723/dgi149bg8.jpg

Guest

4th July 2008, 23:07

Did you read the change log? Change number 21 would be germane.

http://neuron2.net/dgmpgdec/changes150.html

krosswindz

5th July 2008, 00:04

Did you read the change log? Change number 21 would be germane.

http://neuron2.net/dgmpgdec/changes150.html

makes sense now... :thanks:

Ranguvar

13th July 2008, 06:08

The ability to demux the PCM WAV audio from Blu-ray M2TS files... :)

SynchronousArts

14th July 2008, 23:58

I'll go ahead and request this for SynchronousArts and spyros78. Reading of the SMPTE timecodes (http://forum.doom9.org/showthread.php?t=138139).

@mikeytown2,

Thanks for adding this here. I added some idea's to that thread about how the Timecode could function.
I recently was given a project with about 20 HDV camera tapes that have discontinuous Timecode all over them.
Would have been a big help to have this working.

Thanks

SA

Sharc

15th July 2008, 06:57

Would it be possible to add in the information - and log file - the Pixel Aspect Ratio (PAR, SAR) of the source? If it is not specified in the stream one could make a best assumption marked with an * ( similar to the colorimetry).

Blue_MiSfit

24th July 2008, 22:19

Hey folks,

Wondering if DGIndex supports demuxing SMPTE 302M (AES/EBU audio stored in MPEG-2 Transport Stream)?

I may have to handle some of this stuff in the future, and it would be lovely if DGIndex could automagically handle it ;)

~MiSfit

Guest

24th July 2008, 22:36

Do you have the specification?

mikeytown2

24th July 2008, 23:23

looks like someone attempted demuxing SMPTE 302M back in 2006 for ffmpeg
http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2006-September/thread.html#15650

Dead end over on hydrogenaudio, only 3 posts
http://www.hydrogenaudio.org/forums/index.php?act=Search&CODE=show&searchid=f7b75b51c414102a2e1e5b82e19d4db8&search_in=posts&result_type=posts&highlite=SMPTE+302M

The specs looks like it will cost $35
http://store.smpte.org/SearchResults.asp?Search=302

or $22 for a 2002 edition
http://www.techstreet.com/cgi-bin/detail?product_id=1249916

Or wikipedia
http://en.wikipedia.org/wiki/AES/EBU

Which lead to page 22 of this pdf
http://www.ebu.ch/CMSimages/en/tec_AES-EBU_eg_tcm6-11890.pdf
this might be enough to get an implementation going...?

Library Of Congress is of no help
http://www.digitalpreservation.gov/formats/fdd/fdd000208.shtml

Blue_MiSfit

25th July 2008, 00:06

Wow - thanks for doing research for me mikeytown2!

I don't have the spec unfortunately... if I did I would certainly forward it

~MiSfit

Guest

25th July 2008, 00:31

You can buy it if it is that important to you.

Blue_MiSfit

25th July 2008, 02:33

Indeed I will if the need arises!

CruNcher

26th July 2008, 18:53

Donald http://libmpeg2.sourceforge.net/ it was recently updated twice do you think it makes sense to update the core of Dgdecode now with it or are the changes not so important ?

Guest

26th July 2008, 18:55

DGDecode uses an internal decoder, not libmpeg2. It would be a major rewrite to replace it.

stattik

28th July 2008, 05:43

Hi neuron2,

How about support for decoding AC3 to WAV for TS files?

Guest

28th July 2008, 12:36

Hi neuron2,

How about support for decoding AC3 to WAV for TS files?
It already is supported. Set the audio PID and then specify track 0 as the audio track, as described in the users manual.

stattik

29th July 2008, 16:14

It already is supported. Set the audio PID and then specify track 0 as the audio track, as described in the users manual.

Sorry, I was thrown by this in the manual:
A future version of DGIndex may integrate the WAV packaging functionality for transport streams.

Looks like it was referring to M2TS containers.

It's working great now. Thanks for the help.

Guest

29th July 2008, 17:46

That referred to demuxing of LPCM audio from a transport stream and eliminating the need for sox.

mrg

1st August 2008, 10:56

Hi neuron2,
to use AVS templates, i need the filename but without the extension (__vid__ without ".d2v"), because my audio files already exists with the same file name. I believe __file__ would be good.

Ranguvar

1st August 2008, 20:04

Your audio files have the D2V extension!? The heck?

Inventive Software

1st August 2008, 23:02

I spotted a d2a file the other day from some demuxing....

mrg

2nd August 2008, 09:15

Your audio files have the D2V extension!?
No, sorry for my bad English.
I have separate files like "video1.mpv" + "video1.mpa". With the help of the avs-template-function i would that DGIndex produce a avs like:
LoadPlugin("DGDecode.dll")
LoadPlugin("NicAudio.dll")
video=MPEG2Source("video1.d2v")
audio=NicMPASource("video1.mpa")

Underground78

2nd August 2008, 09:34

Maybe you can use __aud__ :

You may also use the __aud__ specifier to generate the audio file name. Note that if you are generating more than one audio file, this specifier will refer to the first audio file that is opened. Therefore, to be sure to get the desired audio file, process just one audio stream.

