TyTool and DVDlab/Chapters

[bato]

If you use DVDlab with tytool please help me find if this problem is present with you too.

- process your ty with TyTool and cut commercials
- import your mpeg into DVDlab 1.3b1 (mux/demux works the same)
- create chapters/ create index/ compile with force-index
- check the resulting VTS_01_0.IFO file with IFOedit to check if the chapters show the same frame number as the green diamonds (chapters) in DVDlab

For me it doesn't work. If this worked then is no need to use the GOP time code fix with mpeg2vcr.

If your souce is a movie and you only cut start/end (not inbetween) then you have no problem with chapters, if you cut several places then you have GOP time code problems that DVDlab can't handle to place correct chapters.

Anyone? Thanks.

[casagw]

I am still trying to figure out how to get DVDLab to work with the chapters all lining up. I have a SA Tivo and am using DVDLab 1.3 BETA1 and TyTool 7r8

I do not have mpg2vcr so I am trying to find a way around it.

What is funny is that some times it works and some times it does not work. It have tried many different methods of compiling the video (using VOBS, IFOEdit, and TMPGenc) to try to fix the GOP timing so that chapters line up. None of them seem to work.

Here is what I did find out though.

I compiled a few episodes of Stargate SG-1. I did not frame index the movie when I entered the chapters, but I did click on the frame index option in addition to the flexible source in the compile DVD options.

The first time I did this, every chapter lined up perfectly. I then added an additional episode to the DVD and recompiled. Interesting though, on the second try, all the old episodes had screwed up chapter points, but the new episode had perfect chapter points.

I am trying to delete the index files that DVDLab creates and hopefully all episodes will line up. I don't know what DVDLab does the first time around to make it work, but it seems that every time I try to recompile or frame index before I make the DVD, the chapters are messed up.

Although I am thankful for everything that JDiner has done for us, it would be nice if he would correct the GOP timing after the cuts, either that or DVDLab would work with it properly.

I am still learning this stuff so if anybody has a better suggestion, please let me know.

[bato]

Originally posted by casagw
...I do not have mpg2vcr so I am trying to find a way around it.
...
The first time I did this, every chapter lined up perfectly. I then added an additional episode to the DVD and recompiled. Interesting though, on the second try, all the old episodes had screwed up chapter points, but the new episode had perfect chapter points.
...

Yea, adding $120 of software just to fix chapters in DVDlab is a lot for most people, I'm seeking another way too.

About what you did, maybe you are into something, that looks like a bug in DVDlab, I'll try to reproduce it, if I can then I report to Oscar, maybe he can find the bug then we no longer need mpeg2vcr to change GOP time code, just as the index process in DVDlab say is not needed.

I new something is broken because I always get the same chapter points with index and without, I'll change my process to match yours and report back.

Thanks.

[casagw]

And so, after re, re disposing of the monster....

Ok, here is what I found out.

1. You must use frame index chapters in the compile DVD menu. It will not work without it.

2. Here is the kicker... The chapters must NOT be frame indexed before hand. To tell, you MUST see the red makers in the movie for the chapter points. If you see the green diamonds (meaning that the chapters have been frame indexed) it will not work!

To get rid of the green diamonds, I
1) Deleted all the index (.idx) files in the movie directory (may or may not be necessary, but that is what I did)
EDIT - It looks like it is necessary, the chapters were messed up when I did not delete them.
2) And this is necessary.
- Right click on the movie and under Chapters, export chapters for each movie.
- Right click on the movie and under Chapters, delete all chapters
- Right click on the movie and under Chaptes, import the chapters
- Do this for each movie and now you will see the red chapter markers instead of the green frame indexed chapters.

Now compile and it should work.

I guess this is an error on DVD-Lab, but this workaround does work for me. If you could pass it on so maybe we will not have to do it in later iterations, it would be great!

[bato]

Thanks, no wonder I always have the chapter problems, I followed the instructions in mediachance.com/dvdlab page, and it say to frame index chapters, so I always get the green diamonds prior to compile.

When I get home I'll try this to take some shots and pass the information to Oscar. I bet if he gets the information prior to next beta this will be fixed for next release.

