[DoVi_Scripts] Multi-Function Scripts for Dolby Vision processing and a lot more...

[en6ads]

I've noticed your script changes source file names sometimes.

For example, I did a 7-2 screenshot with these filenames:

Dune (1984) - 2160p-ESC 2025 France Release-DV P8.1 Baked FEL Encode CRF15 Slow-1000nits MDL.mkv

and

Dune (1984) - 2160p-ESC 2025 France Release-DV 7 FEL.mkv


And after completing the screenshots it had changed the names of the files to these:
Dune (19842 - 2160p-ESC 2025 France Release-DV P8.1 Baked FEL Encode CRF15 Slow-1000nits MDL.mkv

Dune (19841 - 2160p-ESC 2025 France Release-DV 7 FEL.mkv

This has happened before. Any idea what's happening?

[Kuler087]

Yes, the script removes the “)” character from the input filename in some workflows. It caused issues in certain parts of the workflow, and I just never bothered changing the behavior since I never use special characters in my paths or filenames.

[en6ads]

OK, thanks.

All my files include the year in parenthesis to match TMDB.

Yes, the script removes the “)” character from the input filename in some workflows. It caused issues in certain parts of the workflow, and I just never bothered changing the behavior since I never use special characters in my paths or filenames.

[Hellboy.]

I'm using Windows 11. I let the process running in the night and in the morning when i checked it, it was stuck. I did a test with a smaller file and the process finish 100%. Maybe the file is corrupted or my PC has a problem.

[coopzr]

@Kuler087
I tried bake FEL for the first time and the process got stuck. What could be the problem?
I'm using Windows 11. I let the process running in the night and in the morning when i checked it, it was stuck. I did a test with a smaller file and the process finish 100%. Maybe the file is corrupted or my PC has a problem.

Try grabbing the latest 265.exe (patman mod) here (https://github.com/Patman86/x265-Mod-by-Patman/releases). I recall a recent NVIDIA driver update caused cm_analyze to hang. Dolby quickly fixed this with an update. Perhaps the same thing is happening here?

[Hellboy.]

Thanks. Going to try it.

[wyup]

What are the differences between madVR, Resolve and Dolby tool for scene cut detection?
Which is better, more dependable and realistic for scene cuts?

MadVR scene detection is used for plotting.
Resolve use its own neural scene cut detection (best?)
Dolby generates it's own scene cuts for it's plotting and auto trims.

When feeding a external scene cut list such as for av1 encoder (SvtAv1EncApp), which scene cut list would be the most realistic for keyframes?

[Kuler087]

CM_analyze has a scene-cut detection algorithm, but it’s only used for frame-by-frame analysis. It can’t perform shot-by-shot analysis without an external shot list, which is why madVR generates the scene cuts for workflow 3-1.
Resolve has much better scene-cut detection than madVR. madVR is extremely sensitive to fast movement and flashes, and sometimes even misses very obvious scene cuts.

You can use workflow 2-8 to export the scene-cut differences between two sources, and then compare them with an indexed video (9-1).

Dolby generates it's own scene cuts for it's plotting and auto trims.
It doesn’t generate cuts; the DV Plotter workflows(6-3 / 6-4) only plots existing metadata along with the existing scene cuts.

[wyup]

How does the Dolby tool generate auto trims if it doesn't set scene changes? It's supposed to be a self-contained solution, without requiring an external shot list. Dolby Vision is known to adjust its tonemapping by scenes, not by individual frames.
When I said Dolby generates its own scene cuts I didn't mean that it exports them, but that at least it generates them.
It doesn’t generate cuts; the DV Plotter workflows(6-3 / 6-4) only plots existing metadata along with the existing scene cuts.
Where do 6-3 and 6-4 (Measure DoVi Levels) take the 'existing scene cuts' then if the DV plotter only plots existing metadata? Do these workflows also use MadVR scene detection?

