BD Rebuilder Beta - Bug Reports Only

[thegame]

Originally Posted by jdobbs

Unfortunately the real stumbling block for 3D backups right now is a method for writing them. On a 3D disc there are blocks that are referenced by more than one file. Right now there is no popular burning software (that I know of) that would support that. Even if it did, a calling program would have to give a lot of information to the software to tell it which parts of which files are to be shared.

The only current way to get around that is to have the blocks duplicated -- which would be exceptionally inefficient.

Thanks, I assumed it couldn't be done, but I thought I would ask.

[bassnut]

I Am Number Four.

Extras are all encoded twiced as previously posted thopugh some do not have the directors comentary defore the clip. All play fine and I didn't have any freezes as Cap's did. Should not that this is only evident when you click on play all if you select the individual Deleted Sceen it just plays once.

Originally Posted by jdobbs

On a 3D disc there are blocks that are referenced by more than one file. Right now there is no popular burning software (that I know of) that would support that. Even if it did, a calling program would have to give a lot of information to the software to tell it which parts of which files are to be shared.

The only current way to get around that is to have the blocks duplicated -- which would be exceptionally inefficient.

You got me thinking. I've mentioned in another thread my opinions & theories on how it works, but I now think it's more likely the .ssif files are the real files, and the .m2ts files are the virtual files. It makes more sense if you think about how the drive & laser read the disc; It would be much more efficient and not wear out the motor as much, and also be faster, to have the .ssif as the non-virtual file.

[k-c-ksum]

this looks interesting

http://www.sisveltechnology.com/3D_tile_format.asp

720p 3D from a 1920x1080.

Your 3DTV would need to support it but it would get around the need for MVC and siff file creation. AVISYNTH could take care of the tiling

how much res could be created if the source material is 2:40.1 as the black space wouldnt need to be encoded? Something to look at a bit furher down the road prehaps?

[setarip_old]

@jdobbs, @thegame

For what MAY prove to be a pleasant surprise, try specifically using MakeMKV to rip "Dinosaurs Alive 3D" and then perform a FULL DISC backup with BD Rebuilder.

As opposed to using AnyDVD, which copies the .SSIF files at their full 15.2Gb combined filesizes (apparently essentially duplicates of most of the .M2TS files), it appears that MakeMKV is able to extract the pertinent data (likely primarily pointers) from the .SSIF files - and save it as a miniscule total of 56.2Kb in five "SSIF.MAP" files.

File 00008.M2TS contains the 2D full movie (8.86Gb), while file 00009.M2TS (4.67Gb) apparently contains the required additional information for 3D display. I'd speculate that 00009.M2TS likely has to remain uncompressed - and it does when using BD-RB . (I say "I'd speculate" because I do not presently have the required hardware to playback "new 3D" discs)...

Note: After compressing with BD Rebuilder, you'll have to manually copy the five tiny ".SSIF.MAP" files into the "SSIF" sub-folder.

[jdobbs]

Originally Posted by Capsbackup

I have found a strange occurance doing a full movie backup to BD9 of the movie I AM Number Four. The deleted scenes are listed as 18:58 on the original disc. On the BD-RBV03803 backup, the same deleted scenes are listed as 30:39.
These scenes are 1080P, 23.976fps. There are numerous duplicate .m2ts files within the same playlist 00329.mpls.
Could this be the problem, that BD-RB is not making the necessary reference corrections?

I ran this one and was able to repeat it. It was caused by the fact that some CLPI/M2TS pairs were referenced more than once within the same MPLS and one of them was for partial playback. A fix I had added into v0.34.06 (to correct a TSMUXER end-time glitch and resulting "stutter") needed to be tweaked to consider the possibility of multiple-reference/partial playback.

I've fixed it for the next release.

[Capsbackup]

Originally Posted by jdobbs

I ran this one and was able to repeat it. It was caused by the fact that some CLPI/M2TS pairs were referenced more than once within the same MPLS and one of them was for partial playback. A fix I had added into v0.34.06 (to correct a TSMUXER end-time glitch and resulting "stutter") needed to be tweaked to consider the possibility of multiple-reference/partial playback.

