DVD2AVIdg Version 1.3.0 with transport inside! [final]

[Guest]

@dvd2svcd

Great! Sorry for the hassle with the bad archive.

@Zep

Any feedback on my wonderful, reliable PID detection? :D

[len0x]

Originally posted by neuron2

Great! Sorry for the hassle with the bad archive.


It was really lucky of me to spot the error that quickly, otherwise we could've continued tracking the problem in totally different environments :)

[Guest]

Not lucky, len0x. The word is perspicacious, observant, alert, perceptive, diligent...

[len0x]

True. But the luck I was referring to is more about "how quick" it was found. I have no doubts that it would've been solved anyway. You know how I found it? I was testing AutoGK with new DVD2AVI combo and it just happens that DVD2AVI and mpeg2dec3 are placed in a different directories in AutoGK folder. So after you mentioned that d2v structure was changed I was checking if I forgot to unpack mpeg2dec3dg in second folder and I quickly saw that they are the same but didn't notice that date was three days old at first. But it just happens that I don't use Explorer for file navigation but other file manager where the date of current file is shown in a very distinct way, so after a couple of minutes of jusr staring at the screen I noticed difference in dates betweet DVD2AVI and mpeg2dec3. The rest was easy :)

[Trahald]

penicillin was found by 'accident' too. hehe.

[Moitah]

Wilbert: I found a way to fix YV12 playback. DivX was handling YV12, and apparently not working so well :). If you use XviD instead, it should work. You can do this by adding a new string value (or change it if it's already there) in "HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Drivers32" with the name "vidc.YV12" and the data "xvidvfw.dll". This assumes you have XviD installed. I'm running Windows 2000 but it should work in XP as well.

[Guest]

Just some info. I have PVA format video decoding working now. I have to demux the audio and then we have PVA support.

[Zep]

Originally posted by neuron2


@Zep

Any feedback on my wonderful, reliable PID detection? :D

yes i just tested it and it works AWESOME :D


in a 7 stream .ts file it was able to detect them all correctly


Group 1
PCR on PID 11
Video on PID 11
AC3 on PID 14
Other on PID 1A

Group 2
PCR on PID 21
Video on PID 21
AC3 on PID 24

Group 3
PCR on PID 31
Video on PID 31
AC3 on PID 34


i call them groups, you call them programs. I have no idea
if there is an offical title. I have always called them PID groups.

[Guest]

:D :D :D :D :D

[They're officially called programs.]

[Zep]

Originally posted by neuron2
:D :D :D :D :D

[They're officially called programs.]


bahhhh that is aweful!!! who came up with that! grrr..... :D


PID Group much better describes what they are lol



like we don't have enough "programs" already for 50
other things haha

[Guest]

Program in the sense of watching a TV program!

[Zep]

Originally posted by neuron2
Program in the sense of watching a TV program!

yup another reason why it is not good IMHO

what you are listing in the app is not really a TV program
but grouped streams of data that may or not be a TV program.
Stream "Other" in my above post is not Video or Audio (as far
as i can tell) and thus not a TV program :D

and "program" 3 is not a tv program either per se' but a live
computer feed of dobbler radar.

other streams have all sorts of stuff and more and more
streams are coming online that send software and not TV at all.
You can already download 15 gigs an hour over HDTV. It doesn't
have to be a TV program it can be anything.

[HarryM]

Hi,

I have problems with DVD2AVIdg 1.3.
DVD-RB refused AVS script with version 1.3 (1.2 too). :mad:
I get only 10 seconds of obligate error messages (error at line 6 of script).

With version 1.1 final is all O.K.

[Guest]

Originally posted by HarryM
I have problems with DVD2AVIdg 1.3.
DVD-RB refused AVS script with version 1.3 (1.2 too). :mad:
Actually, you have problems with DVD-RB! I assume it needs to be revised for the new D2V file format. Please contact its authors for support.

[vispgraedde]

Originally posted by HarryM
Hi,

I have problems with DVD2AVIdg 1.3.
DVD-RB refused AVS script with version 1.3 (1.2 too). :mad:
I get only 10 seconds of obligate error messages (error at line 6 of script).