Guest

2nd August 2008, 12:44

That won't work because the __aud__ macro is applied only if DGIndex is demuxing the audio, and the poster says the audio already exists. I can add a macro that generates only the base name without the extension. But the __del__ macro won't work then, of course.

J_Darnley

3rd August 2008, 17:43

Use the available Avisynth features
vid_file = "__vid__"
aud_file = LeftStr(vid_file, StrLen(vid_file)-4)
Might need checking but should return the full file path and name but remove the extension and the dot.

Guest

3rd August 2008, 18:04

Then append the desired extension.

Thanks for that. Saves me some work. :)

mikeytown2

6th August 2008, 00:36

Template Request: Using Trim() make each title/chapter from a DVD a different variable in avisynth.

Example Output

#Load AV; true 29.97i video
MPEG2Source("GreenV3.d2v", cpu=6)
AudioDub(WAVSource("GreenV3 Ta0 48K 16bit 2ch.wav"))

#Process Clip
ColorMatrix(d2v="GreenV3.d2v",threads=0,clamp=0,interlaced=true,outputFR=true)

#Split into titles/chapters
DGI_T1C1=Trim(0,999)
DGI_T1C2=Trim(1000,1999)
DGI_T1=Trim(0,1999)
DGI_T2C1=Trim(2000,2999)
ect...

If i load a VOB that has no audio in it when the other VOB's do, the resulting script is out of sync. Add option to pad video with blank audio so it stays in sync?

jeffy

6th August 2008, 06:41

Please add the logging of the "Block error" info if and where exactly in the stream they (the errors) are.
Ref.:
http://forum.doom9.org/showthread.php?p=1167085#post1167085
Thank you.

You Know

10th August 2008, 14:37

Is possible to show in main editor also non Key-frame when arrow keys are pressed?
And jump as before from a key to other with a combo SHIFT + -> for example.

aand

20th August 2008, 17:32

dukey

20th August 2008, 23:13

Add an option to start cutting from the first I frame
Also, make the audio delay 0ms

Export does this, and it's really handy.

If the first frame is a P or B frame, really there is no point in us using it, if we want to re-encode. That is all, thanks for a great program :]

Guest

21st August 2008, 02:18

I'm having a problem with a .VOB file. DGIndex 1.5.2 makes the project file, but not to the end. And the demuxed audio file grows as big as the free space on the hard drive.
Here's a link (http://www.filebox.ro/download.php?key=80cc814e39da958051d83a0e8550e2da) to the file. Your upload host is offering me 2KB/s. Sorry but I cannot wait 30 hours to download your file. Try a decent upload site.

aand

21st August 2008, 12:55

Your upload host is offering me 2KB/s. Sorry but I cannot wait 30 hours to download your file. Try a decent upload site.
:o sorry 'bout that, they used to be decent.

I can't recreate the problem, DGIndex works correctly now.

PS: I tried to register to your forum but didn't get the activation email.

Guest

25th August 2008, 04:06

Is there some way to modify DGMPGDec to support streams that have sequences with different resolutions? For example, in a transport stream, sometimes the commercials come as sequences with lower resolution. Currently, this crashes DGMPGDec.

Fallen

28th August 2008, 11:28

CLI Add option >> -Location[f,startlba,f2.endlba]
Same format as location in .ini

My program is opening up to a hundred vobs (making motion menu thumbnails in Avisynth) and using about 15 sec from each.
Fully processing each vob takes about 45 sec. (or up to 1.25 hr)
Whereas processing (with start end lba's) takes less than a sec per vob) -- and life would be good

Studied the DGIndex source code (actually tinkered the change in) but my old VC++ 6.0 Standard Edition won't allow MASM

Also noticed that DGIndex will open a partial D2V (Header entries with location written in -- No Frame Info) Via GUI Load Project

But the cli won't load a D2v (To bad that also would have done what I need)

Directshowsource is giving me hell (won't load Multiple vobs Codec limitation)-- Even partially hacking a MPG GRF file to replaceable parameters -- works about 80% of the time --missing something

--- DGIndex would solve all my Problems

Thanx

Guest

28th August 2008, 14:07

I'll make it for you this evening.

BTW, I build with VC++ 6.0. You can find NASM on the web.

Fallen

28th August 2008, 14:45

Thanx :thanks::thanks::thanks::thanks: Thanx

Not nasm (I caught that -- searched -- and you explained to another) no problems there
Masm ---idctmmx settings
ml is from the masm compiler

Bet your 6.0 ver is either express or pro and you loaded a processor pack

Guest

29th August 2008, 14:01

I have it working but I want to fix the problem with the mising trackbar selection when using a themed desktop before releasing anything. May take a few days.

I can give you a test build if you like. Send me a PM.

Slammin'

29th August 2008, 18:59

Please enter your feature requests and bug reports for version 1.6.0 here. I would appreciate it if you have already entered them in the 1.5.0 final thread if you could repeat them here. Thank you.

FWIW : it would save me a whole lot of time in front of my screen if DGIndex had a switch to batch output what I batch input.

(ie... open
Video01.m2v
Video02.m2v
etc...