[Kuler087]

As I said, the Dolby Vision algorithm alone cannot create shot-by-shot metadata, it requires an external shot list text file in cm_analyze, or an editing software like Resolve to auto-generate the cuts, or the colorist to do them manually.
When you use cm_analyze without providing a shot list text file, the algorithm produces frame-by-frame metadata with L2 trims using its internal scene-cut detection (I dont know how that works). The generated trims in L2 are just a workaround to mitigate the cmv2.9 bugs.

Where do 6-3 and 6-4 (Measure DoVi Levels) take the 'existing scene cuts' then if the DV plotter only plots existing metadata? Do these workflows also use MadVR scene detection?

all the existing DV movies have scene cuts within the metadata. Madvr is not used in 6-3/6-4

[wyup]

When you use cm_analyze without providing a shot list text file, the algorithm produces frame-by-frame metadata with L2 trims using its internal scene-cut detection (I dont know how that works).
Ok. Allow me to ask, then, what purpose does cm_analize internal scene-cut detection algorithm serve if it produces just frame-by-frame metadata, when not providing a external shot list file.

[Kuler087]

Because that’s how it works, the frame-by-frame algorithm is a different type of DV analysis.
It used to output the same scene cuts as the frame count, but this was changed in the recent CM version. The frame-by-frame metadata now has scene cuts similar to Resolve and madVR, but the metadata still change frame by frame.
Resolve still outputs the same scene cuts as framecount when you export frame-by-frame profile 5 or 8 HEVC.


https://i.ibb.co/TqxwJCBv/explorer-8k-ET5x-XFk6.png

https://i.ibb.co/2Y7bVZ9t/l1.gif

https://i.ibb.co/6cYJyYr8/l2.gif

[MwenDavo]

Hi guys. First of all thanks for making such an awesome tool. I'm looking into making a Linux compatible version of the 8-2-1 flow to convert DoVi P7 into P8.1 (only the video stream, the idea is to be able to automate that and integrate it with other transcoding software). After reading through the script, I'm trying to understand the steps it takes to do this. Currently, I have the following:
- The script separates the BL, EL and RPU from the video stream of the file
- It then checks if EL is either FEL or MEL. In case of FEL it bakes it into the BL, and in case of MEL it discards it
- After that, it combines the baked BL (or untouched BL if MEL was found) with the RPU into a new P8.1 HEVC stream, which then gets muxed with the rest of the original file's streams.

I've seen the script use AviSynth+, dovi_baker, dovi_tool, ffms2 and FFmpegin various steps, though I'm not entirely sure what each one does on its respective part. If someone could explain this I'd be thankful.
Lastly, I've researched into some of the tools that I mentioned earlier, specifically if they have Linux support, and seems that most of them have, so it shouldn't be an issue.

[Kuler087]

that sounds right.

ffmpeg or mkvextract is for demuxing HEVC from the container
dovi_tool is for demuxing the layers, extracting & converting the RPU to P8 and then injecting into the encoded HEVC.
ffms2 or dgindex are the frame indexer
avisynth+ is the whole script to be encoded
x265 is for encoding
mkvtoolnix(mkvmerge) is for muxing to mkv

[alexmorph3us]

@Kuler087: I'm new to this. I have ripped the movie Titanic for which I know that it has an FEL layer because I used Dovi_tool for demuxing it, but it doesn't have the option for baking FEL into the BL in order to use MEGUI for x265 encoding after that. Can you please give me an Avisynth helper script for DoviBaker or at least tell me what to change in this script that I have created using MEGUI?

LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\DGDecodeNV.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\fft3dfilter.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\dfttest.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\masktools2.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\mvtools2.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\nnedi3.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\RgTools.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\avsresize.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\DGTonemap.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\DePan.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\DePanEstimate.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\HDRTools.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\DoViBaker_x64.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\avscube_x64.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\avisynth_plugin\MapNLQ.dll")
LoadPlugin("C:\Users\ALEX\OneDrive\Documents\MeGUI x64\tools\lsmash\LSMASHSource.dll")
LWLibavVideoSource("C:\Users\ALEX\OneDrive\Documents\Titanic - A Cold Breakfast Chapter (Original) [BD]\Titanic - A Cold Breakfast Chapter (Original) [BD].mkv", format="YUV420P8")
#deinterlace
crop(0, 264, 0, -264)
Spline64Resize(1920,816) # Spline64 (Sharp)
#denoise
ConvertBits(16)
z_ConvertFormat(pixel_type="YUV444P16",dither_type="none")
ConvertYUVtoXYZ(Color=0, HDRMode=0, OOTF=false, OutputMode=2)
ConverXYZ_BT2446_C_HDRtoSDR(PQMode=true, Lhdr=10000, Lsdr=100.0, pColor=0)
ConvertXYZtoYUV(Color=2, pColor=0, OOTF=false, OutputMode=2)
ConvertBits(10)
ConvertBits(16)
ConvertToYUV420()