[casagw]

Testing, testing, still testing.

Ok, when I reinserted the chapter points so they were the red markers and did not delete the indes files, the chapters did NOT work!!!

When I redid it, with the index files deleted, red markers on the chapter points, and frame-index chapters selected in the compile menu, it worked like a champ.

So,

- You must delete the .idx files
- You must remove and reinsert the chapter points to make the red markers (as shown on previous post)
- Use the frame-index chapter option when compiling (as well as flexible source)

I have been pulling my hair out trying to figure out why it worked some of the time and I am pretty sure I have it nailed down this time.

[bato]

casagw: I was able to test your method, it works, well sort of, all chapters now work but they start waaay before they should, just for testing I put a chapter almost at the end, after compile I pressed the last green diamond (it turns from red mark to green diamond after compile) and it shows frame 53300, but if I open VTS_01_0.IFO with IFOedit I see that the title is 42:10.29 and the last Frames value is 75928, so the index miss around 22000 frames (75928 / 30frames / 60sec = 42.1822min so that's the right count for my 42:10.29 show)

Your chapters work ok?
I mean they start playing just where you put them?
Are you using Dtivo files or SA tivo?

[texmex]

I used ReStream to fix the GOP timecodes and was able to compile a DVD with good chapters. I'm pretty sure that I did NOT use force-index. I'll run the process again when I get home this weekend and post my results here. I did have to write a small app to translate the TyTool cut points into chapters, at which point I could import those directly into DVDLab.

Here was the process (I think - I'll verify this weekend):
- TyTool to extract, edit, compile to MPEG
- DVDLab: Demux
- Restream: fix GOP timestamps
- Translate TyTool cuts into chapter file
- DVDLab: import fixed vid stream, import chapters
- Make fancy menus
- Compile DVD

[casagw]

Bato,

I am using a SA Tivo. As far as I can tell, the chapters line up correctly and each episode ends where it should. When the chapters are messed up, it loops back to the beginning to make up for the "extra" time (In WinDVD)


- Texmex - What settings are you using for ReStream? Also, is your program to change the cut file into viable chapter points posted? Thanks EDIT: I just played with ReStream, since it only changes the timestamp of the video, what happens to the audio? Does it synce up?

[bato]

casagw: thanks for the info, I'm using Dtivo, so I guess the variable GOP size we have is another problem for DVDlab, I'll try to find more about this problem I have; I put 5 chapters with DVDlab for a 42min show, it display every a chapter every 7 min or so, but in the end all chapters are about 4-5 min so the last chapter is at minute 25 or so and I placed it at minute 35, all work just they start early, when they are messed up the 2 or 3 last chapters point way out (like 12hrs) and the very last point at very end and makes PowerDVD start the video again, then I can't select previous chapters.

texmex: thanks for the info about ReStream, but to use it you need to have a demux video file, then it creates another video file and then you supply that to DVDlab, and casagw confirm (and show a way to reproduce the bug) there is a problem with DVDlab that needs to be fixed. Mpeg2vcr can fix the time code and make me create good chapters and only change some information in the already muxed file so is very fast (less than a minute for Scan then Fix GOP timecode and PTS errors, just don't select fix the GOP size errors because that will take-time/add-size breaking all long GOPs (with Dtivo files) into "legal" DVD size ones. With good timecodes (with ReStream or mpeg2vcr) then you don't need force-index option.

At the end I want DVDlab to work right so we don't need mpeg2vcr or ReStream to change the GOP time code in our mpgs created with TyTool.

I'll do more tests to explain in detail the frame-index problem with DVDlab to Oscar so he can fix it, I just need to find out about the "no-counting-every-frame-index" with Dtivo files in DVDlab.

[bato]

After some tests I decided to call this bug the "Green Diamond Bug", if your movie show Green Diamonds for chapter points then DVDlab will NOT use the index to create frame-index points, if you have Red Marks for chapters points and you select Frame-Index Chapters in Compile step then the index file will be used to create the chapters.

I can't find a way to fix the chapter-start-early problem with my frame-index chapters, but I tell Oscar about this 2 problems so he can take a look at the code.