..save
Video01.d2v
Video02.d2v
etc...)

No biggie, but again, it sure would save me a whole lot of time sitting in front of my screen waiting for an MPEG to process so I can load the next.

Fallen

29th August 2008, 20:05

It already does --study the CLI section of the help file

Slammin'

30th August 2008, 19:38

It already does --study the CLI section of the help file

Heh.. funny, I've been using GUI so much recently, I forgot about that whole section. Thanks for the heads up.

IMHO, I still think there should be a switch. :devil:

Using the CLI "for this" is kind of like having the obvious buried.

Just my 2 pennies..

freezer

1st September 2008, 15:44

I second the need for a gui batch switch. Currently all input files are appended, I would like to output them as individuals.
I didn't even know that you could do that via CLI until now, but anyway - why make it harder for users than necessary?

Guest

1st September 2008, 16:54

I don't plan to support batching. There are existing solutions based on invoking the CLI in a DOS batch file. Search and ye shall find.

stax76

2nd September 2008, 00:51

StaxRip is such a solution for instance.

Guest

2nd September 2008, 22:13

Add mouse scrollwheel support and Copy Frame to Clipboard, as in DGAVCDec.

Add GPU decoding support, e.g., CUDA Video API.

d'Oursse

6th September 2008, 19:45

a request i would like for dgindex would be : separate the gui from the rest of the code, and make the rest of the code a lib.

also, i've read quickly that you use nasm. Maybe using yasm would be better. It's more maintained and it supports 64 bits instructions

mikeytown2

6th September 2008, 21:20

Version 2 of NASM supports (http://www.nasm.us/doc/nasmdo11.html) x86-64

d'Oursse

6th September 2008, 21:34

ha, indeed, the nasm development has been restarted for 1 year

Dark Eiri

8th September 2008, 17:37

Add mouse scrollwheel support and
Add GPU decoding support, e.g., CUDA Video API.

That would be absolutely awesome.

Is would be great if it was possible to use DXVA decoding, so we could deinterlace the video "on-the-fly" via hardware. Is that possible? (I know using DirectShowSource with decoders like Cyberlink works, but the audio gets offsync very often)

GPU accelerated deinterlacing would remove a big bottleneck on encoding, wouldn't it?

Guest

8th September 2008, 18:17

Inventive Software

9th September 2008, 10:26

This may sound a bit pointless at this stage, but what's the maximum throughput you get with using CUDA vs not using it? (MPEG-2)

SledgeHammer_999

9th September 2008, 12:04

Just to inform you the ATI equivalent off CUDA is called "AMD Stream". link-->http://ati.amd.com/technology/streamcomputing/index.html

Ranguvar

9th September 2008, 15:38

Just a note, I generally find that people have very good things to say about the PureVideo deinterlacer.

Dark Eiri

9th September 2008, 15:39

It's very easy to turn on the deinterlacer with CUDA. I'll be experimenting with it to see how much it impacts the frame rate on the low-end and high-end cards, and how good the deinterlacing looks. All I know is that it is claimed to be "motion adaptive". You can Google for "PureVideo deinterlacing" to see marketing hype and reviews.

The deinterlacer you would use in CUDA is the same "Hardware Deinterlacing" used in normal players supporting PureVideo, like PowerDVD Ultra? That one is very nice, actually. I won't say it would be on par to Yadif in all cases, but it's really good.

mikeytown2

12th September 2008, 08:08

output split/cut/trimmed m2t files

stattik

12th September 2008, 17:19

morphinapg

12th September 2008, 18:32

Would you consider correcting A/V continuity errors for TS Files? It would be quicker than using multiple apps to scan each file, demux/fix audio, fix video, and then remuxing audio/video only to have dgindex demux the audio for processing again.

I second this, as well as having the ability to only write a certain format of audio, filling in the gaps with silence. Some TV shows I capture have 5.1 AC3 during the show but 2.0 AC3 during the commercials and right now I have to use project X to demux it and replace all 2.0 audio with silence, and then use DGIndex afterwards. It would be much simpler if it could all be done in DGIndex.

Malcolm

5th October 2008, 14:20

There were some performance issues with DGIndex (1.5.0 RC ff). Maybe you remember (see here (http://forum.doom9.org/showthread.php?p=1133612#post1133612)). My tests with 1.5.2 show that the performance still suffers. I hope you can fix it with the 1.6.0 release.

Thanks,
Malcolm

Guest

30th October 2008, 01:27

- When -HIDE mode is invoked, GUI error messages shouldn't pop up. Instead, it would be more helpful if the error messages were hidden and an error level was returned, similar to x264.

morphinapg

30th October 2008, 03:26

When calling DGIndex from the command line, is it possible to output current percentage done and/or ETA through stdout/stderr? This would be very useful for ASXGui as right now it just shows a waiting screen with unknown how far along it is or how long it will take.

Guest

1st November 2008, 18:08

Improve AVS template management. Support multiple templates and a means to select from them.

check

8th November 2008, 13:43

A quick feature request: the .d2v file lists an absolute path for the source VOBs. If any VOB can't be found, could the current directory be searched for it too? It's a rare occurance, but sometimes I rename the folder the d2v/vob for a particular episode is in.