I would also like to change the tone mapping algorithm in the script so that I can use libplacebo because I heard that it does a better job than what I have put in there.

Any help will be appreciated! Thank you.

[Kuler087]

Just use dovi_scripts workflow 8-2-1 (HDR) or 8-2-2(SDR), cancel the encode when it starts, and go into the temp folder to see what the AVS script looks like.

[en6ads]

An update:
I have DDE now, and converting TrueHD to DDP with Atmos is working well with Deezy (https://github.com/jessielw/DeeZy). It uses TrueHDD (https://github.com/truehdd/truehdd) to create an atmos master audio file from a trueHD atmos movie, then encodes to DDP atmos. I can set bed conformance, get a 1536kbps 7.1 atmos, etc. I check generated file with Cavernize (https://cavern.sbence.hu/cavern/doc.php?p=Cavernize) and it appears to be real atmos (beds and objects).

A feature request: Could you please add Deezy as part of your scripts? So those with DEE can generate DDP Atmos as part of the broader workflows with Dovi Scripts.

Many thanks.

I've been playing with Cavernize and the newly released TrueHDD (https://github.com/truehdd/truehdd). It can take a TrueHD Atmos track from a UHD and convert it to Dolby Atmos Master Format (DAMF). This can be read by Davinci Resolve and also the free Dolby Atmos Conversion Tool - however this only converts to either: .atmos, .rpl, .wav(ADM BWF), or .mxf(IMF IAB).

So we can now edit Atmos tracks from the TrueHD Atmos while preserving the Atmos.

The last step I need to find is how to convert DAMF to DDP Atmos without having to pay the $400 per year for Dolby Encoder Engine. One can hope a free solution can be reverse engineered soon. Or do you know of a way?
Thank you.

[Kuler087]

I’d have to find out how it handles DEE’s inverted channel encoding. Right now, the script can encode DDP (without Atmos) using DEE, and I use FFmpeg to correct the channel order when encoding to WAV PCM as a workaround.
I’m not sure how I could do that with the Atmos master format, because in my limited testing with DDP-Atmos through DEE, the channels were also inverted.

[en6ads]

OK I appreciate you looking into it. Thank you.
I’d have to find out how it handles DEE’s inverted channel encoding. Right now, the script can encode DDP (without Atmos) using DEE, and I use FFmpeg to correct the channel order when encoding to WAV PCM as a workaround.
I’m not sure how I could do that with the Atmos master format, because in my limited testing with DDP-Atmos through DEE, the channels were also inverted.

[Kuler087]

OK I appreciate you looking into it. Thank you.

Ok please try this version. Workflow 8-1-3. (you must put the truehdd.exe in the tools folder)
https://drive.google.com/file/d/128gq8aDUTKA_aT7SQsM9dkjA1EP1sosR/view?usp=drive_link

The default mode is 5.1 Atmos at 1024 kbps, but you can change it to 7.1 at 1536 kbps in lines 82–84. I don’t see much point in using the 7.1 Blu-ray mode though. IMO, the purpose of lossy Atmos is to make it playable on a TV’s internal player or apps like Plex. The 7.1 Blu-ray mode triggers Atmos on Blu-ray players or devices like the Ugoos, but not on a LG TV’s Plex app, where only the 640 kbps 5.1 AC-3 core is played. That’s why the 7.1 mode isn’t the default in the script.

