View Full Version : DGMPGDec 1.4.7 Final
shpitz
21st April 2006, 13:24
neuron2,
great job on the program, but i got the following issues:
1)
i have a recording into a .TS file, the show itself is 5.1, but before the show starts it is only 2.0 .
when decoding the output file will say it's 2_0 while it is actually 5.1 (verified with GSpot).
can you tell the program to set the output file to say 2_0 by default if it is ac3, but change it to 3_2 if dg detects a 5.1 audio during the decoding?
2)
can you not close the information dialog when i do a manual seek using the bar at the bottom? it would be nice to see the time-stamp for example during seek to have an idea when a show is over...
thanks
bob0r
24th April 2006, 16:11
Searched this thread for h264 but nothing found, so here is my question:
Will you support h264 .ts files any time in the future?
Related post:
http://forum.doom9.org/showthread.php?t=110312
Sample file:
http://forum.doom9.org/showthread.php?t=110390 (mirror (http://files.x264.nl/hd/prosieben_hd_20232012022006.ts))
john3voltas
29th April 2006, 23:30
I didn't read the whole thread but I found something in 1.4.7b6 and I thought I would share it with you guys.
I created a project of my Batman Begins PAL from a MainMovieOnly rip with DVDDecrypter.
In the end I used ParseD2V.
Here's the bottom lines:
[GOP]
201429 [B]: 201429,201429.......2
201430 [B]: 201430,201430.......2
201431 [I]: 201431,201431.......2
201432 [P]: 201432,201432.......2
[EOF]
Running time = 2 hours, 134 minutes, 17 seconds
Shouldn't it read only 134 minutes and 17 seconds?
But it reads 2 hours, 134 minutes and 17 seconds.
Strange, I'd say it's a bug...
Am I right?
I would be glad to hear your comments.
TIA
Guest
30th April 2006, 02:26
Strange, I'd say it's a bug... You'd be right. I'll fix it for the next beta. Thank you for pointing it out.
john3voltas
30th April 2006, 02:38
Nah, don't mention it ;)
AVIL
30th April 2006, 21:57
Hi,
With this last beta, delay for audio isn't computed well. Filename says 0 ms and real delay was -288 ms (audio was advanced over video).
It occurs on DVD Chicken Little, after stripping multiangle scenes (over the original footage, dgindex report all the angles sequentially, and I could not remux reencoded m2v file)
Thanks.
There has been no change to the audio delay calculation code for a long, long time.
It's never going to be perfect. For one thing, streams sometimes have discontinuous timestamps, and that really messes things up. Also, the delay is calculated only at the very beginning of the stream, so if there is some anomaly at the start, as there often is, again, things can get messed up. I use the delay value as a starting value to try. Sometimes it is off and I have to tweak it manually.
When the MPEG stream is played directly, the timestamps are used to directly sync the audio and video. So the sync offset can and often does change during a broadcast, especially at content cuts (e.g., show to commercial and back, one program to another, etc.). But when you encode to a format like AVI that doesn't have timestamps you have to guess one offset to use for the entire stream.
Ideas for improving the design are always welcome. Yes, I already know about the idea of inserting/deleting audio to make a stream with a constant (selectable) sync offset. One thing that I will try for the next beta is to have a display of the current sync offset in the information dialog box. I also want to expose all of DGIndex's analysis about the stream via the CLI.
@neuron2
In this case I've found by trial and error a value of -300. So, when ReJig V0.5e (you'll know, a mpeg requantizer from Nic) gives me a value of -288 I think that, perhaps, knowing the algorithm from Nic's app could help.
shpitz
1st May 2006, 13:55
neuron2,
can you please explain how an audio can be out of sync? how can that happen? i just don't get it...
i mean, the audio plays along with the video on the same time-slot on the same frame, i just don't understand how they can get off. if we have a 30min long video file, how come each video frame does not correspond to the same frame in the audio? i.e frame 30 in the video sequence plays along with frame 30 of the audio track, this is something i fail to understand.
the really thing that puzzles me about audio sync issues is the fact the original file capture plays flawless (or so perceived) in a media player yet it goes totally out of sync when demuxed and encoded...
would you please elaborate on that?
thanks
i mean, the audio plays along with the video on the same time-slot on the same frame, No. There are no time slots, the video and audio access units are completely different and do not necessarily correspond one-to-one, and neither one carries access unit numbering (except for temporal order, which is useful only at the GOP level).
i just don't understand how they can get off. if we have a 30min long video file, how come each video frame does not correspond to the same frame in the audio? i.e frame 30 in the video sequence plays along with frame 30 of the audio track, this is something i fail to understand. See above.
the really thing that puzzles me about audio sync issues is the fact the original file capture plays flawless (or so perceived) in a media player yet it goes totally out of sync when demuxed and encoded... That usually happens only in pathological situations, for example, your stream mixes audio at different rates (commercials sometimes have different audio rates than the main program). As I already mentioned, if the player plays the MPEG directly, it can present the audio and video correctly by using the PTS timestamps. These timestamps do not exist in AVI and so their information is lost.