With version 1.1 final is all O.K.

Since it is the avs script it refuse... did you check that you updated the version of mpeg2dec3dg.dll that avisynth use to the one in 1.3?

[Zeul]

@neuron2

I appear to have found a problem, which is hopefully not due to me being an idiot :D
In the latest build of NuMenu4u i am now running dvd2avi (dg) on the demuxed m2v file from dvd decrypter, instead of the VOB file. I am getting a different frame count compared to dvd2avi run against the VOB file.
Example:
Pal project. Coded # = 363, which also corresponds to RB's build (outputs log file). All dg builds report the same. Avifileinfo also reports 363 frames. I load up the VOB and now have 365 frames. This corresponds exactly to the IFO (00:00:14;15). Dropping the demuxed m2v file into Scenarist gives a data end time (00:00:14;14 ie duration = 00:00:14;15) thereby removing a problem with the demuxed file. It would seem that 2 frames have been dropped when using the m2v file. I have tried using a variety of different mpeg2dec3dg.dll all with the same result. Is this plausible? It is sporadic but obviously for menus we must be exact to the frame. From what I can tell the offending files do start with an 'I' frame.

Zeul

[Guest]

Please make the VOB and M2V files available for download. Thank you.

If you can do it quick I'd be grateful as I was hoping to release RC3 tonight.

[vispgraedde]

Originally posted by neuron2
Please make the VOB and M2V files available for download. Thank you.

If you can do it quick I'd be grateful as I was hoping to release RC3 tonight.

I could have sworn RC3 was out already... :p or did you maybe mean marking RC3 as final for this version?

[Guest]

Originally posted by vispgraedde
did you maybe mean marking RC3 as final for this version? Correct. I will try to be more precise in the future.

[BoNz1]

Hi neuron2, I have a bug report for you. The two NTSC DVDs in question are Matrix Revolutions and Master and Commander. What happens is that the audio is totally out of sync with the video. From what I can see the reason is that there are simply too many frames. I will put up links to the vobs and m2v. Unfortunately, they will have to be small hopefully they will be large enough so that you can find the problem.
EDIT: You can get the samples here, http://www.geocities.com/bonzi5252/ it the the samples.zip file, unfortunately the problem is hard to see with these very small samples. The audio gets progressively more out of sync as the video plays. This beta had more frames than dvd2avi 1.76 and dvd2avidg 1.0.

[Guest]

@BoNz1

I don't care about 1.76. But you are saying that with your VOB I will get a different number of frames if I Save Project for 1.3.0 RC3 versus 1.0.0?

Please state exactly how I am to duplicate a problem with the VOBs you are providing.

What is the m2v and why is it relevant?

EDIT: I can detect no difference in handling of your VOB between versions 1.0.0 and 1.3.0. Can you please explain exactly how I am to reproduce your problem? Thank you.

[BoNz1]

I don't care about 1.76. But you are saying that with your VOB I will get a different number of frames if I Save Project for 1.3.0 RC3 versus 1.0.0?

neuron2, sorry I was in a rush to get the vobs up BTW like I said earlier I think it may be difficult to duplicate the problem with such small vobs. I shouldn't have uploaded w/o checking closer. BTW there is a difference in the number of frames between the two dvd2avidg versions for the entire movie but it isn't as big as I thought it was, 187154 frames for 1.3.0 RC3 and 187144 for 1.1.0. For both however the audio is completely out of sync. As for the m2v, I have no clue why I uploaded that I think it probably was because you asked for it from someone else :confused:. Anyway, there is a bug somewhere here. The audio shouldn't be out by 30 seconds or more.

[loni_blues]

@neuron2 and len0x

Hi,
I would like to report a difference, probably not a bug. I use Robot4rip with GKnot. Since Dvd2avidg 1.1.0 the program fails to detect the correct "FILM" percentage (i.e.: for a film that is 99% FILM I get 13% FILM and so, no forced film is applied by R4r). Version 1.0.0 detects the percentage correctly. I am calling this a "difference" because I think it has to do with the d2v estructure analysis Robot4rip does. I think there may be problems with AutoGK as well, though I have not tested it.