5.1 vs 7.1 doesn’t matter when you’re bitstreaming Atmos anyway, and you probably wouldn’t be able to tell the difference between 1024 and 1536 kbps. The AC-3 core just ends up being a waste of space.

[en6ads]

You rock, thank you. It works!

I agree the default should be 5.1 1024kbps. I'm only doing this for tv apps / apple tv compatibility and they need 5.1. I tried 7.1 1664kbps and infuse on apple tv does not pass the atmos part. No such issues with 5.1.

The script breaks if there is a ")" in the filename. I know it breaks for other workflows too, and for those you change the filename to prevent it breaking. Please include this workaround for 8-1-3 also.

Much appreciated. Many thanks again for this work!



Ok please try this version. Workflow 8-1-3. (you must put the truehdd.exe in the tools folder)
https://drive.google.com/file/d/128gq8aDUTKA_aT7SQsM9dkjA1EP1sosR/view?usp=drive_link

The default mode is 5.1 Atmos at 1024 kbps, but you can change it to 7.1 at 1536 kbps in lines 82–84. I don’t see much point in using the 7.1 Blu-ray mode though. IMO, the purpose of lossy Atmos is to make it playable on a TV’s internal player or apps like Plex. The 7.1 Blu-ray mode triggers Atmos on Blu-ray players or devices like the Ugoos, but not on a LG TV’s Plex app, where only the 640 kbps 5.1 AC-3 core is played. That’s why the 7.1 mode isn’t the default in the script.

5.1 vs 7.1 doesn’t matter when you’re bitstreaming Atmos anyway, and you probably wouldn’t be able to tell the difference between 1024 and 1536 kbps. The AC-3 core just ends up being a waste of space.

[alexmorph3us]

@Kuler087: I'm a bit confused. When using Dovi Baker to "bake" FEL with the BL what values I should enter in the x265 encoder for: "--master-display L ( min/max luminance)" and "--max-cll" parameters ? The ones from the base layer or the enhancement layer RPU metadata ? I'm asking this because I'm also reinjecting the RPU file with: "--dolby-vision-profile 8.1 --dolby-vision-rpu" as part of the encoding process and there are cases where the MinMDL/MaxMDL (Mastering Display Luminance) for the full 12 bit signal will be different and usually above the 1000 nits that is required for the base HDR10 static metadata. How I should proceed in this case for the final 10 bit encoding ? What is the correct way ?

[Kuler087]

Just look at how workflow 8-2-1 works; it uses the metadata values from the RPU source PQ (MDL) and the light level (CLL/FALL) to generate the static HDR10 metadata for the FEL-baked encode.

[Kuler087]

The script breaks if there is a ")" in the filename. I know it breaks for other workflows too, and for those you change the filename to prevent it breaking. Please include this workaround for 8-1-3 also.


I just tried 8-1-3 Movie.Name(2025) and it works.
In which other workflows does it break? I know there is one or two workflows that will remove the '')''

[TR-9970X]

Ok please try this version. Workflow 8-1-3. (you must put the truehdd.exe in the tools folder)
https://drive.google.com/file/d/128gq8aDUTKA_aT7SQsM9dkjA1EP1sosR/view?usp=drive_link

The default mode is 5.1 Atmos at 1024 kbps, but you can change it to 7.1 at 1536 kbps in lines 82–84. I don’t see much point in using the 7.1 Blu-ray mode though. IMO, the purpose of lossy Atmos is to make it playable on a TV’s internal player or apps like Plex. The 7.1 Blu-ray mode triggers Atmos on Blu-ray players or devices like the Ugoos, but not on a LG TV’s Plex app, where only the 640 kbps 5.1 AC-3 core is played. That’s why the 7.1 mode isn’t the default in the script.

5.1 vs 7.1 doesn’t matter when you’re bitstreaming Atmos anyway, and you probably wouldn’t be able to tell the difference between 1024 and 1536 kbps. The AC-3 core just ends up being a waste of space.

Kuler, you just keep out doing yourself, constantly adding to this already awesome set of tools, and taking the time to explain pretty much anything people ask of you....