In this case I've found by trial and error a value of -300. So, when ReJig V0.5e (you'll know, a mpeg requantizer from Nic) gives me a value of -288 I think that, perhaps, knowing the algorithm from Nic's app could help. Yes, of course. Is the source code available?
* Improved the misleading error message a user gets if he erroneously loads a file with MPEG2Source() that is not a D2V project file. Previously it complained about a version mismatch; now it says the file is not a project file.
* Fixed the display of the running time in the Parse D2V output.
http://neuron2.net/dgmpgdec/dgmpgdec147b7.zip
Will you support h264 .ts files any time in the future? It's highly doubtful. I think a new generation encoder needs a new generation application. :)
i have a recording into a .TS file, the show itself is 5.1, but before the show starts it is only 2.0. Currently, you have to use multiple projects using range selection to handle this situation. I don't have any plans to improve that right now.
can you not close the information dialog when i do a manual seek using the bar at the bottom? it would be nice to see the time-stamp for example during seek to have an idea when a show is over... Sorry, no. Some of the fields are valid only when a preview or save project operation is performed, so to avoid misleading people I dismiss the dialog box. Yes, I could improve that but the priority of this issue makes it unlikely to be addressed. The source code is there and if you care to develop the changes, I'll be happy to accept them.
I have a feature request - when a .ts file is opened by dragging it over the window and dropping, the File List requester is shown. However, when using Open With in the context menu, DGIndex is launched but fails to open the file. Apparently windows calls the app with "appname.exe" "%1" where %1 is the filename. I'll try to fix this when I overhaul the CLI syntax and functionality.
shpitz
4th May 2006, 13:31
Sorry, no. Some of the fields are valid only when a preview or save project operation is performed, so to avoid misleading people I dismiss the dialog box. Yes, I could improve that but the priority of this issue makes it unlikely to be addressed. The source code is there and if you care to develop the changes, I'll be happy to accept them.[/QUOTE]
thanks for your replies.
can you at least put the time-stamp in the window title so i'll have an idea where i am in the file? no need for any other info, just the time elapsed.
thanks
SeeMoreDigital
4th May 2006, 19:43
Hi Donald.... Did you manage to make any headway with DTS audio streams in .TS?
Cheers
Yes, of course. Is the source code available?
It's here:
http://nic.dnsalias.com/ReJig_src.zip
but after more test I'll see that their algorithm to find the audio delay isn't perfect. Thanks anyway.
I can only second SeemoreDigital here, DTS support for .ts would be highly appreciated! :-)
Thanks
OK, when I finish the timestamp dump I am working on I'll do DTS in TS.
Ok I'm quite clueless atm and most likely this is some stupid user error or glitch in my setup but I cant figure out whats happening and I have used DGindex for a long time without any issues. For some reason 147b7 freezes my computer when I start demuxing .ts stream. Starts demuxing but slows down fast with only 2sec done and my comp freezes. Now when I preview the stream with F5 everything runs fine. I thought first that my comp. had got unstable but I cant get any errors while running Pifast,Prime and so on. Tested my HDDs also if one had fallen to PIO but they are all fine. Tried both IEE-1180 and Skal. My CPU is AMD64 singlecore.
Guest
10th May 2006, 22:50
What do you mean by demuxing? Specifically, were you doing Save Project or Save Project and Demux Video? Were you demuxing audio?
Try disabling audio processing and then do Save Project. Does that cause a problem?
If it happens in a couple seconds, you ought to be able to provide a fragment of the TS that I can use to try to duplicate your issue.
Sorry, saving .d2v and demuxing ac3 from .ts
Most likely the porblem is something else than dgindex although I cant lock my computer with anything else atm. And here is the strange part now it works! I did as you suggested, saved .d2v without audio and it worked. Tried saving .d2v and demuxing audio
and lockup although it went along around 50sec instead of the usual 2sec on earlier tries. Did more stress tests with StressPrime2004 and nothing. After this tried again dgindex and now it worked. I'm just going to put this issue on some Myrphy's Law or something, naah I blame M$ :rolleyes:
Guest
11th May 2006, 19:46
My guess is that you had a wrong PID set for the audio when it failed.
