DGMPGDec 1.5.3 Final

[kumi]

Neuron2: thanks again! :D

[drmpeg]

this sample has problem

the "info" report only 1 audio track but pid detection reveal 2 audio
and dgindex is able to extract the second track

http://www.64k.it/andres/data/a/film_Ac3_5_2.ts (37mb)

is it a stream problem ?

BHH
The stream is a little weird in that the PMT has three audio entries. The third entry has a descriptor for English language, but the PID is the same as audio stream 2 (in German).

The video stream sets progressive_frame correctly. 1 during the movie and 0 during interlaced portions of the commercials.

I also liked the commercial at the end of the clip. ;)

Ron

[Guest]

We don't get commercials like that here in the US. :(

[Sharc]

http://neuron2.net/dgmpgdec/dgmpgdec150b12.zip[/url]

I'll wait and see who cries foul about the second bullet item. I'd prefer not to make an option for it.

Thank you neuron 2.
Yes, let's wait for users' experience with the new defaults.

[Underground78]

Hi,

I sometimes have some ts files with strange audio tracks : for example in this sample (http://michel93.free.fr/data.mpg), it seems to have this tracks :

http://pix.nofrag.com/c/7/1/c9b73dcf36bbd901745084c0be94f.jpg

but the tracks number 150 and 151 can't be demuxed (the file is not created) ... However track n°151 is the default one so it matters when the CLI is used ...

Is there a way to know which audio tracks really exist and to avoid an invalid track being selected as the default one ?

[SeeMoreDigital]

Quite a while ago somebody sent me a std-def .TS sample complete with "TeleText" data for testing with my Ziova CS505 hardware player.

I've just run it thru' DGIndex Beta 12 and a whole host of PID's appeared: -

http://img136.imageshack.us/img136/6219/pidlayouttl8.jpg

But sadly the audio stream was not detected: -

http://img136.imageshack.us/img136/350/streaminfobb2.jpg

Please let me know if you require the 20.2MB sample

[Guest]

Please let me know if you require the 20.2MB sample Yes, I require it. Don't I ALWAYS ask for the stream? :devil:

It's probably the same problem as described in the following post, but I'd like to see it to be sure. You can just set the PID for the audio to 0x79 and it should work, if it is the same problem. Make sure Audio/Output Method is set to Demux all tracks.

[Guest]

Is there a way to know which audio tracks really exist and to avoid an invalid track being selected as the default one ? If the PAT/PMT says it is there I use it. But your stream turned up a bug that makes it work for your stream. When finding initial PIDs with PAT/PMT, the last audio channel was being used instead of the first. When I fixed that it now finds the first MPA stream, which is fine in this stream. It will be in the next beta.

[SeeMoreDigital]

Yes, I require it. Don't I ALWAYS ask for the stream? :devil:LOL :D

Here you go best mate: -

http://www.mytempdir.com/2058091


Cheers

[Guest]

Same problem: last audio used instead of first, and the last one is not really there. Just set the PID for the MPA stream.

[Underground78]

If the PAT/PMT says it is there I use it. But your stream turned up a bug that makes it work for your stream. When finding initial PIDs with PAT/PMT, the last audio channel was being used instead of the first. When I fixed that it now finds the first MPA stream, which is fine in this stream. It will be in the next beta.

Ok, thank you !

[Guest]

* Fixed problem whereby the automatic PID setting for transport streams used the last audio stream encountered instead of the first.

* Added an MRU list to the File menu.

http://neuron2.net/dgmpgdec/dgmpgdec150b13.zip

This version also contains a new feature. The first person to find it gets a brownie point. Then I will add it to the change list. Just for some fun!

[Terranigma]

This version also contains a new feature. The first person to find it gets a brownie point. Then I will add it to the change list. Just for some fun!

Thanks for the update. I should find it in no time (or so i hope :p).

---
The ability to quickly navigate to a previously opened file (i.e., bookmarking) with Recent file list?

[Guest]

That is correct. Your brownie point:

http://neuron2.net/misc/brownie.jpg

[Inventive Software]

Mmm, tasty! :D

[Guest]

Hands off! That belongs to Terranigma.

[Inventive Software]

My bad. :(

[Terranigma]

Hands off! That belongs to Terranigma.

Haha, thanks for the treat (and this wonderful update). :)

[Yoshiyuki Blade]

Just a bit curious, do you make any changes to dgdecode.dll for the betas? I replace them each time I check out a new build just in case, but if I don't have to, that'll save me a good 2 seconds of time :).

[Guest]

They are always changed at a minimum to have the right beta version number. I recommend that you always upgrade all components.

[Inventive Software]

Question for Donald, or anybody who could help. I'm having trouble with noticable interlacing lines on the picture, and I have no idea how to deal with them. Encoding it progressive, it'll look horrible, encoding it interlaced I'm not willing to do unless absolutely necessarily. Sample: http://rapidshare.com/files/70550171/FieldSample.zip. Please assist! :)

[Rainy]

DGIndex doesn't seem to recognize the end of your file; the navigation controls remain disabled, when the timeline end is reached and nothing happens. I have to hit escape to return to the initial state.