Now I haven't used Dovi for quite some time, being too distracted with other stuff, but that's basically all done now :)

And I'm very curious about this truehdd, dee, deezy addition.

You've released a new script that has this new function, and has been tested by the user that asked about it.

You mention that truehdd.exe needs to added to the tools folder, but is that all ??

I know that dee can already be added (from a long time back), but is deezy required as well ??

It's just that I am setting up a fresh Windows 25H2 on the Threadrippers, and it would be nice to get everything organised from the get go.

Regards.

[Kuler087]

The script does not use Deezy.
You only need to add the truehdd.exe and, of course, the paidware DEE folder.

[TR-9970X]

The script does not use Deezy.
You only need to add the truehdd.exe and, of course, the paidware DEE folder.

Too easy, thank you very much.

[TR-9970X]

I just wanted to say that I got everything set up on 1 pc today....took a while, and I think things have changed a little since I used it last (3.0.5).

Since last time, I had to process as an .mp4 to get DV thru the TV's player.
But I bought an nVidia Shield Pro, and I was able to play DV .mkv's (downloaded), playing thru the Shield :)
So the tests I did today, using 3-1, 3.0.9 beta mkv, to convert a short HDR10 clip, tried a couple of different settings.
They played fine as DV thru the Shield, and just for fun I plugged the usb stick into the LG C4, and to my shock, they played as DV....
Be interested to know whats changed.

[Kuler087]

You probably got the WEBOS-25 update, which enables DV MKV support. C2/C3 should also receive that update.

[TR-9970X]

You probably got the WEBOS-25 update, which enables DV MKV support. C2/C3 should also receive that update.

That's interesting to know...I know I've had quite a few updates since having the C4, but there's not much info about the updates.

The Shield does have more options & functions, but it did take me a while to get it all sorted.

So now I've got 2 options. :)

So now onto setting up DS on other pc's :), while it's fresh in my mind, after today.

Cheers.

[Kuler087]

FYI, the shield is buggy as hell, has color issues (https://drive.google.com/drive/u/1/folders/1uaHVLriou0RFuS8DWtVmxYs1tumPl0F3)in DV and about 30% of the FEL bluray disc rip (https://drive.google.com/drive/u/0/folders/1FS42T95TOSpoy4xtwUBIQmziCe_R_IKe) won't be accurate. You should get the Ugoos (https://www.youtube.com/watch?v=HyrA3KmcJBU) if you want the best quality/experience.

[TR-9970X]

FYI, the shield is buggy as hell, has color issues (https://drive.google.com/drive/u/1/folders/1uaHVLriou0RFuS8DWtVmxYs1tumPl0F3)in DV and about 30% of the FEL bluray disc rip (https://drive.google.com/drive/u/0/folders/1FS42T95TOSpoy4xtwUBIQmziCe_R_IKe) won't be accurate. You should get the Ugoos (https://www.youtube.com/watch?v=HyrA3KmcJBU) if you want the best quality/experience.

Maybe so, but I wanted something that played stuff better than the LG player, and I was hearing that the Shield is hard to beat, I did try some pretty cheap "Android" box's, using Kodi, the play back was jumpy, and I tried every setting, but no luck.

The Shield did have some playback issues, but it turned out I had it connected wrong.....the proper connection with a good hdmi cable solved all that.

But now that the C4 plays DV mkv, I can compare for myself.

I bet the C4 still won't play TrueHD, tho.

Thanks for the advise, but I don't think I need to get more "gadgets".

[MwenDavo]

that sounds right.

ffmpeg or mkvextract is for demuxing HEVC from the container
dovi_tool is for demuxing the layers, extracting & converting the RPU to P8 and then injecting into the encoded HEVC.
ffms2 or dgindex are the frame indexer
avisynth+ is the whole script to be encoded
x265 is for encoding
mkvtoolnix(mkvmerge) is for muxing to mkv

Thanks for the reply!

So you first get the HEVC stream with the baked FEL and the RPU converted to P8, and then you combine them to get the final HEVC to mux. I'll take note of that.

