AVS Forum banner

9541 - 9560 of 9581 Posts

·
Registered
Joined
·
116 Posts
I would consider if instant pausing is more important than the potential for stutter and skipped frames. The render queue is definitely there for a reason.
Thanks, I will see how it goes. The actual problem is not instant pausing, that was just the easiest way to describe it. I use Zoom player scene cut editor with approximately 20 to 40 cuts per movie. When a cut happens, you can see action still taking place but can't hear anything, making it appear very strange and not pleasing to the viewer.
 

·
Registered
Joined
·
13,357 Posts
I've been testing the latest Nvidia driver 456.38 over the last few nights, one of the benefits is the improved clock timing. If I recall reading correctly, Madshi worked w/ Nvidia on the improvement.

with the default driver settings i'm getting ~ 1h:45mins before a frame drop which is a great improvement if folks don't want to go digging into the custom timings. I ended up watching Passenger UHD again, what a remarkable looking UHD title.

There is a ton of activity in the VP comparison thread, I hope we see a new release soon to help test.
 

·
Premium Member
Joined
·
3,937 Posts
Deleted.
 

·
Registered
Joined
·
5,392 Posts
that sadly a needed recommended setting for Turing user to a void bugs.

the "number of frames to present in advance" not the render queue to be clear.
Not since many drivers has this been true for most people. nVidia fixed something a while back that fixed this issue.
 

·
Registered
Joined
·
2,135 Posts
Not since many drivers has this been true for most people. nVidia fixed something a while back that fixed this issue.
Yeah I’ve never needed that for my Turing.

I only ever needed to enable vsync and that completely fixed any stuttering on Turing.
 

·
Registered
Joined
·
8,818 Posts
So just starting to really think about how to test MadVR's tone mapping against a reference before I go through the motions of getting an HDR Reference grading monitor here.

I think the first thing I wanted to do it compare MadVR's tone mapping against an SDR Master. To do this, I created an HDR grade of an SDR Photograph, I actually have this as a LUT now in my colour grading software Davinci Resolve.

The idea is, MadVR should be able to look at a 1000 nit, 4000 nit or 10,000 nit grade of the same shot and the end result should be ideally the same as the source master and very comparable, this would also show any colour issues introduced at different brightness levels.

Not just that, and I have already tested this, we will be able to see how much highlight recovery we really need, if ever and perhaps a deeper dive into saturation.

Right now I am completely ignoring P3 and higher. I just started testing with this photo in Rec709, the source is really sRGB obviously but the edit software converted this to Rec709 pretty perfectly, close enough. The grade was done in Rec709 and output as such.

I am learning some interesting things regarding peak nits in colour grading, and I can see why some films shoot way higher in nits than what is listed in the master, this first test is a 4000 nit grade, but during the export conversion process there must be some stray bits introduced during the conversion and MadVR is reading up to 6000 nits. This is NOT visible on my scopes! When I bring the converted video back into Resolve and check it, sure enough there are 'noise bits' showing on the scope reaching up as high as MadVR was reporting. So, I actually think sometimes a lot of what we read as higher peak nits is actually trash pixels that were never intended to be there in the first place. Dynamic Clipping helps a lot, but not all the way.

Anyway, this is really interesting and a good start.

Here is the source photo. The original was from a Hassleblad camera and is 8k. I need to get the original and save a higher bit depth TIFF to create proper 10 bit files for comparison, but these will do for now.



This is what the non tonemapped raw grade looks like. You can see the peak nits showing the top left scopes are just kissing 4000 nits. Slightly under actually.



And, here is the source photo (left) and the MadVR tone mapped image (right). This is with Compromise OFF, No Desat. 100 nits peak, 50 nits dynamic, everything else off.

First of all, this is an INCREDIBLE result in regard to the tone mapping, talk about just about knocking it out of the park, these are almost identical, which is insane given one is pushing 6000 nits and the other only 100. Well done MadVR!!!

The onnly thing I would say based on what Ive been looking at over the last couple days comparing to the SDR Master, I kinda wish MadVR was using a slightly less agressive BT2390 rolloff. You can kinda see that in the face and leg. If it was just a TAD more subdued there, these images would truly be identical.



Compromise OFF, Desat 1+2



Compromise ON.



Based on the above, clearly Compromise On has some incorrect hues, but looking really close, No Comp NoDesat definitely has a tad more saturation in the faces than the original master and the sky is a slightly different more blue, saturated colour.

Early days, but interesting.

There are other examples which have much more of a hue / saturation shift, but it will take me a little bit of time to get this method of comparison optimized.
 

·
Registered
Joined
·
2,382 Posts
This is part of something that is going to be very closely looked at again in the next month or so.

What I can tell you for a fact though, dont Desat is very much NOT the correct way to handle this shot, you lose all information inside the spear.