Guest
15th May 2006, 01:20
* Added an option to dump the audio and video PTS/DTS timestamps encountered during a Save Project operation.
http://neuron2.net/dgmpgdec/dgmpgdec147b8.zip
This feature will be useful for power users as well as for giving me needed debug traces in AV desync situations. I plan to enhance this later to parse the dump and report discontinuous timestamps and other pathologies.
The next beta will add demuxing DTS audio from transport streams. Then I'll make a final release of 1.4.7.
Wilbert
15th May 2006, 10:22
The next beta will add demuxing DTS audio from transport streams. Then I'll make a final release of 1.4.7.
Great!
* Added an option to dump the audio and video PTS/DTS timestamps encountered during a Save Project operation.
I guess it should be possible to convert those to MKV timestamps (at least if some MKV guru is will to invest some time for that)?
MichalHabart
16th May 2006, 15:01
I am looking forward to demux DTS audio from transport streams. Hopefully new version of GDIndex will be released soon :)
:thanks:
Guest
16th May 2006, 15:24
I am looking forward to demux DTS audio from transport streams. Hopefully new version of GDIndex will be released soon The more test streams I have the better. Can you please provide one?
Guest
17th May 2006, 04:29
This beta adds the capability to demux DTS audio from transport streams. Thanks to SeeMoreDigital for providing test streams, and to gabest (Media Player Classic developer) for advice on detecting DTS streams.
http://neuron2.net/dgmpgdec/dgmpgdec147b9.zip
Note to SeeMoreDigital: The second stream you provided has a PAT/PMT table that shows a DTS audio on PID 0x25. But that PID does not appear in the stream, as you can see by doing Raw PID Detection. I don't know why the PAT/PMT table includes that entry. It's always advisable to do an F5 preview to make sure the audio is actually found on a PID before doing the entire Save Project operation. In your stream, the DTS audio is on PID 0x24.
Isochroma
17th May 2006, 05:32
Your work is invaluable and this, just the latest of many improvements to make DGMPEGDec more functional, has helped so many people get the most from their transport streams. So I'd like to thank you for your extensive and ongoing labours. Someone richer than I ought to pay you for this valuable work.
I use the tool constantly to index TS files, on a daily basis. DTS decoding will be an invaluable addition to the mix. From my extensive TS collection, I'd say that about ten to fifteen percent of such files have a DTS track.
A question (maybe asked before): is it feasible to serve the decoded audio stream via the dgmpegdec plugin and avs file to an application?
SeeMoreDigital
17th May 2006, 10:12
Many thanks Donald,
I think the sample I first gave you is the one I ran thru' MPEG2 Repair in order to give the DTS stream priority over the two AC3 streams.
The source was originally sent to me by somebody who wanted to know whether the Zensonic Z500 (I was beta testing at the time) could play "progressive" 1920x1080/88 MPEG-2 video running at 25fps complete with DTS audio!
Needless to say the player could ;)
Now, after all Donald's hard work it would be cool to find a "muxer" that can re-mux the elementary 1920x1080/88 MPEG-2 video and DTS audio streams back together. Imago MPEG-Muxer and MPEG Multiplexer are having none of it :scared:
Cheers
Guest
17th May 2006, 14:29
Thank you for the kind words.
A question (maybe asked before): is it feasible to serve the decoded audio stream via the dgmpegdec plugin and avs file to an application? You can demux the DTS audio and then decode it with NicAudio.dll and serve it in the standard way, as described in the Quick Start guide.
loadplugin("dgdecode.dll")
loadplugin("nicaudio.dll")
vid=MPEG2Source("00 - Fifth_Element_1080p_ Source.d2v")
aud=NicDTSSource("00 - Fifth_Element_1080p_ Source PID 024 DELAY -845ms.dts",2)
audiodub(vid,aud)
laserfan
17th May 2006, 15:35
Your work is invaluable and this, just the latest of many improvements to make DGMPEGDec more functional, has helped so many people get the most from their transport streams. So I'd like to thank you...I must add my admiration and thanks as well. I'm one of those guys who, though not a "newbie", has been stumbling around using one-click tools like AGK and FUW and getting decent-but-variable results, all the while being mostly clueless as to what goes on in the background w/these programs.
Recently I started using StaxRip cuz I needed more codec flexibility, and it comes built defaulting to ProjectX to demux my TS files. Well, I come to find out that DGindex can do this all by itself and without the attendant errors/problems PX has with ATSC streams. Further, it only demuxes the audio saving lots of time & hdd space...
I ramble, but I must say I am sleeping with the DGindex manual under my pillow these days, and in general I have learned more from the neuron2 site than any other single source. I even, thanks to Mr. Graft, know enough now to "Inspect for Motion" any new program, and in most cases I even know what to do about it! A breakthrough folks! I'm finally learning this stuff!!! :D
:thanks:
Isochroma
17th May 2006, 16:27
Thanks for the info, neuron2; I wouldn't actually use this method, but was just wondering...
