Possible custom revision to DVD2AVI?

Does anyone have the know-how, as well as willingness to modify DVD2AVI 1.76 CLI for us? The situation is this: DVD2AVI works great for main movies demuxed by PGC. However, when we try to recreate a seamless-branching DVD, we have to demux by VobID. At this point, DVD2AVI becomes unreliable because the .m2v doesn't start with an I-Frame, and DVD2AVI seems to skip along until it finds one, neglecting several frames in the process of doing so. As a result, the .d2vs created using these .m2vs are almost-always incorrect and when used in CCE result in a transcoded .mpv that is shorter than the original. Can anyone modify DVD2AVI for us to where it will start making the .d2v from the start of the .m2v? From what I understand, it is possible for DVD video to start with a P-Frame also, so this inability of DVD2AVI to do this actually becomes a flaw, right? Also, does anyone know if the VStrip .d2v creation suffers from this same limitation? I haven't been able to get VStrip to output the .d2v file...

I don't think it is really valid to start with anything but an I frame in MPEG2. And also I'm not sure the problem is really this or that it should also skip any audio before the first valid frame. But I'm afraid I still don't understand DVD2AVI audio processing. - Tom

If it just skipped audio we'd end up with a reauthored DVD missing even more data than we do now. The objective is to get a DVD that plays the same as the original. I've also asked Light_UK if it's possible for him to split vobids using the nearest I-Frame. Somehow we need a working solution for this issue or the dream of reauthoring DVD without one-button tools (dvd2one, instantcopy, etc.) will never be a reality :(

Even if DVD2AVI could be modified to do this, the d2v file would need to be changed to accomodate it. The reason being that the lines in the d2v always begin with an I-frame (7). If the d2v is changed, then mpeg2dec will also need to be changed to cope with the new format. Perhaps something stranger is going on here. I did an experiment with a 2-frame m2v, which is the first video stream in a VOB. On loading into DVD2AVI, I could only see one frame - the first frame was missing. After re-encoding, I loaded the resulting mpv into DVD2AVI and guess what... nothing appeared at all - zero frames! But how can this be since my re-encoded video must start with an I-frame?! Previewing each of these files in my software player correctly shows 2 frames and one frame respectively.

Why split the vobs by Vob-/Cell-ID in the first place? If we let DVD2AVI process the whole VTS-set and then use trim in AviSynth then we could create custom AviSynth scripts for each Vob-/Cell-ID anyway. And finding the correct trim-points should be possible by parsing the .d2v. Or does the audio become impossible to handle then?

There are a couple utilities now that allow you to save your vdub scripts as avisynth trims somehow. I wonder if it would be possible to use vdub to trim and save the audio trim controls and then use Avisynth with those trims and whatever other encoding utilities that you use (but I don't)? I use vdub anyway but one of the nice things I like with VirtualDubMod is I can trim the HDTV commercials from the video and ac3 before encoding without worrying about where the key frames are. - Tom