Regards and thanks for any help,
loni_blues

[len0x]

Originally posted by loni_blues
I am calling this a "difference" because I think it has to do with the d2v estructure analysis Robot4rip does. I think there may be problems with AutoGK as well, though I have not tested it.


It very well may be that R4R cannot work with new DVD2AVIdg, but nothing we can do about it since R4R is not being developed anymore.

P.S. latest AutoGK is shipped with latest DVD2AVIdg and has no problems with it.

[Guest]

I have just confirmed problems in RC3.

I ran the complete process on Matrix Reloaded. It had 7 VOBs. The first problem is that DVD2AVIdg 1.3.0 dropped the last GOP of the last file, losing 3 frames. But worse, MPEG2DEC3dg 1.3.0 miscounted the number of frames in the D2V and generated a video that was too long.

Obviously, I will fix these immediately and make an RC4.

Thanks Zeul, BoNz1, and loni_blues, for bringing it to light.

len0x, better to wait for formal releases, I suppose. ;)

[len0x]

Originally posted by neuron2
len0x, better to wait for formal releases, I suppose. ;)

I have not released a stable version fo AutoGK yet, but just beta, so that is supposed to have potential problems by definition :)

[Guest]

Oh goodie. I'm home today and I've already fixed MPEG2DEC3dg. Now I'm fixing that lost last GOP. There'll be a good version very soon. ;)

[Zeul]

@neuron2

excellent news. I spend so much time coding and tinkering with NuMenu4u, i thought i was going insane. Apologies again for not being able to upload the offending files.

Zeul

[Zep]

Originally posted by neuron2
I got bored today and decided to write a PID detector.



Here is sample output:

G:\tmp\Zep>table abc.3.video.3.audio.streams.tp
------------------------------------------------
Table 1.0.1 -- PAT/PMT Parser by Donald A. Graft
------------------------------------------------

Program 1
PCR on PID 0x11
Video on PID 0x11
AC3 Audio on PID 0x14
Other on PID 0x1a
Program 2
PCR on PID 0x21
Video on PID 0x21
AC3 Audio on PID 0x24
Program 3
PCR on PID 0x31
Video on PID 0x31
AC3 Audio on PID 0x34


oh S@#$ i just noticed that was done on MY ABC clip :D


wow I can't believe i didn't spot that first time around lol

[Zep]

Originally posted by neuron2
I have just confirmed problems in RC3.

I ran the complete process on Matrix Reloaded. It had 7 VOBs. The first problem is that DVD2AVIdg 1.3.0 dropped the last GOP of the last file, losing 3 frames. But worse, MPEG2DEC3dg 1.3.0 miscounted the number of frames in the D2V and generated a video that was too long.

Obviously, I will fix these immediately and make an RC4.

Thanks Zeul, BoNz1, and loni_blues, for bringing it to light.

len0x, better to wait for formal releases, I suppose. ;)

has anyone mentioned problems with decode + normalize at 100%?


none of my wavs are normalized to 100%. I will have to check
and see to what they were normalized to first chance i get.
(sounds about oh 70% maybe less)

UPDATE: yup I loaded into cooledit and doubled checked
and instead of 100% it is only doing about 50% to 60%

[loni_blues]

@len0x

Sorry for my complete ignorance -I am one of those who sticks to Gknot- but I thought AutoGknot was using R4r as well.
Regards,
loni_blues

[jimmy basushi]

loni, you should download autogk and have a look yourself, its an interesting program to look at but not for anyone who likes tinkering with thier avs files.

neuron2, until i looked at this thread i was about to update.. is version 1.2.0 considered stable by you? last i looked it was still marked as a rc.

[Guest]

Well, folks, here is 1.3.0 RC4:

http://neuron2.net/fixd2v/decodefix130RC4.zip

I was tempted to go back to beta on it because a substantial rewrite was required. The code that handled running out of bits at the end of the last file was a total kludge (that 'Dry' business). Now it is more rational. I hope I didn't break too much but it was really needed. The problems I was able to duplicate are fixed. I'd appreciate it if the people having the problems can report back with their results.