Where does the FEL baking take place? in the avisynth+ script?

Also, what would frame indexer do?

[Kuler087]

FEL is baked when you encode the AviSynth+ script with x265.exe.
The frame indexer is either DGIndex or FFMS2, it serves as the frame-accurate source for the encoder.

[TR-9970X]

A question about Python installation..

I set up DS on the other Threadripper today, but I thought I'd watch your installation video, and the Python install seems to be different to what I recall with earlier builds.

You did mention that Python 3.9 to be install from the Microsoft Store. and then once that's installed run a command :-

pip install numpy

then during that process it was recommended to run this command:-

python.exe -m pip install --upgrade pip

but then I see another cmd:- (on your site)

python (you also have to install the version in Windows Store) python libraries to install: opencv-python, colour-science, scikit-image, matplotlib,
numpy, colour, opencv-python-headless , PyQt5. Open a command window and type for example: pip install scikit-image, so I did that as well, and that installed a lot of other stuff.

C:\Windows\System32>pip install scikit-image
Collecting scikit-image
Downloading scikit_image-0.24.0-cp39-cp39-win_amd64.whl.metadata (14 kB)
Requirement already satisfied: numpy>=1.23 in c:\users\geoff\appdata\local\packages\pythonsoftwarefoundation.python.3.9_qbz5n2kfra8p0\localcache\local-packages\python39\site-packages (from scikit-image) (2.0.2)
Collecting scipy>=1.9 (from scikit-image)
Downloading scipy-1.13.1-cp39-cp39-win_amd64.whl.metadata (60 kB)
Collecting networkx>=2.8 (from scikit-image)
Downloading networkx-3.2.1-py3-none-any.whl.metadata (5.2 kB)
Collecting pillow>=9.1 (from scikit-image)
Downloading pillow-11.3.0-cp39-cp39-win_amd64.whl.metadata (9.2 kB)
Collecting imageio>=2.33 (from scikit-image)
Downloading imageio-2.37.0-py3-none-any.whl.metadata (5.2 kB)
Collecting tifffile>=2022.8.12 (from scikit-image)
Downloading tifffile-2024.8.30-py3-none-any.whl.metadata (31 kB)
Collecting packaging>=21 (from scikit-image)
Downloading packaging-25.0-py3-none-any.whl.metadata (3.3 kB)
Collecting lazy-loader>=0.4 (from scikit-image)
Downloading lazy_loader-0.4-py3-none-any.whl.metadata (7.6 kB)
Downloading scikit_image-0.24.0-cp39-cp39-win_amd64.whl (12.9 MB)
---------------------------------------- 12.9/12.9 MB 6.5 MB/s 0:00:01
Downloading imageio-2.37.0-py3-none-any.whl (315 kB)
Downloading lazy_loader-0.4-py3-none-any.whl (12 kB)
Downloading networkx-3.2.1-py3-none-any.whl (1.6 MB)
---------------------------------------- 1.6/1.6 MB 6.3 MB/s 0:00:00
Downloading packaging-25.0-py3-none-any.whl (66 kB)
Downloading pillow-11.3.0-cp39-cp39-win_amd64.whl (7.0 MB)
---------------------------------------- 7.0/7.0 MB 6.6 MB/s 0:00:01
Downloading scipy-1.13.1-cp39-cp39-win_amd64.whl (46.2 MB)
---------------------------------------- 46.2/46.2 MB 6.7 MB/s 0:00:06
Downloading tifffile-2024.8.30-py3-none-any.whl (227 kB)
Installing collected packages: tifffile, scipy, pillow, packaging, networkx, lazy-loader, imageio, scikit-image
---------------------------------------- 0/8 [tifffile] WARNING: The scripts lsm2bin.exe, tiff2fsspec.exe, tiffcomment.exe and tifffile.exe are installed in

'C:\Users\xxx\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\Scripts' which is not on PATH.
Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
------------------------------ --------- 6/8 [imageio] WARNING: The scripts imageio_download_bin.exe and imageio_remove_bin.exe are installed in

'C:\Users\xxx\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\Scripts' which is not on PATH.
Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
Successfully installed imageio-2.37.0 lazy-loader-0.4 networkx-3.2.1 packaging-25.0 pillow-11.3.0 scikit-image-0.24.0 scipy-1.13.1 tifffile-2024.8.30

-------------------------------------------------------------------------

But before I did all that, I simply installed Python 3.12.3 (as shown in the video), then installed Vapoursynth R72.

