Can I use DVD2AVIdg/MPEG2Dec3dg 1.3.0 with DVD-RB?

[fritzdis]

I know I can't use MPEG2Dec3dg 1.3.0 with the d2v created by DVD-RB, but could I create a d2v using DVD2AVIdg 1.3.0 (with Field Operation set to Raw Encoded Frames) and use that to replace the d2v created by DVD-RB? I want to do this because I have a d2v created by DVD-RB that gives me an error on the second cell. If I create a d2v from the same VOBs with DVD2AVIdg 1.3.0, the error doesn't occur.

[fritzdis]

Well, I tested it out, and except for the cell where the error had occurred, every single m2v created from DVD2AVIdg 1.3.0's d2v was identical to the m2v created from DVD-RB's d2v. Rebuilding went smoothly and it seems to play fine with IFOEdit, so all that remains is for me to create the image and burn it.

[jdobbs]

Originally posted by fritzdis
I know I can't use MPEG2Dec3dg 1.3.0 with the d2v created by DVD-RB, but could I create a d2v using DVD2AVIdg 1.3.0 (with Field Operation set to Raw Encoded Frames) and use that to replace the d2v created by DVD-RB? I want to do this because I have a d2v created by DVD-RB that gives me an error on the second cell. If I create a d2v from the same VOBs with DVD2AVIdg 1.3.0, the error doesn't occur. No. Two reasons: DVD-RB makes a special version of the D2V; and DVD2AVI adds an additional bit to the values output for each frame making it too large for DVD-RB.

[fritzdis]

Originally posted by jdobbs
No. Two reasons: DVD-RB makes a special version of the D2V; and DVD2AVI adds an additional bit to the values output for each frame making it too large for DVD-RB.
Did you check my second post? It sure seemed to work.

Are you aware of the "Raw Encoded Frames" mode under the "Field Operation" setting? It was introduced in v1.3.0.
From the decodefix130 changes file
11. Added a new field operation mode: 3 = Raw Encoded Frames.
This mode suppresses pulldown and delivers the raw encoded frames.
It differs from Force Film in that no frames are added or
deleted (to maintain film rate) and the frame rate is
left at its coded value. The RFF flags are left intact in
the D2V file but they are ignored during decoding.
Is this different from what DVD-RB does?

As far as any additional bits are concerned, I'm not sure what you mean. Since the M2Vs from DVD-RB's D2V were identical to the M2Vs from DVD2AVIdg 1.3.0's D2V in my test, I wonder if something was changed by the time 1.3.0 came out such that any extra bits were gotten rid of.

[jdobbs]

Yes, I checked your post and I'm aware. In fact I believe that feature was added as a result of the method I created, and described to Neuron2.

Look in your .D2V file and see if you have frame references that are not the number "2". If you see two digit references you'll see what I mean.

My guess (just a guess) is that what actually happened is that you've installed v1.3.0 in your AVISYNTH plugins directory and it is being used as the default decoder for .D2V files. That was probably the cause of the original error -- and the reason it worked when you tried it using mode 3. Either that or you were very lucky with the source you used.


I've created D2Vs with 1.3.0 and tried them with the old MPEG2DEC3DG.DLL, and they fail. Here's and example of 1.3.0 output in mode 3:

7 0 2048 1 1 12 12 12 12 12 12 12 12 12 12 12 12
7 0 57344 1 1 2 2 12 12 12 12 12 12 12 12 12 12
7 0 110592 1 1 2 2 12 12 12 12 12 12 12 12 12 13

Here is that same section from DVD-RB:

7 0 2048 2 2 2 2 2 2 2 2 2 2 2 2 2 2
7 0 57344 2 2 2 2 2 2 2 2 2 2 2 2 2 2
7 0 110592 2 2 2 2 2 2 2 2 2 2 2 2 2 2

[fritzdis]

I think maybe I didn't make it clear enough what I did.

First I'll describe my original attempt to encode which gave me an error:

I have no versions of MPEG2Dec3dg in my AviSynth plugins folder. In the DVD-RB setup, the "Path to MPEG2DEC3DG" was set to "C:\AV\decodefix100\MPEG2Dec3dg.dll" and the "Add to AVS File" option was checked.

After the Prepare phase, 20 AVS files had been created for the movie. After encoding, I realized that there was a problem with the second file, so I loaded every AVS into VirtualDub, and the second was the only one where the error was evident.

Let me describe the error in detail so it is clear why DVD2AVIdg 1.3.0 is needed. When I load the second AVS into VirtualDub, I see that the first 3 frames are all green when they should be black. Also, there is a scene cut from black to an image that is delayed by 6 frames, making the resulting m2v out of sync with the audio. I suspect this error is at least partially a result of one the following:
From the decodefix110 changes file
4. Random navigation performance is greatly improved. The improvement is gained only for dg-style D2V files. It requires the new versions of both DVD2AVIdg and MPEG2DEC3dg. From the decodefix130 changes file
4. All versions of DVD2AVI were not putting correct LBA addresses in the D2V file. Except for LBA=0, the LBAs were stored as one LBA less than the correct value. I corrected this by revising the indexing concept. Instead of storing the LBA I now store an exact address as described in the D2V format document. The result of the bug was that MPEG2DEC3dg had to do extra parsing on every random access, so a minor performance gain is achieved by the fix.
The invocation of Trim in the AVS file is achieved by random navigation. The issue is probably further confused by the fact that the opening GOP of the second cell is open. It may in fact be that the error only occurs when random navigation to a particular point in an open GOP occurs, which could be why this error hasn't popped up before.

