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

[dkangel]

Yes, you can still use 6-2 for MDL selection even though it doesn’t matter in CMV4.0. Just don’t use those scene cuts in 3-1, since Resolve’s detection is much better than madVR’s.

so we can use default p3 1000 for everything ?

[Kuler087]

No, you should match the Mastering color primaries: P3 or BT2020.
For the MDL, it doesnt matter but you can just match your source value.

[Kuler087]

I've been using the scripts for a few weeks now and everything is working well aside from the below error. It seems like a trivial thing to do, but it fails on every FEL stream I check. The source is a M2TS file and I am using the latest tools and BETA release. Am I missing a prerequisite for grabbing a PNG from a HEVC encode?

Checking if FEL expands the brightness... This should take only a few seconds...

[ WARN:0@0.020] global loadsave.cpp:278 cv::findDecoder imread_('E:\DoVi.Scripts\temp.folder1\FEL.png'): can't open/read file: check file path/integrity
ERROR=Unable to read the image. Please check the file path.

I'll check it out but I probably just forgot to add code for the DT-DL input since I only work with ST-DL mkv rip.

[dkangel]

No, you should match the Mastering color primaries: P3 or BT2020.
For the MDL, it doesnt matter but you can just match your source value.

ok i can just take the mediainfo value, i mean not have to workflow 6-2 everytime

[Kontrarian]

ok i can just take the mediainfo value, i mean not have to workflow 6-2 everytime

And/or check the movie brightness Google folder.

[Kuler087]

I've added the same prores cropping to the CM_analyze mode in workflow 3-1. It is a lot faster on my system. new beta (https://drive.google.com/file/d/128gq8aDUTKA_aT7SQsM9dkjA1EP1sosR/view?usp=drive_link)

I upgraded my GPU to the RTX 5080, and to my surprise, Dolby Vision analysis and scene cut detection aren’t any faster than on my 3070. The main difference is utilization: the 5080 only runs at about 50%, while the 3070 was consistently at 95–100%.
However, for heavy editing tasks in DaVinci Resolve, like multiple nodes or AI upscaling, the 5080 completely outperforms the 3070, delivering roughly 3× the performance.

[Korbus]

Hi! I'm in the process of overhauling my remux collection and as such I started upgrading the CMv.2.9 RPU of some of them with the CMv4.0 ones from your RPU collections. So far I've hunted for "retail cmv4.0 restored" or "cmv4.0 added", but recently I've noticed a lot of titles with "original L1 CMv4.0 flag" and I'm not very sure what to make of it. Would you please explain what it means (and how they differ from the other two cathegories), as I would gladly use them to continue upgrading my library? Thanks in advance.

[Kuler087]

These new RPUs are effectively identical to using the Ugoos “append CM v4.0” option. For RPUs that were originally mastered in 4.0 but delivered in 2.9, and where no web source exists to restore the 4.0 metadata, it is preferable to retain the original L1 values rather than generating new metadata.

because of:

https://i.ibb.co/2738txqX/firefox-MTh9-Jw-N24-I.png

[Korbus]

Thank you! In this case, for those titles I'll just keep them as they are and let the Ugoos append the CMv4.0 dynamically when watching them.

[ac777103]

could anyone point me to the Ugoos CE build with the append CMv4 option? thanks.

[Gatorman3385]

I think it's the avdvplus builds that have it, but I'd be surprised if they all don't soon.

[Kuler087]

https://github.com/avdvplus/Builds/releases

[ac777103]

https://github.com/avdvplus/Builds/releases

thanks.

[Korbus]

This is one of them (avdvplus): https://github.com/avdvplus/Builds/releases
There is a p3i build too, but that's still a test (dev) build; if you want that one, you should join this Discord channel: https://discord.com/channels/492590071455940612/1462235711552880761

[ac777103]

so much cool stuff in that build! And PGS subs are presented correctly for the first time ever in my experience of Kodi boxes playing HDR content. Only glitch I see is that lipsync on interlaced VC1 is a mile out, but I have other options for that content.

[Kuler087]

Yeah, the Ugoos with those latest CE builds is amazing, pretty much the perfect player... It's even better than my 3K USD Kaleidescape.
I can’t believe some people are still watching Blu-ray rips on a crappy Shield or Zidoo players…

[SamuriHL]

I've added the same prores cropping to the CM_analyze mode in workflow 3-1. It is a lot faster on my system. new beta (https://drive.google.com/file/d/128gq8aDUTKA_aT7SQsM9dkjA1EP1sosR/view?usp=drive_link)

I upgraded my GPU to the RTX 5080, and to my surprise, Dolby Vision analysis and scene cut detection aren’t any faster than on my 3070. The main difference is utilization: the 5080 only runs at about 50%, while the 3070 was consistently at 95–100%.
However, for heavy editing tasks in DaVinci Resolve, like multiple nodes or AI upscaling, the 5080 completely outperforms the 3070, delivering roughly 3× the performance.

Yea my new laptop has a 5080 (obviously not the same as a desktop version) and it's quite a bit faster than my previous laptop's 4080. Not sure if the 16gb of memory makes a difference. I'd say probably.