[symm]

Originally posted by bato
casagw: thanks for the info, I'm using Dtivo, so I guess the variable GOP size we have is another problem for DVDlab, I'll try to find more about this problem I have; I put 5 chapters with DVDlab for a 42min show, it display every a chapter every 7 min or so, but in the end all chapters are about 4-5 min so the last chapter is at minute 25 or so and I placed it at minute 35, all work just they start early, when they are messed up the 2 or 3 last chapters point way out (like 12hrs) and the very last point at very end and makes PowerDVD start the video again, then I can't select previous chapters.

texmex: thanks for the info about ReStream, but to use it you need to have a demux video file, then it creates another video file and then you supply that to DVDlab, and casagw confirm (and show a way to reproduce the bug) there is a problem with DVDlab that needs to be fixed. Mpeg2vcr can fix the time code and make me create good chapters and only change some information in the already muxed file so is very fast (less than a minute for Scan then Fix GOP timecode and PTS errors, just don't select fix the GOP size errors because that will take-time/add-size breaking all long GOPs (with Dtivo files) into "legal" DVD size ones. With good timecodes (with ReStream or mpeg2vcr) then you don't need force-index option.

At the end I want DVDlab to work right so we don't need mpeg2vcr or ReStream to change the GOP time code in our mpgs created with TyTool.

I'll do more tests to explain in detail the frame-index problem with DVDlab to Oscar so he can fix it, I just need to find out about the "no-counting-every-frame-index" with Dtivo files in DVDlab.

Bato:

I am also using a DTIVO and want to use Mpeg2vcr to make my cuts. The problem I'm having is Mpeg2vcr wants to transcode the 480 x 480 mpg into "legal" DVD size when I try to save the mpg.

The interface works beautifully and makes accurate cuts that play back well within Mpeg2vcr. How can I save the cut/editted mpg without Mpeg2vcr transcoding it to a "legal" resolution? Does it make any difference if I use a .vob source file instead of a .mpg (MPEG2) file?

I've read numerous threads here but have not come up with the solution yet. Any ideas?

TIA!

[bato]

symm: can you post your steps? I edit with TyTool, I think is faster than mpeg2vcr but only GOP accurate.

I used mpeg2vcr with some dish files, they came separate video/audio, so I mux them, mark start and end points with the blue arrows (next to the button with a N), then select scissors with the blue line in the middle, then the commercial is gone, after all commercials are gone I select the start with the arrow, the end with the next arrow and then clic the (red/yellow) icon to save as program stream. This was a couple of months ago, I'll try again because mpeg2vcr never tried to save it as a legal resolution, dish is 544x480 (most).

[symm]

Bato:

Thanks for the response. I edited with the same steps you outlines (mark in and out, scissor to cut between marks) and the resulting clip plays great within mpg2vcr. When I press the save icon, a dialog box appears with program stream selected. The orginal file is Dtivo 480 x 480. Checking the video settings in mpg2vcr it attempts to save with transcode to CBR 15000000 480 x 480, etc. Nowhere can I find a way to tell it to not transcode (which takes longer, takes more storage space and produces an inferior image quality).

I've tried just deleting only at the start and end of the original file so as to have a uncut clip to save and it still wants to transcode. I've tried running the MPEG GOP fixer in mpgvcr first (time code not size) but still no luck.

I've read through the help files, documentation and any online discussions I could find on mpg2vcr editing and I'm still stumped. I considered editing the vob first header to show 720 x 480 since mpg2vcr can edit vob files from tytool, etc. Perhaps this will avoid the transcode/re-encode process and just chunk the uncut video together like tytool does.

I would just use Tytool since it's almost as easy for editing but the GOP index in my Dtivo files are too course to get anywhere near accurate cuts (my indexes must be near 2 seconds). Basically I want the frame accurate cuts and ease of use (in editing) of mpg2vcr but not the transcoding.


TIA!

[TRILIGHT]

I do not know what MPEG2VCR is capable of or not. However, I've always used M2-Edit from Mediaware Solutions and it has been flawless in it's frame accurate MPEG editing. It allows me to mark all of my edit points and then output the resulting video to a single file. Perfect quality (no re-encoding) and no messing with the header (still 480x480).