Once I'd done that and all the other requirements I started a 3-1 job, and it processed as expected, but then I noticed that it stated that Python wasn't installed, so I stopped the script, and went back and did the Python 3.9 procedure.

So even tho it showed this "error" with Python, it still seemed to be processing properly, but maybe if I had of kept it going, there may have been an error at the finish.

To sum up, is ALL the above still necessary, or is there now a simpler way ??



As for LAVfilters install, simply run that, but then go to Video Configuration and set Hardware Decoder to "D3D11". and Hardware Device to "Auto" (or you could choose the installed GPU)

[Kuler087]

yeah i need to make an updated installation video one day...

Just install python 3.12 from windows and the regular version. then run this cmd:
pip install opencv-python colour-science scikit-image matplotlib numpy colour opencv-python-headless PyQt5

for lavfilters: nothing changed, do like in the video.

[TR-9970X]

yeah i need to make an updated installation video one day...

Just install python 3.12 from windows and the regular version. then run this cmd:
pip install opencv-python colour-science scikit-image matplotlib numpy colour opencv-python-headless PyQt5

for lavfilters: nothing changed, do like in the video.

Excellent, thankyou.

I'll be interested to check processing times between the 4080 Super & the 5070Ti.

Cheers.

[DMD]

If it may be useful, I will explain my procedure for installing Python.

After installing Python, check the version using the cmd command:
python --version

If Windows returns the following
error message:
Python not found; Run without arguments to install from the Microsoft Store or disable the shortcut from Settings > Manage app launch aliases.


To avoid this problem, set the path in the environment variables, locating the installation path of the Python program and the Scripts subfolder.

If you have kept the default installation path, the paths are located in:
(Example for Python39 version)

C:\Users\USERNAME\AppData\Local\Programs\Python\Python39
C:\Users\USERNAME\AppData\Local\Programs\Python\Python39\Scripts
For convenience, copy these paths to a notepad

To open the environment variables window, type the following in the search box: Environment variables, selecting the one related to the system

https://i.postimg.cc/sBY9MY9x/v1.png (https://postimg.cc/sBY9MY9x)

https://i.postimg.cc/jwC5pK5V/v2.png (https://postimg.cc/jwC5pK5V)



Select “New” and copy the previously noted paths one at a time.

https://i.postimg.cc/4nQfWmwR/v3.png (https://postimg.cc/4nQfWmwR)

https://i.postimg.cc/Kkr2pH7x/v4.png (https://postimg.cc/Kkr2pH7x)


Checking the version with the command prompt
python --version

https://i.postimg.cc/BvxgMsGJ/v5.png

Install all necessary modules manually.




[OPTIONAL]
Personally, for this procedure, I opened Notepad and named it “Requirements.txt,” in which I entered the names of all the necessary modules:
Colour
colour-science
matplotlib
numpy
opencv-python
opencv-python-headless
PyQt5
scikit-image

Then I created a launch batch file so that all the modules are installed at once.

@echo off
REM Request elevation to administrator rights
:check_admin
NET SESSION >nul 2>&1
if NOT %ERRORLEVEL% EQU 0 (
echo Request for administrator rights...
goto elevarsi
)
goto continua

:elevarsi
echo. > "%TEMP%\elevate.vbs"
echo Set UAC = CreateObject("Shell.Application") >> "%TEMP%\elevate.vbs"
echo UAC.ShellExecute "%~s0", "", "", "runas", 1 >> "%TEMP%\elevate.vbs"
cscript //nologo "%TEMP%\elevate.vbs"
exit /b

:continua
REM ====================================================================
REM SCRIPT LOGIC START (EXECUTION WITH ELEVATED RIGHTS)
REM ====================================================================

REM Temporarily set the working directory to the script folder
pushd "%~dp0"

echo Execution with administrator rights.