[Kontrarian]

My laptop has a GTX 1060... :cool:

[TR-9970X]

Yeah, the Ugoos with those latest CE builds is amazing, pretty much the perfect player... It's even better than my 3K USD Kaleidescape.
I can’t believe some people are still watching Blu-ray rips on a crappy Shield or Zidoo players…

I have a "crappy" Shield !!!!

Works well enough for me.

[Kuler087]

Well, it's by far the worst DV player of all time... If you want the best quality, there is only one option...
Doesnt make much sense to me to buy a 2000USD+ TV ( and with calibration) and ruin the quality with an inaccurate player.

- incorrect colors with any P8/P7 rip(red artifacts (https://drive.google.com/drive/u/1/folders/1uaHVLriou0RFuS8DWtVmxYs1tumPl0F3) )
- All of these FEL movie (https://drive.google.com/drive/u/0/folders/1FS42T95TOSpoy4xtwUBIQmziCe_R_IKe)s will have incorrect brightness and tone mapping metadata
- some dv metadata not working (positive lift and L5 (https://drive.google.com/drive/u/1/folders/1LbuwBDZj5x3i1bdbAntsZlr5CiYJ9KKj))
- ycbcr clipping and colorspace conversion issues (https://slow.pics/c/xLgSTMc3)
- dropped frames
- bloated GUI

[Korbus]

I have a "crappy" Shield !!!!

Works well enough for me.

I also have one, but it's collecting dust ever since I've bought the first AM6B+ (yeah, I got two now, one as a spare :)), in July 2024.
I thought I would still use it occasionally, but I was wrong about that.

[Kuler087]

I mean, it's ok for streaming Netflix, Disney etc... But for watching a bluray rip: absolutely not!

[TR-9970X]

I bought the Shield as my ONLY way of playing DV anything, as the LG C4 didn't support DV mkv, at the time.

To my old eyes, the Shield displays 4K DV plenty good enough.

I see no need to "retire" it, until it stops working.

I'm watching Avatar - Fire and Ash 4K DV mkv thru the Shield, and I can't fault it.

[Kuler087]

Some people like their TV in Vivid mode, I prefer accurate Filmmaker Mode that tracks properly and displays movies as they were intended. Taste isn’t the point here, it’s about accuracy, and the Shield isn’t.

[TR-9970X]

Some people like their TV in Vivid mode, I prefer accurate Filmmaker Mode that tracks properly and displays movies as they were intended. Taste isn’t the point here, it’s about accuracy, and the Shield isn’t.

Yes, I'm using Vivid mode, Filmmaker mode looks like shit, if thats the way the movie was meant to look like, then I wouldn't be watching it.

Is that why you go to great lengths to correct them ?

I'm not going to spend $400 + dollars on another unit...it would have be pretty amazingly better.

You are a major videophile, and seem to demand the very best, some of us are not !!!

[Kuler087]

Then this thread isn't for you and go talk about how you like to destroy the image quality somewhere else.

[TR-9970X]

Then this thread isn't for you.

As the first to support you (under a different username) when you started on Doom9, it saddens me for you to say that :(

I still use several of the awesome functions of your app, but I certainly haven't bothered to go to the lengths you do.

I can't relate for basically a one off viewing.

I'm not goin' anywhere !!!

[daffie]

These new RPUs are effectively identical to using the Ugoos “append CM v4.0” option. For RPUs that were originally mastered in 4.0 but delivered in 2.9, and where no web source exists to restore the 4.0 metadata, it is preferable to retain the original L1 values rather than generating new metadata.

because of:

https://i.ibb.co/2738txqX/firefox-MTh9-Jw-N24-I.png

Hmm, so for all my cmv2.9 movies (originally mastered in 4.0) which I converted to cmv4.0 using your Generated RPU's, it is better to retain the L1 values and use this new method?

Big shame I did not keep the original cmv2.9 RPU's of those converted movies :(

[Kuler087]

As the first to support you (under a different username) when you started on Doom9, it saddens me for you to say that :(

I still use several of the awesome functions of your app, but I certainly haven't bothered to go to the lengths you do.

I can't relate for basically a one off viewing.

I'm not goin' anywhere !!!

Watch movies however you want, but don’t come here telling us how great it looks on your Shield with your TV in vivid mode. This isn’t the thread for that.

[Kuler087]

Hmm, so for all my cmv2.9 movies (originally mastered in 4.0) which I converted to cmv4.0 using your Generated RPU's, it is better to retain the L1 values and use this new method?

Big shame I did not keep the original cmv2.9 RPU's of those converted movies :(

Most of the generated RPUs I made are from movies mastered in 2.9.
Most movies mastered in 4.0 have a web source available to upgrade them with the original L3 and L8 trims.

[daffie]

Ok, that is good to hear, thanks!

[Gatorman3385]

Yes, I'm using Vivid mode, Filmmaker mode looks like shit, if thats the way the movie was meant to look like, then I wouldn't be watching it.