[Guest]

I had to cut some orphaned data at the end to keep DGIndex from doing that. Bad cut.

Um, this is pure interlaced video (not film). You either have to encode it as interlaced, or deinterlace it.

[Boulder]

Don,

could you please check what is causing the problems when using SetMTMode(2) with tsp's Avisynth build that has multithreading support?

http://forum.doom9.org/showthread.php?p=1065701#post1065701

[SeeMoreDigital]

Sadly the detail in many NTSC TV shows often get completely destroyed when converted to PAL. Especially if the original NTSC source was interlaced to begin with :(

[elfurbe]

I'm not sure if this is specific to the beta or not. I noticed a "new to me" feature, not sure when it showed up, Tools->Analyze Sync. I just decoded a vobset with multiple audio streams, all of which show delays of 0 in their output filenames, but when I ran track 1 through the Analyze Sync tool, the output log file showed delay on every entry, except the first. I'm not sure what to make of that, so I'm wondering a couple things.

1: Does the analyze sync tool do what I think it does or am I misunderstanding? It seems to do a frame-by-frame check of audio sync to video. Is that accurate?

2: If that's what it's doing and it's finding delay, should I/how would I use this info to get proper sync with my video and audio after encoding to a different format? Average all the reported delays perhaps?

[Guest]

could you please check what is causing the problems when using SetMTMode(2) with tsp's Avisynth build that has multithreading support?[/url] I don't have a multi-core machine.

[Guest]

I'm not sure if this is specific to the beta or not. I noticed a "new to me" feature, not sure when it showed up, Tools->Analyze Sync. I just decoded a vobset with multiple audio streams, all of which show delays of 0 in their output filenames, but when I ran track 1 through the Analyze Sync tool, the output log file showed delay on every entry, except the first. I'm not sure what to make of that, so I'm wondering a couple things. When you see a new feature, the first thing to do is to read the users manual:

"This tool allows the user to create a log of the audio delay values that would be generated if the project were to be started at all the I frames of the stream. ... Note that for a displayed delay value to be usable and correct, you must start your project at the corresponding I frame."

1: Does the analyze sync tool do what I think it does or am I misunderstanding? It seems to do a frame-by-frame check of audio sync to video. Is that accurate? See above.

[elfurbe]

Many apologies. RTFM indeed. Thanks anyway!

[remiche]

Number 13 is fatal without doubt-that!

I riping 3 DVDs for 6min!!!
Log file say: no find audio!
HoHoHooo...
I install 1.45 Stable... and all OK :) (successful RIPs OK)

[Guest]

Number 13 is fatal without doubt-that!

I riping 3 DVDs for 6min!!!
Log file say: no find audio!
HoHoHooo...
I install 1.45 Stable... and all OK :) (successful RIPs OK) Please post a link to a portion of the source stream that fails. My guess is that you aren't doing something right.

[ilhyfe]

@neuron2:

I have a feature request:

Sometimes it would be nice if the information window wouldn't close it self so that I can see the timeline while I'm scrolling through the video. Just like dvd2avi.

[squid_80]

Looong standing bug to report in dgdecode: If there's a problem opening the avisynth script (for example the source mpeg file has been moved/deleted or a version mismatch is reported) and an error is thrown, the .d2v file isn't closed and can't be deleted or modified until the host application is closed.

[Guest]

Sometimes it would be nice if the information window wouldn't close it self so that I can see the timeline while I'm scrolling through the video. Just like dvd2avi. You can set single step and then step through with the info window. The problem for jumping around is that some of the fields are not valid unless decoding linearly.

[Guest]

Looong standing bug to report in dgdecode: If there's a problem opening the avisynth script (for example the source mpeg file has been moved/deleted or a version mismatch is reported) and an error is thrown, the .d2v file isn't closed and can't be deleted or modified until the host application is closed.
When I close the D2V and then call ThrowError(), Avisynth takes a system exception. I'd have to trace through Avisynth to find out why and I'm not motivated to do that right now.

[squid_80]

If I had to guess, I'd say the code in CMPEG2Decoder::Close() is wrong:void CMPEG2Decoder::Close()
{
int i;
CMPEG2Decoder* in = this;

if (in != NULL)
fclose(in->VF_File);

Of course in is not going to be NULL. I think the check should be if (VF_File != NULL) and NULL should be assigned to VF_File after it is fclosed.

[squid_80]

Actually now that I'm back from voting and have some time to sit down, I see a better way - change CMPEG2Decoder::Close to CMPEG2Decoder::~CMPEG2Decoder. This requires a few other changes, I've zipped the modified source code files and they're at http://216.75.63.164/andrew/dgdecode_mods.zip (Based on 1.4.9)

[buzzqw]

@squid_80
while you are in modifying.. how about ifo parsing (of decrypted dvd) with pgc selection for demuxing.. ? :thanks:

(i hope Neuron2 don't be angry)

BHH

[squid_80]

I don't know the first thing about IFO parsing.

