Re: [Dvdauthor-users] Vamps

[Erik S. <er...@ol...>]

> That sounds great, but how did you solve the problem with dvdauthor
> choking on the VOBs afterwards?

You mean the "SCR moves backward" problem? I had that on one vob from my
testing DVD. I tried various things with little success, in the end I
simply demuxed it, remuxed it and re-added the subtitles. The a/v sync
appeared to be OK, which I did not expect. Only the subtitles' colours
are wrong now, so I also had to hack dvdauthor a bit, to be able to
really set the palette using the palette file.

> I've had issues with VOBs extracted from a DVD with transcode when
> trying to re-author them back with dvdauthor.

I either use dvdunauthor (which now should work together with vamps) or
a script that uses vamps_play-title to extract all chapters individually
and pipes them through vamps directly. I quite like the approach because
it easily keeps chapter info, gives me the opportunity to delete
chapters and also allows me to remove some audio and subtitles
languages. The drawback is that you loose original menus and chapter
picking etc. I don't mind, I make my own using the same script. The last
dvd I run dvdunauthor on, gave me over 80 vob files, most of them being
a part of menu's. I choose to have it simple, only kept the titles and
made my own menu.

> # re-author the title
> dvdauthor -o ${proj} -c `cat chapters.txt` ${proj}.vob
> dvdauthor -T -o ${proj}

My script scans titles, chapters and palette from lsdvd -x, rips each
chapter in a seperate vob and then builds an xml file for dvdauthor,
with the proper attributes.

> Problem is, dvdauthor doesn't like the VOBs generated by tccat. That's a
> really big bummer, because tccat is such a convenient way to extract
> exactly one title in exactly one VOB - easy to process with vamps and
> the like.

What are the symptoms?

>> Q&D patch available on request.
> Ok, this is a request. :-)

I lost the original source file, so this will be a handcrafted "patch":

The only thing it does / appears to do is to step over padding stream
packets which otherwise upset vamps. It seems to work, but no guarantees
;-)

------------------------------------------------

     case 0xbb:
	// system header/private 2: PCI/DSI
	copy (SECT_SIZE);
	break;

+	  case 0xbe: // skip padding stream EMS
+		skip(SECT_SIZE);
+		break;

      default:
	fatal ("Encountered stream ID %02x at %llu, "
	       "probably bad MPEG2 program stream", id, rtell (ptr));
    }

-----------------------------------------------

      case 0xbd:
	// private 1: audio/subpicture
	copy_private_1 (ptr);
	break;

+		case 0xbe: // skip padding EMS
+			skip(SECT_SIZE);
+		break;

      case 0xbb:
	// system header/private 2: PCI/DSI
	copy (SECT_SIZE);
	break;

-----------------------------------------------

      case 0xbb:
	// system header/private 2: PCI/DSI
	nav_packs++;
	break;

+	  case 0xbe: // skip padding stream EMS
+		break;

      default:
	fatal ("Encountered stream ID %02x at %llu, "
	       "probably bad MPEG2 program stream", id, rtell (ptr));
    }