Is that why you go to great lengths to correct them ?

I'm not going to spend $400 + dollars on another unit...it would have be pretty amazingly better.

You are a major videophile, and seem to demand the very best, some of us are not !!!

This is a bad April Fools joke, right?

[TR-9970X]

Watch movies however you want, but don’t come here telling us how great it looks on your Shield with your TV in vivid mode. This isn’t the thread for that.

Just know that you started this, with your totally unprovoked and unnecessary comments about the Shield !!

I'm sure there are way more Shield users than Ugoos users.

And from what I've read, the Ugoos are very basic in default configuration.

Each to their own.

[daffie]

Just know that you started this, with your totally unprovoked and unnecessary comments about the Shield !!

I'm sure there are way more Shield users than Ugoos users.

And from what I've read, the Ugoos are very basic in default configuration.

Each to their own.

You are an employee of Nvidia and he hurt your feelings?
My God bro, grow up.

[Kuler087]

Just know that you started this, with your totally unprovoked and unnecessary comments about the Shield !!



False.
The fact that the Shield is inaccurate with DV Blu-ray rips must be repeated as many times as needed, and unlike you, who is obviously ignorant about what quality is, I always provide clear examples of why it shouldn’t be used.

I'm sure there are way more Shield users than Ugoos users.
.
What a great argument.

And from what I've read, the Ugoos are very basic in default configuration.

stop talking about things you dont know or understand.


Each to their own.
The only valid thing you’ve said so far… You have the right to destroy the image as much as you want with your Shield and vivid mode.

PLEASE MOVE ON. This is my last warning!

Edit: I'm starting to understand why your other account got banned...

[Kontrarian]

I'm sure there are way more Shield users than Ugoos users.

I love this fallacy!

You are more than welcome to own, and ever prefer, a Shield and watch your TV in vivid mode, but you do understand this thread, and one's like it, are generally populated with purists.
So just know that when you post your comments.
You can't act surprised when you get these type of responses from other users... it's the internet.

[TR-9970X]

I love this fallacy!

You are more than welcome to own, and ever prefer, a Shield and watch your TV in vivid mode, but you do understand this thread, and one's like it, are generally populated with purists.
So just know that when you post your comments.
You can't act surprised when you get these type of responses from other users... it's the internet.

So what do you have ??

I'm very aware of the "purist's".

Not surprised, I just thought the comment was unnecessary.

And then it blew up. :eek:

[Gatorman3385]

Just know that you started this, with your totally unprovoked and unnecessary comments about the Shield !!

And then it blew up. :eek:

Get 'em out.

[Swede]

Ok, time to cool things down. All of you. If you don't agree at least be constructive in you comments!

[dkangel]

hello everyone
just a question, when i ask info about a mkv dv file, workflow 1, the script give me a number of frames
for example Thor has 165165 frames
the number is good in the mediainfo.json but the script show me frames: 554
what is the problem ?
thank you

[Gatorman3385]

Workflow 1 will take a sample of the file, not the entire file, and give you the info. This is normal.

[dkangel]

Workflow 1 will take a sample of the file, not the entire file, and give you the info. This is normal.

ok but why not show the good number that is in the mediainfo.json file created with workflow 1 ?

"FrameCount": "165165",

[Gatorman3385]

Because it's showing DV info, and grabbing that info from a small sample is quicker.

[dkangel]

i'm very sorry but i don't understand why it's quicker

when i make workflow 1, there are some files created in temp folder

in this files there are a mediainfo.bl.json and a mediainfo.json

554 is the number of frames in mediainfo.bl.json and 165165 is in the mediainfo.json

165165 is the good one

the two files are created instantaneously at the same time, so it's not quicker

why showing the wrong number ?

sorry but english is not my native langage

[Kuler087]

The script displays the DV RPU frame count/summary, not the video itself. To get the full frame count, you would need to extract the entire RPU, which takes time so the script instead generates a small sample that can quickly display the RPU information.

[dkangel]

ok it's the rpu frame not the video, i understand this time :)

[Kontrarian]

@Kuler087

Going through my children's movies and I've come along the Pirates of the Caribbean films which are HDR10 UHDs and it looks like the WEB DVs are a mismatch.
The movies themselves are sub 1,000 nits so generating DV seems like I'm just doing this to help the LG with gradation and posterization, which is fine, but now that the latest CoreELEC builds are enabling stuff like VS10 HDR to HDR is there really any point in generating DV for a 600 nit HDR10 movie?

[Kuler087]

yes, the true DV is still better than vs10.
vs10 works well for bright content, but crushes details a bit in dark content.

https://slow.pics/c/LCQKMpdr

[SamuriHL]

Before the G2 croaked on me, I was seeing brightness flashes in some of the titles I've converted. I always choose option 3 for Resolve for what is supposed to be a balanced profile. I hadn't seen that in any of the ones I've done manually so wasn't sure if it was something in the new automated process or simply my G2 being a jerk. LOL Thoughts?

[Kuler087]

did you check if the RPU scene cuts were aligned with the video in workflow 2-1