Guest
20th May 2006, 18:59
Good news for Fusion card owners!
Fixed a problem that prevented DGIndex from processing
audio streams from private stream 1 when the stream is not
DVD-structured (i.e., when there is no audio substream
number and associated DVD audio header). For example, AC3 audio
could not be processed from MPG files captured by a
"FusionHDTV RT Gold 5" card.
http://neuron2.net/dgmpgdec/dgmpgdec147b10.zip
My previous analysis was all wet. The 0x0b was the start of the AC3 audio sync word!
http://forum.doom9.org/showthread.php?t=101507
Yeah, great work neuron2!
Guest
21st May 2006, 02:18
Oops, beta 10 had a bug in Decode to WAV. Please ditch it and get beta 11:
http://neuron2.net/dgmpgdec/dgmpgdec147b11.zip
Carpo
21st May 2006, 11:40
glad i checked here before is started to do my next backup :p
buzzqw
21st May 2006, 13:50
pardon me Carpo...
but why decode audio to wav ? there is plenty of possibility for trancoding ac3 (for example) to mp3/ogg/aac...
in any case for anyone using this option i am glad that is fixed
BHH
Guest
21st May 2006, 14:14
in any case for anyone using this option i am glad that is fixed It was broken in one beta that was available only for a few hours. Such are the perils of betas.
buzzqw
21st May 2006, 15:51
afaik i don't remember any serius bugs within Dgindex not fixed in few hours :thanks: again
BHH
Carpo
21st May 2006, 16:15
pardon me Carpo...
but why decode audio to wav ? there is plenty of possibility for trancoding ac3 (for example) to mp3/ogg/aac...
in any case for anyone using this option i am glad that is fixed
BHH
neroaac encoder only supports wav atm - and im testing it so decode to wav helps :) - i could use bepipe but i dont wanna get ahead of myself, learn one thing at a time :)
buzzqw
21st May 2006, 16:58
Carpo try something like this
create a file named audio.avs
LoadPlugin("your_plugin_path\NicAudio.dll")
NicAC3Source("your_audio_path\audio.ac3")
#for multichannel, NicAC3Source("your_audio_path\audio.ac3",2) for quick stereo setup
EnsureVBRMP3Sync()
ConvertAudioToFloat()
then create an audio.bat
bepipe.exe --script "import(^audio.avs^)" | neroaacenc.exe -q 0.3 -if - -of audio.mp4
launch audio.bat
BHH
Carpo
21st May 2006, 17:40
thanks i'll give that a try on my next backup :)
backing up dvds was soooo much easier ;D
"import(^audio.avs^)" are the ^^ required or do they need to be removed ?
goes to show i should look the rest of the forum over as
Simple NeroAACEnc GUI - SNG v0.7.1 - will do most forms of audio to neroaac ;P (although it is faster if you use wav and use megui to do it - from my perspective anyway)
Guest
21st May 2006, 18:37
Please take OT discussion to appropriate forum or PM. Thank you.
cc979
22nd May 2006, 15:08
think i've found a bug with dgindex 1.4.7 beta 11
using megui with it
http://forum.doom9.org/showthread.php?p=830585#post830585
Guest
22nd May 2006, 16:14
Yeah, I found that the new detection code to support Fusion is not quite ready for prime time. Back off to beta 9 for now. I'm working on making the Fusion support robust.
Guest
23rd May 2006, 00:53
Technical difficulties prevented me (at this time) from implementing reliable automatic detection of Fusion-style audio, so I implemented a GUI option to force it. It is described in the DGIndex User Manual. It would be better if it could be made automatic, so I will continue to try to find a reliable way, but for now this should at least allow Fusion owners to decode their MPG streams. (For geeks: One gotcha is that LPCM streams do not have a sync word and so it difficult to distinguish LPCM from a stream corruption that clobbers the audio substream number. That's why beta 11 could erroneously think it had PCM audio.)
* Added an option to force proper parsing of audio when the stream is not DVD-structured (i.e., when there is no audio substream number and associated DVD header). For example, AC3 audio could not be processed from MPG files captured by a FusionHDTV RT Gold 5 card.
http://neuron2.net/dgmpgdec/dgmpgdec147b12.zip
Sharktooth
24th May 2006, 02:00
today i added b12 to megui autoupdate.
if there are still problems ill report them here.
Guest
24th May 2006, 02:16
Thank you, Sharktooth. I'd like to final release this version, so your absorption of it will provide useful testing.
BTW, version 1.4.8 will revamp the CLI and make all parsing information available to 3rd-party apps.
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.