Guest

8th November 2008, 13:54

Can't just turn off the Full Paths option when you make your projects?

Chumbo

16th November 2008, 15:46

- When -HIDE mode is invoked, GUI error messages shouldn't pop up. Instead, it would be more helpful if the error messages were hidden and an error level was returned, similar to x264.
Would love to see this when -MINIMIZED is used too, i.e., turn off UI errors/warnings. May be an option that would log any error messages to a log file when in CLI. So an option to specify a log file would help too.

Improve AVS template management. Support multiple templates and a means to select from them.
Just last night I ran some stuff that used a template on the command line that was not set up as a default template in the DGIndex UI and no templates were created.

I figured the problem may be the location? If we don't specify a path for the template, does it look in the folder where the source is or where dgindex.exe is? Sorry for the stupid question, but I want to be sure. I'm sure it's the former. Thanks.

Guest

16th November 2008, 16:19

It's the latter.

Chumbo

16th November 2008, 16:32

It's the latter.
How about that? ;)

btw, I put the template in both locations and the CLI still did not create the avs file. I'm now setting the template file using the UI first and then running the command line.

Also, I ran it hidden, i.e., with the -HIDE and the "field detection" error/warning still popped up.

[EDIT] Update. Looks like the -AT command requires a fully-qualified path. Just putting the name of the template does not work regardless of where it is, i.e., in the source folder or the .exe folder. The UI setting has no impact on the CLI either.

Guest

16th November 2008, 17:18

It does work. If a full path is not specified, the current directory is used. Perhaps the current directory is not what you think it is.

Chumbo

16th November 2008, 17:38

It does work. If a full path is not specified, the current directory is used. Perhaps the current directory is not what you think it is.
Yep, you're right. It does use the source directory if a path is not specified. It was user error on my part, sorry about that. I have two drives I work with and had put the template on the wrong one thinking it was the source. Ugh...

Chumbo

16th November 2008, 21:42

@neuron2,
This may be a DOS thing, but thought I'd let you know about it in case it's not. I happen to have a file name with an ampersand in it and here's the error I received when my CLI batch file got to that file:
'Consequences.ts]' is not recognized as an internal or external command,
operable program or batch file.
The original file's name is: Heroes.S02E10.Truth & Consequences.ts
The & is only in the input file name on the command line like this:-IF=[Heroes.S02E10.Truth & Consequences.ts]

J_Darnley

17th November 2008, 00:16

-IF=[Heroes.S02E10.Truth ^& Consequences.ts]
It is a DOS thing, the ampersand will tell it to run another command. Escaping it with ^ should fix it.

Chumbo

17th November 2008, 04:59

-IF=[Heroes.S02E10.Truth ^& Consequences.ts]
It is a DOS thing, the ampersand will tell it to run another command. Escaping it with ^ should fix it.
Sweet, thanks for the tip. :)

josey_wells

19th November 2008, 15:11

It's very easy to turn on the deinterlacer with CUDA. I'll be experimenting with it to see how much it impacts the frame rate on the low-end and high-end cards, and how good the deinterlacing looks. All I know is that it is claimed to be "motion adaptive". You can Google for "PureVideo deinterlacing" to see marketing hype and reviews.

a version of DGMPGDec based on the NVIDIA Cuda would be fantastic. The difference in DGAVCDec and DGAVCDecNV was fantastic. The DGAVCDecNV gave much better framerates and perfect decoding.

A beta would be great to try out :)

Dare's User Name

19th November 2008, 22:34

Dare's User Name

19th November 2008, 22:37

It would be nice if DGIndex had a batch mode, so I could save project files for a large number of mpeg files that I know should be treated the same way, such as episodes of a TV series.

morphinapg

19th November 2008, 23:11

a version of DGMPGDec based on the NVIDIA Cuda would be fantastic. The difference in DGAVCDec and DGAVCDecNV was fantastic. The DGAVCDecNV gave much better framerates and perfect decoding.

A beta would be great to try out :)

this and would it be possible to detect whether a user has a capable card and use NVIDIA if available, otherwise just use the regular version?

Chumbo

20th November 2008, 02:25

I would like DGIndex to have an option to parse a d2v file in a way that lists ranges of frames with and without pulldown flags. The output might look something like this:

0,499 pulldown
500,999 clean
1000,15999 pulldown
16000,20000 clean
etc.

That would be easy to convert into an AVS script such as this, in order to treat pulldown and non-pulldown material separately:

a=MPEG2Source("filename.d2v").trim(0,499)
b=MPEG2Source("filename.d2v").trim(500,999)
c=MPEG2Source("filename.d2v").trim(1000,15999)
d=MPEG2Source("filename.d2v").trim(16000,20000)
Great idea. I second this one. :)
It would be nice if DGIndex had a batch mode, so I could save project files for a large number of mpeg files that I know should be treated the same way, such as episodes of a TV series.
This would be nice too, but no big deal since you can write your own batch file. Just use one as your template and then copy the rest. The CLI (Command-Line Interface) is not that complicated.