@Jimmy

The last guaranteed version is 1.1.0. Try 1.3.0 RC4. I need the testing. :)

EDIT: One known bug found in RC4. Play and Preview do not work after Save Project. I'll fix this in the next version. I'd like to get feedback on whether this version fixes the frame count issues that were reported.

[Guest]

Originally posted by Zep
has anyone mentioned problems with decode + normalize at 100%? I'm not too bothered because all that audio stuff is going to be replaced (or removed). But thank you for pointing it out.

[Guest]

Here's the fix:

http://neuron2.net/fixd2v/decodefix130RC5.zip

Sorry for all the versions. I will wait for feedback on this one. I can't find anything wrong with it.

[Cyberia]

I can't find anything wrong with it.

ditto

[Zep]

Originally posted by neuron2
Well, folks, here is 1.3.0 RC4:

http://neuron2.net/fixd2v/decodefix130RC4.zip

I was tempted to go back to beta on it because a substantial rewrite was required. The code that handled running out of bits at the end of the last file was a total kludge (that 'Dry' business). Now it is more rational. I hope I didn't break too much but it was really needed. The problems I was able to duplicate are fixed. I'd appreciate it if the people having the problems can report back with their results.



I noticed it changed because first time ever I saw all green frames at end of file.
No big deal as the last few frames are trim() out but I'm hoping it fixes my crash
on each pass also in VDub. For some reason avisynth was feeding VDub/Xvid
data it choked on, on the very last frame (even with a trim)

find out in 20 minutes as pass one with RC5 does its thing (I missed RC4)

[Zep]

Originally posted by neuron2
I'm not too bothered because all that audio stuff is going to be replaced (or removed). But thank you for pointing it out.

NP I try to spot them :)


woah really? I hope you leave it in. if I have to use dvdtoavi (i.e. step 1)
I would rather the normalize be done there on my wavs instead of in avisynth.

Hmmm...

[Zep]

Originally posted by neuron2
Here's the fix:

http://neuron2.net/fixd2v/decodefix130RC5.zip

Sorry for all the versions. I will wait for feedback on this one. I can't find anything wrong with it.

doing my first cap now with this version. might take a few days to really put
it through its paces :D so far so good though I did notice the green frames
in last GOP unlike prior versions. trimmed out like always since I go a few minutes
long when raw capping.

thanks for the update. I really wanted this one :)

[Guest]

The green frame (sometimes pink) is fixed in my latest for release. Actually, it appears only at the end of Save Project sometimes, and never appears in MPEG2DEC3dg output. Stuff like that is always a big deal for me. :)

The problem is that at the end of the Save Project the last frame is displayed but the previous frames were not decoded, so it may not have its reference frame. Now at the end of a Save Project I don't display the last frame, but just write the coded frame count into the info dialog instead. Then I raise a popup for the user to dismiss, after which I navigate back to frame 0. This avoids ever positioning on a frame without its reference frame having been decoded.

[Lord_KiRon]

Sorry for being a bit off topic and newbie , however I have a question directly related to DVD2AVIdg and you looks to the kind of people that can help :

I made a rip of DVD (NTSC 16:9 anamorphic 3 movies on disk) using roborip , the "problem" is that when I run DVD2AVIdg (any version) from roborip the 1st and 3rd movie (it's ~40min series) all the time "jump" NTSC/FILM , while it looks like 2nd movie is set to FILM (forced ?) all the time and this causes different field operation descission in the resulting .D2V file - "None" for 1st and 3rd and "Force FILM" for the 2nd.
Also somehow the system detects FPS as 29.7...fps while DVD is NTSC , shouldn't it be 30 fps?

It wouldn't get me worry but the 2nd movie is compresses WAY better then first and 3rd , I set 900 bitrate for all 3 movies (XviD,Fast recompress to remain in same color space,2 passes,.AVS created with GK) and 1st and 3rd I got 800+ bitrate as it was supposed to in resulting AVIs while with this 2nd movie the resulting bitrate is 670 no matter what bitrate I set even up to 3000.
Can some one explain me that ?