I've fixed it for the next release.

Thanks again! I am impressed with your continued vigilance.

[Acerjen]

Jdobbs,

I have a question about DTS Express and how it differs from
AC3 Plus.

I have backed up a few movies that use AC3 Plus instead of DTS Express for the PIP track and it works fine in the final BD-25 disc. Everything is in sync. I assume this works because TSMUXER supports AC3 Plus but not DTS Express? So, I was curious what the difference is. I was also wondering if it would be possible to convert DTS Express tracks to AC3 plus and then use TSMUXER to put it back together? I was curious and thought maybe you would know the answer.

I would also like to thank you for such an awesome program and all the time you put into it. Thanks.

Later,
Acerjen

[PurpleMan]

Originally Posted by jdobbs

Unfortunately the real stumbling block for 3D backups right now is a method for writing them. On a 3D disc there are blocks that are referenced by more than one file. Right now there is no popular burning software (that I know of) that would support that. Even if it did, a calling program would have to give a lot of information to the software to tell it which parts of which files are to be shared.

The only current way to get around that is to have the blocks duplicated -- which would be exceptionally inefficient.

Actually, wouldn't it be sufficient to have a program that can write such ISO files? (then use any burning software to write that ISO)

[jdobbs]

Originally Posted by PurpleMan

Actually, wouldn't it be sufficient to have a program that can write such ISO files? (then use any burning software to write that ISO)

Yes, but typically the ISO is created by the same (or similar) program that writes it.

[jdobbs]

Originally Posted by Acerjen

Jdobbs,

I have a question about DTS Express and how it differs from
AC3 Plus.

I have backed up a few movies that use AC3 Plus instead of DTS Express for the PIP track and it works fine in the final BD-25 disc. Everything is in sync. I assume this works because TSMUXER supports AC3 Plus but not DTS Express? So, I was curious what the difference is. I was also wondering if it would be possible to convert DTS Express tracks to AC3 plus and then use TSMUXER to put it back together? I was curious and thought maybe you would know the answer.

I would also like to thank you for such an awesome program and all the time you put into it. Thanks.

Later,
Acerjen

It would work fine to convert DTS Express to AC3 Plus. But TSMUXER can't demux DTS Express, and I don't know of a freeware CODEC that will play it back (so you can reencode it).

[drmih]

Originally Posted by Acerjen

Jdobbs,

I have a question about DTS Express and how it differs from
AC3 Plus.

I have backed up a few movies that use AC3 Plus instead of DTS Express for the PIP track and it works fine in the final BD-25 disc. Everything is in sync. I assume this works because TSMUXER supports AC3 Plus but not DTS Express? So, I was curious what the difference is. I was also wondering if it would be possible to convert DTS Express tracks to AC3 plus and then use TSMUXER to put it back together? I was curious and thought maybe you would know the answer.

I would also like to thank you for such an awesome program and all the time you put into it. Thanks.

Later,
Acerjen

This is one of those things which keeps cropping up and you revisit it and then think there may be an easy solution, only to find there's always one link missing. I understand that eac3to will read / demux DTS-E and E-AC3, but I've never seen an ability to encode to E-AC3. eac3to picks up commercial codecs (which may be on your machine if you've installed TMT or Power DVD) but I can find no mention of, for instance, demuxing DTS-E and encoding it to E-AC3 - it seems to have to be to a format such as ac3, wav etc. Given the number of great brains who would have their eye out for a solution to the PiP issue, I guess it's not possible, rather than nobody's tried. With TsMuxer frozen, maybe the author of eac3to might be able to do what you suggest, as I think the remux of DTS-E might be too problematic.