Inventive Software

21st November 2008, 19:06

@bug! Frame Size and Display Size are exactly the same for all PAL DVDs I put through DGIndex, I can't say the same for HD content because I don't have any. My assumption is that with the SAR applied, the Display Size would change from the Frame Size 720x576 to 1024x576 / 768x576 (for PAL DVDs).

Guest

21st November 2008, 19:08

Post stream showing the claimed bug.

Inventive Software

22nd November 2008, 02:40

/me goes to find DGSplit and an upload site you can use. Watch this space.

EDIT: Please (hopefully) find attached the information window.

I'm currently uploading a 50MB sample, from a decrypted VOB file from a DVD done by my standalone DVD recorder, to Rapidshare. If you prefer it done somewhere else, please speak now and I'll set it on its way.

Uploaded 50MB VOB sample: http://rapidshare.com/files/166142623/VTS_01_1_split.VOB

Dare's User Name

23rd November 2008, 19:45

This would be nice too, but no big deal since you can write your own batch file. Just use one as your template and then copy the rest. The CLI (Command-Line Interface) is not that complicated.

I understand that it is not that complicated in theory, but I'm retarded. I can barely get my head around Avisynth scripts as it is, and searching online for a working batch script takes me longer than if I just babysat the indexing process for a hundred files.

I think many lives, not just mine, would be a lot easier if we could specify a folder for DGIndex to demux and index, and let it do its thing.

J_Darnley

23rd November 2008, 20:27

Expand it with the options you need and the correct path of DGIndex.
cmd prompt: FOR %A IN (*.vob) DO dgindex -IF=[%~fA] -OF=[%~dpnA] -exit
batch file: FOR %%A IN (*.vob) DO dgindex -IF=[%%~fA] -OF=[%%~dpnA] -exit

Dare's User Name

24th November 2008, 06:44

Sorry, you lost me at "Expand". I am merely a mortal! :)

You are very skilled at command line incantations, but like it or not, this program does have a GUI, and I see no fault in wanting to expand its usefulness.

On a buggish note, I got a nasty error on one file, "Too many pictures per GOP (>=500). DGIndex will terminate." Click OK, and it quits. There doesn't seem to be anything wrong with the video, and it plays just fine. If I create an avisynth script to open it using DirectShowSource, it opens in VirtualDub without any problems. I can create and post a log if that would be useful.

J_Darnley

24th November 2008, 13:06

I mean add in the options you need such as audio demuxing, audio decoding, audio track(s), avs template. Any other options you need beyond the basic structure I provided with just the input and the output. If you want to try it, provide specifics on the names of the input files and where you want the output and I will adapt it for you. If you want to continue slogging through your files with the GUI, you are welcome to do so. You can also post requests.

As for your problem, you will have to wait and see what neuron2 says.

Inventive Software

24th November 2008, 13:50

OCedHrt

12th December 2008, 19:01

Add mouse scrollwheel support and Copy Frame to Clipboard, as in DGAVCDec.

Add GPU decoding support, e.g., CUDA Video API.

Is this limited to CUDA cards (aka nVidia) or will it also work for ATI cards?

laserfan

5th February 2009, 20:28

Add GPU decoding support, e.g., CUDA Video API.In your own forum, you said "H264 is supported in DGAVCDec. I don't intend to merge DGMPGDec and DGAVCDec. At some point all three will be supported in one application in my Nvidia enabled application."

Presumably by "all three" you meant MPEG2, H264, and VC-1. Now that I am Nvidia-enabled, I want to *bump* the idea of a DGMPGDecNV. But merging these three into one application would result in almost more joy than I could bear; indeed t'would be a consummation devoutly to be wish'd. ;)

smok3

6th February 2009, 08:32

a feature request, for command line, from

'0=None, 1=Demux Tracks, 2=Demux All Tracks, 3=Decode AC3 to WAV '

to

'0=None, 1=Demux Tracks, 2=Demux All Tracks, 3=Decode ANYTHING I CAN to WAV' (including mp2 for example)

p.s. and a question: are the ac3s supposed to be fixed regarding the delay when they are decoded?

kypec

6th February 2009, 10:23

p.s. and a question: are the ac3s supposed to be fixed regarding the delay when they are decoded?
I think not, only the filename reflects DELAY if detected. I ceased to use DGDecode for AC3->WAV decoding, eac3to (http://forum.doom9.org/showthread.php?t=125966) is much more suitable for this job, it handles conversion and fixes delays automatically! :)

Guest

6th February 2009, 15:13

p.s. and a question: are the ac3s supposed to be fixed regarding the delay when they are decoded? Yes, the previous poster is incorrect.

kypec

9th February 2009, 06:17

Oops, I'm sorry for providing wrong info. Has this feature been enhanced (auto-adjustment of AC3 delay) when compared to v1.5.3?

Guest

9th February 2009, 06:33

Maybe you are confusing demuxing with decoding?

