Video Stream size?

[DelBoy83]

Hi,

I demuxed the raw video stream from a blu-ray folder with eac3to, dgdemux and makemkv. I was expecting the streams to be identical but the demuxed video stream extracted from makemkv was slightly bigger. Could anybody tell me why the stream size could be different?

[Rocky]

I don't really have time right now for a general exposition about it. So much going on and we're behind on some pending issues. It's hard to find time for general stuff not related to a specific problem with a DG tool. If you want to know for a specific case why the demuxed files differ, then please link to a disk that shows that behavior. Or better yet, learn to use a binary compare tool and post the results. Doom9 is generally better for open-ended questions like this.

[DelBoy83]

No problem, I understand your busy atm with other stuff. I just thought I'd ask on the off chance it was something simple. I will try doom9 forum.

[Curly]

Nothing is simple.

You didn't link a disk so how are we supposed to help you?

[DelBoy83]

Hi Curly, I didn't link a disk as its the same with all video streams and not an individual disk.

[flyordie]

So on this topic...
Using eac3to_3.51 and dgdecnv_255
I took a source 901ms video
https://coleman-it.com/dg/mkvDemux/1sec.mkv
dgindex demux gave me
https://coleman-it.com/dg/mkvDemux/1sec.m2v
and eac3to.exe 1sec.mkv 1: 1sec_eac3.m2v gave me
https://coleman-it.com/dg/mkvDemux/1sec_eac3.m2v
using HxD to compare I found these differences. I had to delete each difference in order for future comparisons to happen so the later offsets are minus the removal of the past differences.
Offset
0x01E4D 01 00 00
0x01FCD 01 00 00
0x0214C 00 01 00
0x022CD 01 00 00
0x0244C 00 01 00
0x025CC 00 01 00
0x0274C 00 01 00
0x028CC 00 01 00
0x02A4C 00 01 00
0x02BCC 00 01 00
0x02D4C 00 01 00
0x02ECC 00 01 00
...

Maybe this helps?
¯\_(ツ)_/¯

[Rocky]

Thank you, yes, that helps.

I assume you meant demuxed with DGIndexNV as DGIndex does not do MKV.

It points to an issue. Please be patient as we have a backlog right now after our down time. Gonna catch up as quick as we can.

Setting thread back to unresolved.

[Rocky]

I slipstreamed the fix for this file into DGIndexNV build 255, so please re-download and test.

This fix affects only M2V and VC1 in MKV. I'll be happy to look at any other mismatch cases you may run into. Please post them here.

[Rocky]

@DelBoy83

Does this fix your issue?

[DelBoy83]

Hi Rocky, i've never used DGIndexNV build 255 so i wouldn't know if the fix was actually fixed. The problem i had was when a raw video stream is demuxed with DGdemux, eac3to and MakeMKV all 3 tools gave a different output size. I was just wondering why these would be different sizes if they are all untouched raw streams?

[Rocky]

OK, I can run some tests. Please tell me how you demux with makemkv. I wasn't aware that it could demux.

[DelBoy83]

I just make the mkv with MakeMKV and then use gMKVExtractGUI to extract the raw streams.

[Rocky]

I tested with Dragon_Ball_Z. First I demuxed 00008.mpls with dgdemux to get dgdemux.avc. Then with eac3to:

Code: Select all

D:\tmp\DGDemuxGUI test>eac3to G:\Rips\Dragon_Ball_Z
1) 01005.mpls, 00008.m2ts+00010.m2ts, 0:29:23
   - Chapters, 5 chapters
   - h264/AVC, 1080p24/1.001 (16:9)
   - RAW/PCM, Japanese, stereo, 48kHz
   - RAW/PCM, Undetermined, stereo, 48kHz

2) 00008.mpls, 00008.m2ts, 0:29:23
   - h264/AVC, 1080p24/1.001 (16:9)
   - RAW/PCM, Japanese, stereo, 48kHz
   - RAW/PCM, Undetermined, stereo, 48kHz

3) 01004.mpls, 00007.m2ts+00010.m2ts, 0:19:54
   - Chapters, 5 chapters
   - h264/AVC, 1080p24/1.001 (16:9)
   - RAW/PCM, Japanese, stereo, 48kHz
   - RAW/PCM, French, stereo, 48kHz

4) 00007.mpls, 00007.m2ts, 0:19:54
   - h264/AVC, 1080p24/1.001 (16:9)
   - RAW/PCM, Japanese, stereo, 48kHz
   - RAW/PCM, French, stereo, 48kHz

D:\tmp\DGDemuxGUI test>eac3to G:\Rips\Dragon_Ball_Z 2) 1:eac3to.avc
Running in fast mode
Keeping dialnorm
analyze: 100%
M2TS, 1 video track, 2 audio tracks, 2 subtitle tracks, 0:29:23
1: h264/AVC, 1080p24/1.001 (16:9)
2: RAW/PCM, Japanese, 2.0 channels, 24 bits, 48kHz, dialnorm: 0dB
3: RAW/PCM, Undetermined, 2.0 channels, 24 bits, 48kHz, dialnorm: 0dB
4: Subtitle (PGS), French
5: Subtitle (PGS), French
v01 Extracting video track number 1...
v01 Creating file "d:\tmp\eac3to.avc"...
process: 100%
Video track 1 contains 42264 frames.
eac3to processing took 14 seconds.
Done.

Finally I compared the two demuxed files:

Code: Select all

D:\tmp\DGDemuxGUI test>fc /b dgdemux.avc eac3to.avc
Comparing files dgdemux.avc and eac3to.avc
FC: no differences encountered

If you want me to go any further with this you will need to specify a case that produces a difference. I'll mark this resolved for now and open it again if you show a case with differences.

[DelBoy83]

Hi Rocky, i've just realised that all video steam sizes are exactly the same once the demuxed files have been remuxed with mkvmerge. I took the remuxed files done by mkvmerge and then demuxed into raw streams and all were exactly the same. I just thought i'd let you know.

[Rocky]

Good to hear. Thank you for the update. The demuxed elementary streams should be the same without all the remuxing/demuxing, but if you are happy then so am I.

[flyordie]

Yes you were correct and I meant DGIndexNV.exe
Thank you for the slipstream. I see the only changed file is DGIndexNV.exe in the dgdecnv_255.zip and I now get the same/identical .m2v elementary stream from using DGIndexNV.exe eac3to and mkvextract (via mkvcleaver).

[Rocky]

Good to hear and thank you for the update. Now I have to head over to your other thread.