Now, let me explain what I changed that resulted in error-free files:

First, I changed the "Path to MPEG2DEC3DG" to "C:\AV\decodefix130\MPEG2Dec3dg.dll". After preparing with DVD-RB, I loaded the movie VOB files into DVD2AVIdg 1.3.0, set the Field Operation mode to Raw Encoded Frames, and saved the project as V01.d2v (replacing the d2v created by DVD-RB) in the D2VAVS folder in DVD-RB's working folder. I then opened V01.d2v in WordPad and changed the framerate from 29970 to 23976 (sorry I forgot to mention this before). When I loaded the second AVS into VirtualDub, the green frames were gone and the scene changed was no longer delayed. I then encoded the files and checked them all for errors, and they all seemed fine. Next I did a binary comparison between the newly created M2Vs and the old ones, and except for the second one, they were identical.

So here's what I saw:
When reading the d2v created by DVD-RB with MPEG2Dec3dg 1.0.0, the error was evident.
When reading the d2v created by DVD2AVIdg 1.3.0 with MPEG2Dec3dg 1.3.0, the error was gone.

If you want to check any of this for yourself, the disc is the R1 Widescreen edition of Intolerable Cruelty.

[fritzdis]

I've done several more encodes, all of which have turned out fine from what I can see. At this point I'm convinced I can use DVD2AVIdg 1.3.0 along with MPEG2Dec3dg 1.3.0 if I want to. What I'm more interested in now is if the error I encountered is replicable, and if so, is it serious enough for jdobbs to consider updating the d2v generating portion of his code?

[jdobbs]

I gave my answer as to whether it is compatible earlier and you've decided I was wrong. I still think you're mistaken, but it's your call. Of course it will work if you want to reencode the output a second time (for each VTS) using DVD2AVIdg v1.3.0 every time you prepare a disc -- but why would anyone want to do that? There really is no advantage.

I would recommend you use v1.1.0 -- which is compatible with DVD-RB (and the one I use). As for the output of the DVD2AVI in your second quote... again, DVD2AVI isn't used with DVD-RB -- DVD-RB generates its own .D2V.

But -- to anyone else who attempts to use this version based on this thread, and if fails... send your complaints to fritzdis, not me.

[fritzdis]

Originally posted by jdobbs
I gave my answer as to whether it is compatible earlier and you've decided I was wrong.
I'm very sorry if it sounded like I said you were wrong. I'm saying that if I want to go through the hassle of running the VOBs through DVD2AVIdg 1.3.0 with Raw Encoded Frames set, then I can use MPEG2Dec3dg 1.3.0. Is that wrong?

Originally posted by jdobbs
Of course it will work if you want to reencode the output a second time (for each VTS) using DVD2AVIdg v1.3.0 every time you prepare a disc -- but why would anyone want to do that? There really is no advantage.
By reencode the output a second time, do you mean put the VOBs through DVD2AVIdg after the Prepare phase? Because I wouldn't encode with CCE twice, if you thought I meant that. I'm not saying you thought that, I'm just a little confused.

Anyway, I would want to do what I've described because I get an error if I don't. That's a big advantage.
Originally posted by jdobbs
As for the output of the DVD2AVI in your second quote... again, DVD2AVI isn't used with DVD-RB -- DVD-RB generates its own .D2V.
I'm sorry. The error I saw using DVD-RB's D2V was replicable with DVD2AVIdg 1.0.0's D2V so I assumed there was shared code. If that were the case, fixes to DVD2AVIdg could apply to DVD-RB as well. Those quotes were only meant as possible explanations for the error. Whether or not they are in any way relevant to DVD-RB doesn't change the fact that the error occurred with the D2V created by DVD-RB.

[fritzdis]

I've realized that I may have jumped the gun when placing any blame on DVD-RB's creation of the D2V. Since neuron2's site describes the random navigation problems as being related to MPEG2Dec, it could easily be MPEG2Dec3dg 1.0.0 that is entirely responsible for the error I am seeing. I can't find any place to download decodefix110, so I can't check if the error would go away with 1.1.0.

jdobbs,
I'm really sorry if you were upset by anything I said. That was never my intention. I just want to figure out what is causing the error I am seeing.

[jdobbs]

I'm never upset -- but I am bull headed. ;)

BTW -- my intention is to skip over version 1.3.0 completely and go to Neuron2's DGMPGDec as the decoder. 1.3.0 is pretty hard to find these days.

[Fr4nz]

Why not include it in your package (with the author's permission obviously) ??? :D