From the users manual:

Note that when AC3 tracks are decoded to WAV, their filenames do not contain a delay correction value because the correction is applied during decoding.

smok3

9th February 2009, 14:50

neuron2, tnx, how about this:

'0=None, 1=Demux Tracks, 2=Demux All Tracks, 3=Decode ac3 to wav, if it is ac3, else just demux'

Zarxrax

26th April 2009, 18:53

I'm not sure when it changed, but I think the current "select tracks" dialog is much less user friendly than the old one. I frequently decode audio streams to wav, so it's annoying to have to look up the IDs for the stream I want and then type it in. It would be fantastic if you could just show a list of all audio streams and then let the user put a checkbox beside the ones they want to use.

Guest

5th June 2009, 16:01

hajj_3

8th June 2009, 11:34

neuron2, tnx, how about this:

'0=None, 1=Demux Tracks, 2=Demux All Tracks, 3=Decode ac3 to wav, if it is ac3, else just demux'

i'd add a number 4 which encodes to 2 channel 128bit mp3 using lame.exe, would be pretty handy!

7oby

15th June 2009, 18:03

I have been using DGMPDec to transcode some MPEG2 movies to MPEG4-AVC with x264 for mobile use. I recongized that the MPEG2 decoding part takes a significant part of the available CPU cycles.

I was wondering if there's a way to speed up the decoding process by incorporating the GPU to some extend. DGMPGDecNV uses nVidia CUDA, but unfortuantely I have a notebook with intel integrated graphics (GM965/X3100). It exposes its MPEG2 decoding facility only by means of DXVA. From reading neuron2s dialog with nVidia (click (http://neuron2.net/dgavcdecnv/cuda/cuda.html)) and some postings in the CoreAVC release thread it seems to me the DXVA (regardless of DXVA 1.0 or 2.0) doesn't easily allow to decode already parsed streams and redirect the output to non GPU memory or files. Is that correct?

Intel documentation for the MPEG-2 decoding is available (http://intellinuxgraphics.org/documentation.html), but actual code libraries or code snipplets only for linux (click (http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=52054b6a4c1ca5117c9750361f71aedd91220c39)). So I guess access to the an intel MPEG-2 decoder is not easily possible? Actually windows code snipplets for the H.264 intel decoder are easier to find (click (http://software.intel.com/en-us/blogs/2009/03/15/intel-clear-video-and-h264avc/)).

Anything else how I could speed up the MPEG-2 decoding process? Is it possible to do at least the video scaling 720 x 576 -> 320 x 240 or 480 x 272 on the GPU? Are there tools available? EDIT: Meanwhile I tested the japanese _GPU2.5 rev53 (http://www.avisynth.info/?%A5%A2%A1%BC%A5%AB%A5%A4%A5%D6#x44606bb) plugin which still sees me be maintained since there's also a download for ATI HD48x0 chipsets. The result on intel grraphics using GPU_BilinearResize() was slower than CPU based resize. It seems the penalty of reading back the resized data from gpu allocated memory is too expensive in this case compared to the cpu cycle win.

Thanx for comments.

P.S.: DGMPGDecNV and DGAVCDecNV are both great tools - regardless of whether I am able to use it or not. I also like the idea that you have to pay for it.

B@Ron

3rd July 2009, 09:09

Something like add another segment and another and another in the making of one .d2v file so it would be possible to cut out the commercials from MPEG .ts file directly in DGIndex.
Is it possible to make something like this? (Sorry if it had been asked or answered before:))

Thank you neuron2 for this truly great program!

halsboss

5th July 2009, 01:59

Our PAL 1080i HDTV channels are generally rubbish and are really really blocky. "Pretend" HDTV, really. I'm using
#ipp=true interlaced post-processing
#cpu=4 DEBLOCK_Y_H, DEBLOCK_Y_V, DEBLOCK_C_H, DEBLOCK_C_V
#cpu=6 DEBLOCK_Y_H, DEBLOCK_Y_V, DEBLOCK_C_H, DEBLOCK_C_V, DERING_Y, DERING_C
MPEG2Source("G:\HDTV\q1.d2v",info=0,ipp=true,cpu=4)

Would it be possible to add functionality along the lines REALLY_STRONG_DEBLOCK=true ?

I have an unfounded feeling it'd be better done in here than in avisynth using deblock_qed_mt trying to deal with fields with fast action interlaced material.

MrMod

24th August 2009, 01:30

I've been puzzled by occasional "A field order transition was detected" with some of my DVDs. I have since found the culprit which leads to an issue currently in DGIndex.

I first noticed something fishy when I was using mkvtoolnix to mux some vobs into a MKV. There was an odd Padding stream packet with a very large packet length specified after the header. This caused mkvtoolnix to warn that audio/video sync was lost and may be noticable during playback. As it turns out, I find this occasionally on DVDs. It appears the authoring sometimes puts these weird Padding stream packet lengths. The good news is that if you detect the >2048 length and search for the next packet_start_code_prefix (000001), you will find that the PTS for the audio and video remain correct.

I confirmed this with a tool of my own and indeed, if you detect the Padding stream packet length >2048 and just search for the next 000001, instead of skipping the bytes indicated by the bogus packet length, then the subsequent PTS for audio and video lines up with what is expected.

