[catav]

Can someone with an LG OLED please look at the UHD "Battleship " ...Timestamp...1:09:41 and compare the explosion to madVR.

Thanks.

Hey... I have never seen this movie before.. but I really liked it… was so uplifting to watch for a Sci-Fi flick…

The LG OLED with madVR pass-through has way more HDR impact on that frame… (if that is what you’re asking)… Pretty status quo from my experience. Sorry no Lumagen here for testing. But that would make a great compare.

My limited testing shows.. TM’ing @ 65 nits renders a “yellow” explosion and pass-through to OLED @700 nits has a much more “orange” look. I really don’t expect that to change even with the expected destat adjustments.. But would be pleasantly surprised. Would love to see the colorist master on that scene.

 

[Javs]

Hey... I have never seen this movie before.. but I really liked it… was so uplifting to watch for a Sci-Fi flick…

The LG OLED with madVR pass-through has way more HDR impact on that frame… (if that is what you’re asking)… Pretty status quo from my experience. Sorry no Lumagen here for testing. But that would make a great compare.

My limited testing shows.. TM’ing @ 65 nits renders a “yellow” explosion and pass-through to OLED @700 nits has a much more “orange” look. I really don’t expect that to change even with the expected destat adjustments.. But would be pleasantly surprised. Would love to see the colorist master on that scene.
View attachment 3077832

View attachment 3077836

How come you are using 65 nits on a 700nit oled?

 

[catav]

How come you are using 65 nits on a 700nit oled?

Too lazy to turn on 65 nit JVC mostly.. shouldn't the results be close enough for testing?

edit: I'll fire up that bad boy if you want me to (lol).

 

[Javs]

Too lazy to turn on 65 nit JVC mostly.. shouldn't the results be close enough for testing?

edit: I'll fire up that bad boy if you want me to (lol).

if you lowered your oled to approx 65 nits light output then its perfectly valid.

 

[catav]

if you lowered your oled to approx 65 nits light output then its perfectly valid.

Battleship (1.09.41) Rec709, 400 nits

LG OLED - HDR Passthrough

 

[ddgdl]

I think there are issues with grain with max.

We've been busy looking at digital sources and forgot about film. Sigh.

There are probably loads of scenes like this....


At 500 nits (La La Land) frame 69263



sep vs max




edit:

max and sep mix has it to a degree




At 1300 nits (Black Hawk Down)



sep vs max




At 400 nits (La La Land)



sep vs max




At 980 nits (Black Hawk Down)


sep vs max



@Neo-XP , do you have that Great Gatsby scene sample that you used in the past? Same thing with sep vs max there?

Well that would rule out max for me

 

[aron7awol]

if you lowered your oled to approx 65 nits light output then its perfectly valid.

This is not necessary. All he has to do is "output video in HDR format" and as long as the display tracks the EOTF (a low bar to clear with the vast majority of HDR displays) it will be valid. This is how I've tested low nits many, many times, and how I've done desat comparisons. It works extremely well.

 

[quietvoid]

I think there are issues with grain with max.

We've been busy looking at digital sources and forgot about film. Sigh.

There are probably loads of scenes like this....


At 500 nits (La La Land) frame 69263



sep vs max

FWIW I've looked at this frame on my TV and it's hard to see a difference with passthrough at high DPL.
At 500 DPL, max is making the grain very slightly darker but above that there is no difference.
Using 300, the difference with sep is noticeable.

Makes it seems like an issue only at lower nits? Or at least a target that's far from the peak (though this frame measures a bit over 1000 nits).
There was a mention earlier that the gradients were not smooth with max, and that's visible with the 2nd La La Land scene.

 

[woofer]

The real diff is not using madVR TM and just going with pass-through.

Thanks.... could you take a pic of the same scene using "pass-through" with the OLED doing the mapping and taking madVR out of the equation....again thanks.

 

[catav]

Thanks.... could you take a pic of the same scene using "pass-through" with the OLED doing the mapping and taking madVR out of the equation....again thanks.

Yes... I will do that very soon (both phone photo (oh-boy) and PC screen grab).. that should make a good compare... Give me a couple days to produce.

 

[James Freeman]

What is happening here with "max" around all the clouds?

non-tone-mapped / @ 100nits + sep 1.000 50-175 / @ 100nits + max 1.000 50-175

This is really awful. Back to sep for me

Yep.
I'd rather have slightly flatter colors and well behaved greyscale, than artifacts, kinks, bumps, and glitches.

I don't think colors are bad or lacking with Sep because DPL has to be used with DTN, using just 100 DPL to compress a high nit scene does not leave room for proper color.

 

[Neo-XP]

@madshi if you do not plan to review this option again soon, could you please add a "very low" preset for the "highlight recovery" option in the next build?

It is hardcoded like this atm:

1) low: 3.25
2) medium: 4.5
3) high: 7.0
4) very high: 12.0
5) are you nuts!?: 22.0

But the gap between "disabled" and "low" is HUGE:

disabled / low / medium / high / very high / are you nuts!?