[buzzqw]

something could be in vstrip source or pgcdemux source

(i want help.. but don't know a single fprint of c/c++)

BHH

[Guest]

No IFO parsing. It's been discussed before.

[Guest]

I think the check should be if (VF_File != NULL) and NULL should be assigned to VF_File after it is fclosed. I did it that way. Thanks.

[ilhyfe]

You can set single step and then step through with the info window. The problem for jumping around is that some of the fields are not valid unless decoding linearly.

I tried. but when I scroll through the video the info window closes itself.

EDIT: I don't need many information, a timestamp only would e enough.

[Guest]

You didn't follow me.

Position on the timeline to the area of interest. Then set Options/Playback Speed to Single Step. Then hit F6 to play. Now you use the > button to single step. The Info window will stay open.

[ilhyfe]

You didn't follow me.

Position on the timeline to the area of interest. Then set Options/Playback Speed to Single Step. Then hit F6 to play. Now you use the > button to single step. The Info window will stay open.

Hm yeah but I don't want to use the single step function. I want to scroll fast through the video to time xyz.

[Zep]

Did I find a bug? (1.5.0 beta 13)

I capped a HDTV show in 6 segments stopping recording at each ad break etc... then editing lead in and lead out ads. basic stuff.

if I choose all 6 edited final segments in DGindex then output to 1 d2v file it is 59997 frames.

mpeg2source("D:\wholeshow.d2v")

and the audio is in perfect sync.




If I index each segment separately then do this

v1=mpeg2source("D:\part1.d2v")
v2=mpeg2source("D:\part2.d2v")
v3=mpeg2source("D:\part3.d2v")
v4=mpeg2source("D:\part4.d2v")
v5=mpeg2source("D:\part5.d2v")
v6=mpeg2source("D:\part6.d2v")
return(v1+v2+v3+v4+v5+v6)

OR

If I encode each section to xvid first pass then merge the first pass video.pass files then run a second pass well.... both of these ways give me exactly 60006 frames and the audio is off more and more as the encode reaches the end.

in fact the audio is BEHIND by just about 320ms (give or take)
in this example as the audio is shorter in each segment by 1 or 2 frames than the video. (about 41ms per frame the audio falls behind and it adds up more and more over the length of the encode as more segments are hit)

is there something I am missing as to why the video frame count would be so different? They should match right? Can this be fixed?

[Guest]

@Zep

If the frame count was less when doing them separately, I could understand it. But you say it is more that way.

You won't like this answer, but you're going to have to give me at least two sections so that I can duplicate your result.

[Guest]

Hm yeah but I don't want to use the single step function. I want to scroll fast through the video to time xyz. You can't always get what you want. I'm sure that with a little creativity, perhaps with the F6 key and different playback speeds, you can find your desired location without too much difficulty.

[ilhyfe]

You can't always get what you want.

I know but since xmas is near and it worked fine in *old* dvd2avi it should be possible to add :P

[Zep]

@Zep

If the frame count was less when doing them separately, I could understand it. But you say it is more that way.


You won't like this answer, but you're going to have to give me at least two sections so that I can duplicate your result.

Ok I have been doing a ton of tests and I am narrowing it down. I no longer think DGindex
is adding frames but only subtracting frames like you said you thought it would do. After a
avisynth fresh install I now only see LESS frames when I do this

v1=mpeg2source("D:\part1.d2v")
v2=mpeg2source("D:\part2.d2v")
v3=mpeg2source("D:\part3.d2v")
v4=mpeg2source("D:\part4.d2v")
v5=mpeg2source("D:\part5.d2v")
v6=mpeg2source("D:\part6.d2v")

video=v1+v2+v3+v4+v5+v6

video=video.Selecteven()
video=video.Crop(8,8,-8,-8)
video=video.Decimate(5)
return(video)

compared to this

mpeg2source("D:\wholeshow.d2v")
Selecteven()
Crop(8,8,-8,-8)
Decimate(5)

same test chunks the first way returns 2 less frames. I have now done it over a dozen times
and always returns 2 less frames. I am curious as to why less frames are returned
and why above you thought that would be? it makes me wonder WHERE the 2 frames are
cut from. (note I use projextx and VRD and edit at the GOP level) If at the very end then no big
deal but if somewhere in the middle then the audio would be about 84ms longer than the video
after it hits those 2 frame cuts)


Anyway, I am still seeing 9 frames added to the final encode. So I will continue to drive on
as it now appears Vdub/Xvid are adding frames and not sticking to what DGindex ---> avs
is sending to them. (I did a fresh install of both Vdub and Xvid also and it didn't help)


I'll post here what I do find in hopes it helps others.


Thanks