So I wondered if this was the cause of these strange "A field order transition was detected" messages. I took a VOB with a known bad Padding stream packet length, ran DGIndex, and sure enough it notified me of the field order transition. I then hex edited the VOB, corrected the Padding stream packet length, and re-ran DGIndex. This time it found the VOB was good. You can also tell something is up in the fixed d2v file since the lines in question are unusually long.

So for now I've been hex editing the bogus Padding stream packet lengths before using DGIndex, but thought I'd report my findings to make the awesome DGMPGDec tool even better.

Guest

24th August 2009, 01:40

Interesting. I would appreciate you posting a link to a VOB fragment that I can use to see this issue in operation. You can use DGSplit to split VOBs.

MrMod

24th August 2009, 02:33

Interesting. I would appreciate you posting a link to a VOB fragment that I can use to see this issue in operation. You can use DGSplit to split VOBs.

Here's a hex dump of an example (clip.txt):
108064000 PACK header
10806400E PES packet (video stream 0, length 0068)
10806407C PES packet (padding stream, length 6A71 <-yikes!)
108064800 PACK header
10806480E PES packet (video stream 0, length 07EC)

So if you respect the padding stream length you end up skipping the next valid PES packets. So in this example the padding stream length should be changed to 077E so that the next PACK header located at 108064800 is used. This is what I do manually to fix these bad padding stream packets when I run across them.

Another interesting thing to note is that padding stream packets are supposed to be stuffed with FF bytes, but in this case you can see it appears to be filled with garbage. A sanity check you can do is to make sure the next PACK header you find is on a page boundary, that way if the garbage bytes happen to contain the packet_start_code_prefix of 000001, you can ignore it and keep searching.

Guest

24th August 2009, 02:42

I'm not doubting your analysis.

I need the stream to test a fix!

Varies

26th August 2009, 14:57

If I understand correctly, credits and other bad lookin scenes appear because frames marked as progressive and not processed by deint, i'm right?
you may write functions for manual select interlaced frames ?

Guest

26th August 2009, 15:01

There is no deinterlacing functionality in DGMPGDec, so I can't make any sense out of your posting. And DGDecode delivers decoded frames, so there's no way to mark them.

Varies

26th August 2009, 15:23

I know, but unless DGMPGDec does not write in d2v the information on fields which further is used in deint?

Guest

26th August 2009, 15:30

Even if I revised the progessive_frame indication in the D2V file, it would be lost in DGDecode and could not reach a deinterlacer.

Varies

26th August 2009, 16:53

I think so... Source->DGIndex->d2v->DGDecode->video+flags (i,p) for deint ->Deint or IVTC->Progressive Video... i'm not right? :)

Guest

26th August 2009, 21:03

Ah, you're talking about the hinting data in the video.

I find it hard to believe that the pf flag is wrong on a per frame basis. Show me a stream where it is so, that requires such a manual capability.

Varies

27th August 2009, 10:24

As sample 30p with 60i subtitles, the first sample source that I sent you... it will help make deint better?

Guest

27th August 2009, 13:22

As sample 30p with 60i subtitles, the first sample source that I sent you... it will help make deint better? You can use a deinterlacer with manual overrides to force deinterlacing on a per-frame basis. I'm not going to hack up a decoder to do that.

vlada

1st September 2009, 09:29

Hi, neuron2. Is there any option to turn on progress reporting in stdout?

I'm using DGIndex from CLI in my application. To make the look of my program consistent, I'd like to hide the DGIndex's window and report the progress directly in my program's GUI. I use following command line:

dgindex.exe -IF=[video.vob] -OM=0 -OF=[output] -HIDE -EXIT

If I try to read from stdout, there is no output at all. Is there any CLI switch which would turn on progress reporting to stdout? If not would it be possible to add such an option?

Another option for me (not preferred) would be to show the DGIndex main window, which displays the progress. But the users may close this window (cancel indexing) and it will cause serious issues. First problem is, that the returncode of DGIndex is 0 even if the indexing is canceled, so I have no information that the user aborted the process. Then if I use the partial .d2v file in AviSynth, it will cause the host application (VirtualDub, avs2yuv) to hang and use all memory of my PC.

Could you add a feature to better handle this situation? One possibility is to delete incomplete .d2v files or (if possible) check whether the file is complete and if not, report it as an error through AviSynth.

If I'm unclear in some points, please let me know. Thank you.

Guest

1st September 2009, 12:55

Source code is available. Feel free to submit revisions.

vlada

3rd September 2009, 06:35

I'm sorry, but unfortunately I have absolutely no knowledge and experience with C/C++ programming. I can only hack some GUI in PyGTK or PHP-GTK. I'll try to look into it, but I'm afraid I won't be able to find out how to do it. :-(

morphinapg

18th November 2009, 03:39

How about Windows Media Center format support (dvr-ms/wtv) obviously non-DRM files only.

Currently I have to use DVRMSToolbox and then DGIndex.

Guest

18th November 2009, 04:22

Isn't DVR-MS obsolete now?

WTV might be a possibility if a spec is available.

morphinapg

18th November 2009, 04:36

Isn't DVR-MS obsolete now?

