Dolby Vision now possible through MP4 Mux.

[RESET_9999]

@RESET_9999
Using 3-1 in 1.9.1.5 beta
Input Martian.mov and for cutscenes hdr10plus.json generated in DaVinci

It seems, for some reason, the dovi_tool can't generate a rpu from a Resolve hdr10plus json directly but if you just inject the resolve json to an hevc file and then extract it to a new json file, it works.
So either you do that or export a DV XML instead, or you can use my old RPU scene cuts: https://drive.google.com/file/d/15jug36 ... drive_link

[quietvoid]

It seems, for some reason, the dovi_tool can't generate a rpu from a Resolve hdr10plus json directly but if you just inject the resolve json to an hevc file and then extract it to a new json file, it works.

Does Resolve not do automatic scene cut detection?
Do you have an example file?

[RESET_9999]

It seems, for some reason, the dovi_tool can't generate a rpu from a Resolve hdr10plus json directly but if you just inject the resolve json to an hevc file and then extract it to a new json file, it works.

Does Resolve not do automatic scene cut detection?
Do you have an example file?

Wait, I just tried with a sample (resolve 18.5) and it worked but I'm pretty sure I had a Resolve hdr10plus JSON that I couldn't convert directly to RPU in the past(had to inject/extract again).
He wants to generate HDR10plus in resolve and use the same scene cuts to generate DV with cm_analyze.exe .

@thirdmoon which Resolve version do you use?

[RESET_9999]

Does Resolve not do automatic scene cut detection?
Do you have an example file?

Ok, strangely I can only reproduce with a full movie.
Original resolve json throws an error but if I inject and then extract it, no more error. The hdr10plus_tool json filesize is also much smaller.

files: https://drive.google.com/drive/u/1/fold ... LhERrp1rL0

[thirdmoon]

@thirdmoon which Resolve version do you use?

@RESET_9999 I used 18.5 for HDR10+
I tried generating DV by letting it run for a few min, but DaVinci wouldn't let me export it and gave some error. I ended up generating DV for the entire movie and then used it in 3-1 which worked fine.

But I don't think I can use 3-1 as it takes way too long. I couldn't even finish the full process as it kept getting slower and slower. It was already 4+ hours.
While it takes 30mins in DaVinci.

Comparison of 4K DV vs 1080p + cubic scaling - https://slow.pics/c/kjdYIsBI

I'm pretty happy with the 1080p + cubic scaling result and will stick to it for time being.

[Draco1544]

Hi, how I can bake the FEL into the BL. This is included in the @RESET_9999 script ?

[chronomac]

With option 6-2-7, to create heatmaps and gamut visualizations, is there anyway to have the script read a directory full of screenshots instead of dragging and dropping one at a time?

[RESET_9999]

With option 6-2-7, to create heatmaps and gamut visualizations, is there anyway to have the script read a directory full of screenshots instead of dragging and dropping one at a time?

Workflow 6-8-4. Input a mkv file and the script will automatically export heatmaps and gamut png for 50 frames. You can change the number of frames to export at line 76 in the latest beta.
set frame_number=50

Hi, how I can bake the FEL into the BL. This is included in the @RESET_9999 script ?

6-9-2 in the latest beta

[chronomac]

Workflow 6-8-4. Input a mkv file and the script will automatically export heatmaps and gamut png for 50 frames. You can change the number of frames to export at line 76 in the latest beta.
set frame_number=50

What's the maximum number you can use?

[RESET_9999]

What's the maximum number you can use?

the script exports a screenshot every 2000 frames starting at frame 1000. (1000-3000-5000-7000-9000-11000 etc..)
So it depends on your input length.

you could remove a 0 in %%i000 around line 11495 and that would allow a lot more screenshots to be extracted.

