DGDemux development
Re: DGDemux development
OK, thanks. Standby for new version...
Re: DGDemux development
Just tried with a lower directory (T:\New Folder\) and got the same error
Re: DGDemux development
Please re-download and try again. Thank you.
Re: DGDemux development
It's working on the DaVinci Code
I'll try it on Inferno, the other movie, in a sec
I'll try it on Inferno, the other movie, in a sec
Re: DGDemux development
Great, thanks. I had my define for the maximum THD frame size too low. I should malloc/free it instead of having a fixed buffer size. Glad you found that directory bug before I released that!
Re: DGDemux development
Working fine on the second movie
Looks like its all good
Looks like its all good
Re: DGDemux development
I'll make releases tomorrow morning. Thank you for your testing, gonca.
Re: DGDemux development
Have a good night
Re: DGDemux development
Thank you, you too!
Re: DGDemux development
We're back! We met Rosalind Franklin. She had so many insights on "modern physics" and the sociology of science. And she went to Cambridge, too.
Anyway, I thought up a real simple, robust way to do episode demuxing without all kinds of hullabaloo. Everything falls out automatically. OK, first implement a 'skip first n M2TS files' option. Also implement stopping the demux when a new M2TS (or EOS) is asked for, when in episode mode. Then a high-level process simply starts and when the first M2TS hits NextFile() it re-spawns DGDemux.exe with skipping 1 M2TS and then the current instance dies. Then at each NextFile() call, just bump the number to skip by one, re-spawn, and die. Performance should be fine because each stage only needs to parse one M2TS file. Pretty sneaky, huh? The alternative of closing and re-opening files on-the-fly while parsing all the M2TS files straight-through is too nightmarish to contemplate.
What do you think?
Anyway, I thought up a real simple, robust way to do episode demuxing without all kinds of hullabaloo. Everything falls out automatically. OK, first implement a 'skip first n M2TS files' option. Also implement stopping the demux when a new M2TS (or EOS) is asked for, when in episode mode. Then a high-level process simply starts and when the first M2TS hits NextFile() it re-spawns DGDemux.exe with skipping 1 M2TS and then the current instance dies. Then at each NextFile() call, just bump the number to skip by one, re-spawn, and die. Performance should be fine because each stage only needs to parse one M2TS file. Pretty sneaky, huh? The alternative of closing and re-opening files on-the-fly while parsing all the M2TS files straight-through is too nightmarish to contemplate.
What do you think?
Sherman Peabody
Director of Linux Development
Director of Linux Development
- Bullwinkle
- Posts: 338
- Joined: Thu Sep 05, 2019 6:37 pm
Re: DGDemux development
Sherman, that's pretty good for a kid. However, coordinating that process with the GUI may not be so trivial. I expect you to have it coded by the end of the day tomorrow. That is the real test of a Moose.
Re: DGDemux development
Sherman is like wheee!
Re: DGDemux development
Guys, I have the first phase completed. I added the 'skip n M2TS files' option. It uses an extra episode number on the file names and adjusts chapter times, etc. Seems to work fine. Now I will write the high-level controller that does all the episodes via:
DGDemux ... -sk n
starting at n = 0 and incrementing it until done.
DGDemux ... -sk n
starting at n = 0 and incrementing it until done.
Sherman Peabody
Director of Linux Development
Director of Linux Development
Re: DGDemux development
I got the high-level process working. You just add -ep to the end of the DGDemux command line and that will enable per-episode demuxing. Now, I have to add the -ep option to DGDemuxGUI. Then regression test everything very thoroughly and make sure non-episode mode is still working properly.
Sherman Peabody
Director of Linux Development
Director of Linux Development
Re: DGDemux development
Now I have DGDemuxGUI working fine. There is one wart. For each M2TS the progress goes 0-100 and then after the last one everything finishes. It's hard to make the progress span the entire M2TS list so I'm not going to bother.
Now code review and regression testing.
Now code review and regression testing.
Sherman Peabody
Director of Linux Development
Director of Linux Development
- Bullwinkle
- Posts: 338
- Joined: Thu Sep 05, 2019 6:37 pm
Re: DGDemux development
You better hurry up, Sherman. There are only about 13 hours left in the day.
Re: DGDemux development
Sherman has turned it over to me for release, so he made the schedule. Great job!
Probably make slipstream tomorrow.
Probably make slipstream tomorrow.
- Bullwinkle
- Posts: 338
- Joined: Thu Sep 05, 2019 6:37 pm
Re: DGDemux development
Not so fast, Sherman. There is no proper error checking and recovery. And the progress doesn't span the entire playlist? You have all the in and out times for all the M2TS. Figure it out.
Re: DGDemux development
Thank you for the guidance, Bullwinkle. I have added error handling and fixed up the progress indication. Handing off to Rocky.
Sherman Peabody
Director of Linux Development
Director of Linux Development
Re: DGDemux development
Thank you, guys. The changes are non-trivial so let's go with a test version for people to bang on. Note that it has been tested only with CC_SAKURA.
http://rationalqm.us/dgdemux/binaries/D ... isodes.rar
Update both DGDemux and DGDemuxGUI. The manual explains the new options -ep and -sk. Feedback will be appreciated.
http://rationalqm.us/dgdemux/binaries/D ... isodes.rar
Update both DGDemux and DGDemuxGUI. The manual explains the new options -ep and -sk. Feedback will be appreciated.
Re: DGDemux development
Quick test on Stargate Atlantis disc
Works fine, chapters are right and everything else seems ok
Works fine, chapters are right and everything else seems ok
Re: DGDemux development
Great to hear, gonca. Thank you for your testing.
Re: DGDemux development
Released as slipstream 28. Thank you everybody for suggestions, coding, and testing.
What should we ask Sherman to do next? Solve the quantum measurement "problem"? After he saves history, of course.
What should we ask Sherman to do next? Solve the quantum measurement "problem"? After he saves history, of course.
Re: DGDemux development
Don't worry, Rocky, I can multitask.
Sherman Peabody
Director of Linux Development
Director of Linux Development
Re: DGDemux development
Beautiful.
Made a 9-angle-BD test project in DVD Architect 7.0 just to see if it would work:
All 9 angles are found nicely (but did not test demuxing for now).
Small suggestion: Once it comes to cosmetics, the second video stream can have its PID.
Found 0x1015 (dec 4117) for Dolby Vision Enhancement Layer (UHD 2001 - A Space Odyssey)
Somewhere I happened to come across 0x1012 (dec4114) for a MVC stream, don't remember where...
Made a 9-angle-BD test project in DVD Architect 7.0 just to see if it would work:
All 9 angles are found nicely (but did not test demuxing for now).
Small suggestion: Once it comes to cosmetics, the second video stream can have its PID.
Found 0x1015 (dec 4117) for Dolby Vision Enhancement Layer (UHD 2001 - A Space Odyssey)
Somewhere I happened to come across 0x1012 (dec4114) for a MVC stream, don't remember where...