IMO, "disabled" lacks details, and "low" is already too strong for a natural result.

I tried to get back to build madVRhdrRestoreDetails18 to find the best compromise, but it is deleted unfortunately: http://madshi.net/madVRhdrRestoreDetails18.zip

Thanks!

 

[markmon1]

Sorry, but no. GUI design costs a LOT of time. And doing this for options which are only supposed to stay in the short term does not seem like a wise investment to me.

Have you considered, since this is just suppose to be for testing anyway, ditching the GUI for all these options entirely and forcing people to just use registry keys? I guess that'd require individually addressable registry keys rather than a single blob of data, but you could do it for all the temp options and stop updating the GUI until release time perhaps?

 

[madshi]

I think there are issues with grain with max.

We've been busy looking at digital sources and forgot about film. Sigh.

There are probably loads of scenes like this....

Ouch. I mean we don't want to hide grain. But that seems excessive!!

@madshi if you do not plan to review this option again soon, could you please add a "very low" preset for the "highlight recovery" option in the next build?

It is hardcoded like this atm:

1) low: 3.25
2) medium: 4.5
3) high: 7.0
4) very high: 12.0
5) are you nuts!?: 22.0

But the gap between "disabled" and "low" is HUGE:

Well, we carefully picked different filter widths, but you can't actually choose the "strength" of the filter atm. I've already talked with Javs about this earlier: Choosing the strength might make sense. E.g. you could use "are you nuts" but with only 20% strength or so. The tricky part is that I don't really want to offer multiple different options for Highlight Recovery. But maybe we can combine it. E.g. increasing not only the filter width, but also the strength at the same time?

But it might make sense to tackle this after custom tone mapping curves, because we may have to revisit the Highlight Recovery filter settings again, anyway, if we do indeed change the tone mapping curve.

I tried to get back to build madVRhdrRestoreDetails18 to find the best compromise, but it is deleted unfortunately: http://madshi.net/madVRhdrRestoreDetails18.zip

It's back up now.

I've removed all older test builds, but I can make any build you guys want available again. The reason why I removed them is because all the zip files didn't have a license in them. And I didn't really like the idea of updating 100+ zip files manually to add a license file. So if you want a specific build to be available, just let me know. But please don't say "all of them".

Have you considered, since this is just suppose to be for testing anyway, ditching the GUI for all these options entirely and forcing people to just use registry keys? I guess that'd require individually addressable registry keys rather than a single blob of data, but you could do it for all the temp options and stop updating the GUI until release time perhaps?

Ditching all the GUI and switching to registry would in itself be a pretty huge change and cost several hours of development time, plus it would require me to spend even more hours on writing a documentation on which registry options are available, and which values each option supports and what it means. Plus changing all that would likely introduce new bugs. That seems rather counter productive atm.

 

[bdht]

MMFR 35192

max:

sep:

 

[AmigoHD]

max:
View attachment 3078133
sep:
View attachment 3078134

how does this frame look with max & sep mix?

 

[bdht]

how does this frame look with max & sep mix?

o sorry forgot to add the frame, edited that last post its 35192

max+sep is right in the middle

 

[stevenjw]

I've removed all older test builds, but I can make any build you guys want available again. The reason why I removed them is because all the zip files didn't have a license in them. And I didn't really like the idea of updating 100+ zip files manually to add a license file. So if you want a specific build to be available, just let me know. But please don't say "all of them".

@madshi, I believe you set the license / public use to expire next month. Is that the case with all builds or only since v113? Have your latest test builds extended this date?

For me, changing my HTPC's date is not an option, so need to know if I have to revert back to v112b, or can I continue to use / test with the latest ones like 120 if you extended that cut-off?

 

[Neo-XP]

Well, we carefully picked different filter widths, but you can't actually choose the "strength" of the filter atm. I've already talked with Javs about this earlier: Choosing the strength might make sense. E.g. you could use "are you nuts" but with only 20% strength or so. The tricky part is that I don't really want to offer multiple different options for Highlight Recovery. But maybe we can combine it. E.g. increasing not only the filter width, but also the strength at the same time?

It would be nice if we could review this again, after the custom curves and desat topics of course.
I am sure there is room for improvements here too. Maybe just lowering the width would do, or using a lower strength, but this will need to be tested properly, in due time.

The "very low" preset would be at 2,625 width to follow the logical sequence, if you want to add it in the meantime.
It could already improve the experience for some users, without hurting anything or requesting any development time, so why not? 🤷‍♂️

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

Considering all the pros and cons of all the options, I stopped with this for now as an "all purpose" or "set it and leave it" configuration:

Thanks to the "don't add peak nits" option, I was able to lower the shadow recovery and the sky detection strengths a bit
Edit: Obviously, you enter your own DPL in the right place.

I am taking a little break now from testing, spending more time to actually watch new HDR movies, patiently waiting for the next build

PS: if you have any questions about my configuration, feel free to send me a PM. I will not answer here, unless it is on topic (about desat options).

 

[aron7awol]

^^^ Blank DPL? Is that a cheat code?