[shon3i]

eac3to (in order is much better than tsmuxer for demuxing, and safer for streams that will be direct stream copyed) can demux/decode dts-e IIRC and here is freeware e-ac3 encoder (aften extension/patch from developer) http://forum.doom9.org/showthread.php?t=161383, maybe is not best but will satisfy needs here (i hope that secondary audio not need core track as primary e-ac3 does, i am not sure but i can check), it probably will create same quality like ac3 at same bitrate, but that is not bad at all for first time

[drmih]

Thanks - sounds interesting. Currently if I process an m2ts file with dts-express I can use the .dts setting within eac3to. Does eac3to automatically pick up aften, and thus should you be able to just put a suitable suffix in the command line for it automatically encode to e-ac3?

**As an initial test I have extracted a dts-express PiP audio track and I got a perfect ac3 files using Aften, therefore if the beta version of Aften does make a E-AC3 file, can this be remuxed back into the m2ts file to test?

eac3to p:\LET_ME_IN_2010\BMDV\STREAM\00001.M2TS 6: audio.ac3
DTS Express, 2.0 channels, 1:55:48, 24 bits, 192kbps, 48kHz
Decoding with ArcSoft DTS Decoder...
Encoding AC3 <448kbps> with libAften...
Creating file "audio.ac3"...
eac3to processing took 2 minutes, 42 seconds.
Done.

[shon3i]

Originally Posted by drmih

Thanks - sounds interesting. Currently if I process an m2ts file with dts-express I can use the .dts setting within eac3to. Does eac3to automatically pick up aften, and thus should you be able to just put a suitable suffix in the command line for it automatically encode to e-ac3?

No there is no compiled bulds of aften that support eac3 (i didn't see) and eac3to uses aftenlib, so no eac3 encoding inside eac3to at least not yet. But still i am not sure (someone correct me if i wrong) and don't know that eac3to can decode dtse or just can demux it.

eac3 encoding will be now easy, but still need to check that secondary audio need core (normal ac3), that will broke completely idea and back to beginning

[jdobbs]

My plan has always been to manually pull the DTS-E audio packets from the source M2TS, manipulate the timestamps appropriately, and reinsert it in the output stream. It's already very efficiently encoded, so I see no need to go through all the decoding/reencoding. I wrote the code to do the extract/insert, but haven't been able to make it work without causing "stutter" in the output, so I haven't released it for any testing outside myself. I'm not sure yet what I'm doing wrong.

[drmih]

Just out of interest, is the stuttering limited to the secondary audio, or does it impact the video? Is it possible that the stuttering is player dependent?

[shon3i]

Originally Posted by jdobbs

My plan has always been to manually pull the DTS-E audio packets from the source M2TS, manipulate the timestamps appropriately, and reinsert it in the output stream. It's already very efficiently encoded, so I see no need to go through all the decoding/reencoding. I wrote the code to do the extract/insert, but haven't been able to make it work without causing "stutter" in the output, so I haven't released it for any testing outside myself. I'm not sure yet what I'm doing wrong.

I agree, but from this point it's completely irrelevant since quality will not drop (reencoding will not affect that much since bitrate can be up 256kbps and eac3 encoding is efficient as much ac3 is), and it's secondary audio, who cares anyway. Reinserting is not good idea (there is ton of PES/DTS extensions,flags, ...) and tsmuxer support eac3, but you maybe success, you may need more samples ?

[jdobbs]

Originally Posted by drmih

Just out of interest, is the stuttering limited to the secondary audio, or does it impact the video? Is it possible that the stuttering is player dependent?

It's everything.

No. I have four different players that I use for testing, two Sonys and two Samsungs. It does play ok on the computer -- but that's always a bad test.

[jdobbs]

Originally Posted by shon3i

I agree, but from this point it's completely irrelevant since quality will not drop (reencoding will not affect that much since bitrate can be up 256kbps and eac3 encoding is efficient as much ac3 is), and it's secondary audio, who cares anyway. Reinserting is not good idea (there is ton of PES/DTS extensions,flags, ...) and tsmuxer support eac3, but you maybe success, you may need more samples ?

I'm very aware of all the timestamps -- I manipulate them already in other areas. There are times when you just can't use the timestamps created by TSMUXER (especially in menus or sources with out-of-mux substreams) and you have to set the stream's timestamps back to match the original.