Also does anyone know a good guide for newbies like me explaining use and meaning of DVD2AVIdg options/parameters , I was using roborip before but I a outgrown it now and it's no longer updates , while AutoGK is too automatic for my taste so I need to lurn to control DVD2AVI manually.

[Zeul]

@neurnon2

Is there any chance of you adding another commandline switch which would output a log file similar to the RB build. The information provided by this log file has proven to be very useful in determining field order / scanning order, and is now extensively used by NuMenu4u.

Thanks
Zeul

[len0x]

Originally posted by Lord_KiRon
Also does anyone know a good guide for newbies like me explaining use and meaning of DVD2AVIdg options/parameters

essential reading:
http://www.doom9.org/ivtc-tut.htm

[Lord_KiRon]

Ops missed this one :( , and also the "essential reading" stycky thread here :( , sorry guys next time I will put more attention .

EDITED:
OK , now after reading all the guids I am a smarter person now ;) , so this brings me to the next question : if I have a series and some episodes detected as 100% film and some not (percentage is high but bellow 95%) should I force "Force FILM" option on them ?

[Zeul]

@neuron2
As an update dvd2avidg (RC5) now reports the same number of frames for the offending VOBs and M2V, when run through avs script. The main window (coded #) still shows a difference in preview mode. So thanks for your hard work on this.

@Lord_KiRon
Only use force film if >95% film. anything less use no force film and apply telecide to your avs script.

Zeul

[Guest]

Originally posted by Zeul
Is there any chance of you adding another commandline switch which would output a log file similar to the RB build. The information provided by this log file has proven to be very useful in determining field order / scanning order, and is now extensively used by NuMenu4u. I've answered this several times. It's on my work list.

Now you please tell me specifically what information you want to see in there. But please do it in Cyberia's "development" thread.

[Guest]

@Zeul

The DVD2AVI GUI and navigation is a mess. A slight inaccuracy on frame counts in Preview is a consequence of it. I plan to redo all that stuff so it's not worth bothering with right now. The main thing is to get a correct D2V file written. Thanks for reporting your results.

@Lord

You can consider encoding the movies separately, so that you can apply FF to the parts that are all film.

I plan to write a tutorial after the formal release.

[Lord_KiRon]

Originally posted by neuron2

@Lord

You can consider encoding the movies separately, so that you can apply FF to the parts that are all film.

The "problem" is that it "looks" like the whole episode is a film , the only thing I can think of is that episodes that are not detected as film were edited after telecine ...

[Zep]

Originally posted by neuron2
The green frame (sometimes pink) is fixed in my latest for release. Actually, it appears only at the end of Save Project sometimes, and never appears in MPEG2DEC3dg output. Stuff like that is always a big deal for me. :)

The problem is that at the end of the Save Project the last frame is displayed but the previous frames were not decoded, so it may not have its reference frame. Now at the end of a Save Project I don't display the last frame, but just write the coded frame count into the info dialog instead. Then I raise a popup for the user to dismiss, after which I navigate back to frame 0. This avoids ever positioning on a frame without its reference frame having been decoded.


hmmmm... I saw at the end of MPEG2DEC3dg output feed to avisynth--->VDub.
The good news NO crashes on last frame on either pass.
That is a first with the 1.3.0 versions.


ahhh yeah that seems like a good way to do it. But what
MPEG2DEC3dg output? I am getting green frames at end.
I will double check it all later when i get back from work
and try to give more details and maybe UL a clip for you
if i can get it to re produce on a small .ts.

[Guest]

@Zep

If you pass a stream to MPEG2DEC3dg with incomplete or corrupted final pictures, you will get what you deserve. The only thing I can do is not crash.

EDIT: It is *possible* to suppress the last picture but I think it runs against my general philosophy of delivering all pictures in the stream. If there is a picture header, as far as I am concerned, there is a picture. :p