WTV might be a possibility if a spec is available.

idk, I have to convert to DVR-MS before I can use DVRMSToolbox with it. Windows 7 has an option when I right click the wtv file to convert to DVR-MS. But yes, WTV would be much more useful to me, I just thought some people might still use dvr-ms.

SomeJoe

24th November 2009, 09:14

If I try to read from stdout, there is no output at all. Is there any CLI switch which would turn on progress reporting to stdout? If not would it be possible to add such an option?

It has been a while since you posted, but I handle this as follows:

In my front-end application that calls DGIndex/DGIndexNV, I periodically (every 3 seconds or so) peek into the .d2v/.dga/.dgm/.dgv file that is being created and grab the last several lines of the file, and then extract a frame number from the data within. This will be very near the current frame that DGIndex is working on.

My front-end app has knowledge of the total number of frames in the file (as reported from other tools, such as eac3to), so this allows me to derive a progress %.

vlada

24th November 2009, 11:59

SomeJoe> Thanks a lot for you advice. But last week I asked a friend of mine if he would be able to implement the CLI progress reporting to DGIndex and he did it. Yesterday he sent a modified .exe which prints percentage numbers to stdout and this evening he will send me the modified sources.

Guest

24th November 2009, 16:05

Please submit the sources to me so I can absorb it into the mainline. Thank you.

vlada

24th November 2009, 23:23

The modified sources are here (http://www.mediafire.com/?zmmznmzn3m3). It should be modification of version 1.5.6. I hope it can be merged into the mainline easily.

d62ks821

30th December 2009, 18:16

Great tool. One suggestion, dgindex crops bottom 8 of 1088 HDTV streams to get to 1080, however, all my local broadcast station place the extra 8 lines differently. some the extra 8 lines are at top rather than bottom. So I'd like to see command line options to either (1) pass the 1088 through so I can crop as needed with avisynth/vdub, OR (2) allow selection of which 8 lines to crop. Thanks!

stax76

5th January 2010, 15:18

I have a VOB file where MediaInfo and DGIndex report 70 seconds delay but the actual delay is 14 seconds. I'm pretty clueless about the structure of VOB files otherwise I would provide more info, I can however upload a sample. People using DGIndex directly would just use the DGIndex trim feature and it would be fine but this is not a option trying to automate things, use subtitles etc.

Guest

5th January 2010, 18:40

Post the VOB sample please.

A delay of 14 seconds is prima facie absurd.

stax76

5th January 2010, 19:03

Are 50 or 100 MB enough? Problem is they start talking at 350 MB and I have only 576 kbit/s upstream so uploading 400 MB would take a while.

Guest

5th January 2010, 19:07

It's not enough if it doesn't include an AV sync event that can be used to determine the actual sync, which might be different from the sync determined from the timestamps.

Is this a commercial DVD?

stax76

5th January 2010, 19:48

Yes it's a commercial DVD, is that a problem?

Guest

5th January 2010, 20:17

Not a problem. It just sounds highly unlikely for a commercial DVD.

Go ahead and upload as much of it as you can, with the beginning intact.

stax76

5th January 2010, 22:11

My hoster isn't very fast (but free :) ) so you will need to have some patience, should download at 100-300 kB's.

http://www.stax76.bplaced.net/files/VTS_01_0.VOB

Guest

6th January 2010, 16:06

Is the site down? I have a dead link.

SeeMoreDigital

6th January 2010, 16:11

My hoster isn't very fast (but free :) ) so you will need to have some patience, should download at 100-300 kB's.

http://www.stax76.bplaced.net/files/VTS_01_0.VOBAs a matter of interest. How big is the sample?

stax76

6th January 2010, 16:23

It's 380 MB, unfortunately the site is down atm., where else could I upload it?

MuLTiTaSK

6th January 2010, 16:51

@stax76

try here (http://tinyurl.com/2fzusz)

stax76

6th January 2010, 18:09

The site runs again, it's downloading at 160 kB/s here. I'm currently uploading it to file dropper (thanks MuLTiTaSK) as mirror and I'll also upload it to another hoster as second mirror but this time I've split it into two parts using 7-Zip in case the upload aborts.

stax76

6th January 2010, 21:37

mirror 2

http://www.filedropper.com/vts0107z
http://www.filedropper.com/vts0107z_1

mirror 3

http://rapidshare.com/files/331338784/VTS_01_0.7z.001.html
http://rapidshare.com/files/331364281/VTS_01_0.7z.002.html

Guest

6th January 2010, 21:46

I have the stream now.

It's clear what is happening. The timestamps reset to 0 after the leader. Normally we just start the project in from the start skipping the leader but it requires a manual step. I will look for a possible automatic handling of it.

stax76

6th January 2010, 23:32

It reported 70 sec delay while it was 14, the subtitles have a delay of -56 sec so even if you manage to put it in sync I still need to understand what's happening with the subtitles, what's worse is VSRip isn't maintained. To understand what's going on and how to deal with it I'll have to learn a lot more about the DVD structure and this don't seem to be easy.

ProjectX seem to be able to sync video, audio and subtitles perfectly, I tried it just with the first VOB but it can't give the subtitle streams a language ID since this info is in the IFO file and it just work on VOB files.