Dont Desaturate



Desat 1+2



No Compromise

ok sorry for my horrible camera. dont desaturate does show more detail than in your photo, and while in this photo desaturate looks brighter i want to say the more saturated green looks comparably bright in person but that may not be the case. The dimming didnt seem to affect the scene though. What other combinations to test? compromise tone/gamut?

dont desaturate
3043333


Desaturate
3043334
 
  • Like
Reactions: Reezepees

·
Registered
Joined
·
8,818 Posts
ok sorry for my horrible camera. dont desaturate does show more detail than in your photo, and while in this photo desaturate looks brighter i want to say the more saturated green looks comparably bright in person but that may not be the case. The dimming didnt seem to affect the scene though. What other combinations to test? compromise tone/gamut?

dont desaturate


Desaturate
You have still lost 90% of the information inside the spear. Sorry, but more saturation is not the answer here... We have looked at this spear 1000 times, in fact one of the reasons the desat option exists is exactly because of this spear!
 

·
Registered
Joined
·
2,382 Posts
You have still lost 90% of the information inside the spear. Sorry, but more saturation is not the answer here... We have looked at this spear 1000 times, in fact one of the reasons the desat option exists is exactly because of this spear!
right i agree, there were alot of scenes with fire and effects in bvs where you could see the differences with desaturate, ive been trying desat 1 with compromise unchecked, though i often find myself wanting to go back to dont desaturate as if some fine detail/sharpness/depth is lost, but that shouldnt be the case, right?

but is there some other combination of 1+2 or compromise tone/gamut that i should be trying?
 

·
Registered
Joined
·
2,382 Posts
o hmm, these look quite different, your dont desaturate and desaturate vs my dont desaturate. Remember my camera is terrible and obscuring alot of detail and it looks comparably bright to desat... maybe.. @zombie10k @tnaik4 have you tested this on the mico?
3043365

3043366
 

·
Registered
Joined
·
8,818 Posts
o hmm, these look quite different, your dont desaturate and desaturate vs my dont desaturate. Remember my camera is terrible and obscuring alot of detail and it looks comparably bright to desat... maybe.. @zombie10k @tnaik4 have you tested this on the mico?
View attachment 3043365
View attachment 3043366

What are these four images of? Are both the ones on the right mine?

The image I posted are directly off the disc as it were, no projector, the most pure type of comparison, also depends what settings you used to make your comparison, please post your settings.

This is what inside the spear, no tone mapping.



120 nits / 50 nit target / Dont Desat



Desat 1+2



Compromise On (Dumb)



Bluray



These are the source frames, it does not matter if you have a Christie Eclipse, you are not going to get more information that whats in these source frames. You must have paised on a different frame, the spear is pulsing.
 

·
Registered
Joined
·
2,382 Posts

·
Registered
Joined
·
8,818 Posts
gotcha that mustve been it wrong frame. so try desat 1+2 over desat 1?
Personal preference, generally not dont desat though.

Dont desat can look good most of the time, hell, even the comparison I made above only looks correct vs the source master with Dont Desat ON, however some other scenes will absolutely require it.

Hopefully this can be readdressed soon, this has not been looked at for a year or more now.
 

·
Registered
Joined
·
8,818 Posts
Leaving this post here for when we finally start looking at this very soon.

Had to go way back to find it, I could write it in a text doc, but others can be reminded, this is what I think needs to be specifically done as a starting point moving forward.

Saturation Correction should probably be added as its own new option drop down with varying strengths, while also being 're-calibrated' in general since its too strong anyway, so people who like it can use it, and people who dont can turn it off just like Contrast, Shadow etc...

I very much like what 100% hue correction was doing without adding any saturation.

 

·
Registered
Joined
·
116 Posts
Hey all - just making the dive into MadVR 113 up from the last 'stable' release.

Looks like I have some settings to play with now!

Quick thing I noticed - when I tried to use NeoXP High 1.0 my rendering times SHOT up - started losing frames. All other settings identical, if I switch to mirror, log/hill/loghill strong they all play fine.

Having a tough time with so many options. I am not a professional calibrator, so don't know what things are 'supposed' to look like.

Where do I select the NO Compromise option discussed above? Is that in the trade quality for performance tab?
 

·
Registered
Joined
·
116 Posts
Hey all - just making the dive into MadVR 113 up from the last 'stable' release.

Looks like I have some settings to play with now!

Quick thing I noticed - when I tried to use NeoXP High 1.0 my rendering times SHOT up - started losing frames. All other settings identical, if I switch to mirror, log/hill/loghill strong they all play fine.

Having a tough time with so many options. I am not a professional calibrator, so don't know what things are 'supposed' to look like.

Where do I select the NO Compromise option discussed above? Is that in the trade quality for performance tab?
I largely follow SirMaster's settings.
if you want specific advice you should post your specs though.
 

·
Registered
Joined
·
116 Posts
Computer Specs:
Core i7-3770 @ 3.4Ghz (all the time, C states disabled)
16GB ram
GTX 1650 Super - High performance power state selected (never downclocks)
1TB Samsung SSD (evo 860)
Watching with a Sony 285ES Projector on a 130" Screen

I am watching a movie now using Mirror (silversomething's mirror) for the contrast setting. Any opinions on this curve? So far it seems nice and bright.

I noticed that if I uncheck compromise for hdr luminescence that my rendering times JUMP by 10. From 30 to 40ms - and frames begin to be lost. So NO compromise may not be possible for me.

I already had to drop chroma from AA High to AA Medium with beta 113 in order to keep rendering times around 30.
 
9541 - 9560 of 9581 Posts
Top