mike admin wrote: ↑Wed Mar 11, 2020 6:11 pm
Of course, EL is always 12 bit on blu-ray. The question is what exactly is in these 2 extra bits. My hypothesis - in MEL stream all pictures in EL are filled with the same value (zeroes). This is easy to verify though - one just has to decode entire EL from any MEL disc (Joker) and confirm that ALL frames in EL are indeed dark zero pictures. From the sizes of slice NALU it looks that way.
I believe you are right.
I have rendered every single frame of a FEL (Terminator) and an MEL (Joker) stream.
It was clear that there were differences between the FEL frames.
The MEL frames are instead all the same (I also tested another MEL layer and after rendering it I get the same MD5 hash).
However, I believe it is appropriate to keep it.
If you notice, several users have reported that the Shield was able to play my files with profile dvhe.07.
There seems to be a fallback mechanism that allows you to take advantage of the profile best suited to the situation (in this case dvhe.04 given the presence of BL, EL and RPU): are we sure that, by removing the information relating to the EL layer, the hardware decoder is not initialized with the specifications of the profile dvhe.05 (BL + RPU)?
In this case, the decoder would expect a 12bit stream with IPTPQc2/IPT proprietary color space (instead of 10bit YCbCr).
shawnc22 wrote: ↑Thu Mar 12, 2020 1:18 am
... I originally tried to combine the two original HEVC (BL+EL) files using v2 of yusecope's tool, but it kept hanging on me after dropping in the files, so I just used the one he had already created in the post above...
I can assure you that the tool works: the batch file (the one with the .bat extension) starts the program expecting to work with huge files (and not with 90MB files). To combine the two example HEVC raw streams you have to launch my tool using the command line:
Code: Select all
python-3.7.6.amd64\python src\bl_el_rpu_builder.py -mode 1 -lbf 1000 -bl \path\to\Joker_BL.hevc -el \path\to\Joker_EL.hevc -of \path\to\output_file.hevc
mattmarsden wrote: ↑Wed Mar 11, 2020 4:59 pm
Arbigator wrote: ↑Wed Mar 11, 2020 12:07 am
shawnc22 wrote: ↑Tue Mar 10, 2020 11:40 pm
I was going to give that a try, too, when I got home, but I can try First Man also. I'm wondering if it might be because the Doctor Sleep disc is also in HDR10+. The EL might be a little different in that case?
That's interesting, I didn't realize it's in HDR10+, though that explains the descriptor in mediainfo. The file did not give any errors in the tool, and it did finish to 100%.
Same with Power Rangers
Whoever had the "EL ANALYSIS: 0.0%" problem could kindly upload a piece of the two original layers
Code: Select all
ffmpeg -i \path\to\bluray\folder\or\m2ts_file -map 0:0 -ss 0 -t 20 -c copy \path\to\BL.hevc
ffmpeg -i \path\to\bluray\folder\or\m2ts_file -map 0:1 -ss 0 -t 20 -c copy \path\to\EL.hevc