REM Verify that the Requirements.txt file exists
if not exist "Requirements.txt" (
echo.
echo ERROR: File "Requirements.txt" Not found in the script folder.
popd
pause
exit /b
)

echo Installing Python modules from Requirements.txt...

REM Performs pip install -r
pip install -r Requirements.txt

if errorlevel 1 (
echo.
echo ERROR: The installation of the modules failed.
echo Check the previous error messages.
echo.
) else (
echo.
echo Installation completed successfully.
echo.
)

REM Back to previous directory
popd

pause

[alexmorph3us]

Can someone give me an updated installation guide? I've downloaded Python 3.14 from their site. Will it work with this version? Also, is Vapoursynth still required in order for Dovi Script to work properly?

[Kuler087]

avisynth+
https://github.com/AviSynth/AviSynthPlus/releases/download/v3.7.5/AviSynthPlus_3.7.5_20250420_vcredist.exe

python 3.12 windows store
https://apps.microsoft.com/detail/9ncvdn91xzqp?hl=en-US&gl=US

python 3.12 regular version
https://www.python.org/ftp/python/3.12.10/python-3.12.10-amd64.exe

python libraries:
you can do as DMD said but it is just simpler to run this in a Windows command prompt:
pip install opencv-python colour-science scikit-image matplotlib numpy colour opencv-python-headless PyQt5

lavfilters:
https://files.1f0.de/lavf/nightly/LAVFilters-0.80.0-9.exe

madvr:
go in the dovi_scripts/tools/madvr folder and with admin rights, click on the install bat.

vapoursynth no longer required

[alexmorph3us]

@Kuler087: Thank you so much for this! I have already installed LAV filters and Madvr in my system as part of K-lite Codec Pack. In this case, do I still need to run install.bat for Madvr?

I appreciate your help.

[TR-9970X]

Well, I'm glad I wasn't the only one asking about initial installation :)

Thanks Kuler, that makes it a LOT easier :)

[Kuler087]

it depends on which version it is. try 6-2 if it works.
if it doesn't, you will need the version included with the tools pack. 113b IIRC

@Kuler087: Thank you so much for this! I have already installed LAV filters and Madvr in my system as part of K-lite Codec Pack. In this case, do I still need to run install.bat for Madvr?

I appreciate your help.

[alexmorph3us]

it depends on which version it is. try 6-2 if it works.
if it doesn't, you will need the version included with the tools pack. 113b IIRC

It's the latest version 0.92.17 (February 2025).

[Kuler087]

That's one of the beta time-limited builds? If yes, then be careful when it expires, it will break ''madMeasureHDR.exe''...
The version included in the tools pack is the latest beta without this time restriction.

[alexmorph3us]

That's one of the beta time-limited builds? If yes, then be careful when it expires, it will break ''madMeasureHDR.exe''...
The version included in the tools pack is the latest beta without this time restriction.
No. It's an official build
https://www.videohelp.com/software/madVR

But I've installed it a while ago as part of K-lite Codec Pack 19.2.5 which can be found here:
https://codecguide.com/features_mega.htm

[Kuler087]

then it will probably not work.. try 6-2

[kyrunner1]

I know the ugoosamb6+ can do HDR to Dolby vision 8.1 on the fly is there any benefit to re-encode the file using the Dovi scripts?

[Kuler087]

I know the ugoosamb6+ can do HDR to Dolby vision 8.1 on the fly is there any benefit to re-encode the file using the Dovi scripts?

Generating native Dolby Vision will be slightly more accurate and a bit brighter, but it might not be worth the extra effort since the VS10 HDR10-to-DV conversion is already surprisingly good.

Check out my comparison on it:

https://www.youtube.com/watch?v=2N6uBDmSwec

[kyrunner1]

Just finished the video I have a G4 so I think I’ll just keep running with on fly HDR to Dolby Vision. When I was wa5ch the G5 couldn’t tell any difference so I assume G4 would be about the same.