Code: Select all

for /L %%i in (1,2,%frame_number%) do (
  echo LoadPlugin("%ffms2%"^) > "%TEMP%%%i.avs"
  echo FFVideoSource("%filepath%%filename%%fileext%", cachefile="%TEMP%1.ffindex"^) >> "%TEMP%%%i.avs" 
  echo scriptClip(""^"  >> "%TEMP%%%i.avs" 
  echo     subtitle("%filename%", size=14, align=7, 4, 25, text_color=$606060^) >> "%TEMP%%%i.avs" 
  echo     subtitle("Resolution: " + string(width^) + " x " + string(height^) + " @ " + string(framerate^), size=14, align=7, 4, 40, text_color=$606060^) >> "%TEMP%%%i.avs" 
  echo     subtitle("Frame: %%i000", size=14, align=7, 4, 55, text_color=$606060^) >> "%TEMP%%%i.avs" 
  echo     subtitle("Picture type: " + Chr(FFPICT_TYPE^), size=14, align=7, 4, 70, text_color=$606060^) >> "%TEMP%%%i.avs"
  echo ""^"^) >> "%TEMP%%%i.avs"
  echo trim(%%i000,%%i000^) >> "%TEMP%%%i.avs"
)

change it to

Code: Select all

for /L %%i in (1,2,%frame_number%) do (
  echo LoadPlugin("%ffms2%"^) > "%TEMP%%%i.avs"
  echo FFVideoSource("%filepath%%filename%%fileext%", cachefile="%TEMP%1.ffindex"^) >> "%TEMP%%%i.avs" 
  echo scriptClip(""^"  >> "%TEMP%%%i.avs" 
  echo     subtitle("%filename%", size=14, align=7, 4, 25, text_color=$606060^) >> "%TEMP%%%i.avs" 
  echo     subtitle("Resolution: " + string(width^) + " x " + string(height^) + " @ " + string(framerate^), size=14, align=7, 4, 40, text_color=$606060^) >> "%TEMP%%%i.avs" 
  echo     subtitle("Frame: %%i00", size=14, align=7, 4, 55, text_color=$606060^) >> "%TEMP%%%i.avs" 
  echo     subtitle("Picture type: " + Chr(FFPICT_TYPE^), size=14, align=7, 4, 70, text_color=$606060^) >> "%TEMP%%%i.avs"
  echo ""^"^) >> "%TEMP%%%i.avs"
  echo trim(%%i00,%%i00^) >> "%TEMP%%%i.avs"
)

Also line 11528 and 11626:
echo noms_images = ['%filename%_Frame_%%i00.tiff'] >> "%TEMP%gamut_map.py"

echo noms_images = ['%filename%_Frame_%%i00.tiff'] >> "%TEMP%heat_map.py"

[Draco1544]

How I can check if my mkv is a FEL dual layer or single layer ?

[RESET_9999]

How I can check if my mkv is a FEL dual layer or single layer ?

workflow (2-3)

FYI, there's a small explanation of all the workflows in the TOOLS_READ_ME.txt

[quietvoid]

It seems, for some reason, the dovi_tool can't generate a rpu from a Resolve hdr10plus json directly but if you just inject the resolve json to an hevc file and then extract it to a new json file, it works.

Resolve generates HDR10+ metadata with SceneFirstFrameIndex starting at 86400, while it's currently expected to start at 0.
I guess maybe a solution is to offset the list of the first frame indices by whatever is the first frame index.

Should be fixed in next version, which is very much due.
Maybe by this weekend.

[thirdmoon]

@RESET_9999
How do I create screenshot comparison between HDR.mkv and HDR-DVinjected.mkv to see the difference between HDR and DV.
Thank you.

[RESET_9999]

Resolve generates HDR10+ metadata with SceneFirstFrameIndex starting at 86400, while it's currently expected to start at 0.
I guess maybe a solution is to offset the list of the first frame indices by whatever is the first frame index.

Should be fixed in next version, which is very much due.
Maybe by this weekend.

thank you

@RESET_9999
How do I create screenshot comparison between HDR.mkv and HDR-DVinjected.mkv to see the difference between HDR and DV.
Thank you.

AFAIK, you can't.

The way I did the HDR/DV comparisons below is by capturing the LLDV signal of my bluray player using an HDMI capture card. And different EDIDs will produce different outcomes.
https://mega.nz/folder/1C1gFLYD#lKYXNitCf0Tzg_nf5vhN9g