Idea: is it practical to color grade on a low resolution screen?

[Wayne Morellini]

You are calling it calibrated Marc

I'm sure next years will be so much better against new standards, which is what it is about. Even LCD will have advantages in some areas compared to this years LG OLED. But agreed. If all you are looking for is P3 it is well worth looking at this years LG or Panasonic 900

 

[Björn Benckert]

What about the mac lg 5k screens that will be released this month 10bit 5k for 1k usd seams lika a a good jump start to up the monitor quality levels for the comming year no? They state the price is only for a limited time... normally monitors get cheaper over time but the price seams really good. Apperently its possible to drive even over tb2 with a adapter.

Any one with more knowledge that knows more about this screen?

 

[jake blackstone]

I'm reposting this from LGG.
The LG OLEDs do make decent client monitors, but LGs high-resolution LCD computer monitors like the 5K Ultrafine, and the 4K 31MU97, all have off-axis viewing angle issues that makes them a no-go for anyone other than the person sitting directly in front of them. That's why they're often marketed to gamers. They give off a magenta or green cast and the contrast decreases pretty quickly as you move off-center.

 

[Wayne Morellini]

I presume they are P3 compliant, but what about percentage of rec2020 or HDR performance?

 

[Marc Wielage]

I presume they are P3 compliant, but what about percentage of rec2020 or HDR performance?

I don't think you can get anything even vaguely approaching "true" Rec2020 or Rec2100 performance out of a $600 computer display. It's tough to get it with a $3000 display, and even the $30,000 Sony X300 can only hit about 90% of all the parameters. All this stuff is much, much harder to do than a lot of people grasp.

HDR is far more difficult because of voltage regulation and issues with ABL. The set might be able to handle (say) 1000 nits in a small square area, but no way can it handle it across the entire face of the display. This has been heavily documented all over the web. It also makes calibration really difficult, because it boils down to where do you measure the display for brightness parameters? If it's different in the center than in the corners, how do you average it? Then when you weigh the relative lack of panel consistency into it, you wind up with a monitor that's a little darker and bluer on the left side than the right (or worse).

 

[Hrvoje Simic]

The 55" LG B6 OLED is currently $2200, and there are suppliers that sell it about 10% cheaper:

https://www.amazon.com/LG-Electroni...TF8&qid=1481853342&sr=8-1&keywords=LG+B6+oled

How affordable does it need to be? Where I come from, $2K is dirt cheap for a good, reliable, calibratable monitor.

I've seen the image and it's great but the reliable part is something I'm still considering.
OLED has burn-in issues and we typically don't grade just 16/9 material. And projects take time. It has a system built-in to deal with burn-in in typical scenarios but this is not a typical usage scenario.

 

[M.Halsell]

The 8 bit HDR discussion had me wondering about something I saw but dismissed and couldnt get back to. Senario was a Samsung 4K BluRay player connected to a non HDR 8 bit panel. The output was surprisingly smart looking. I was under the general impression that in the absence of HDR display device, the HDR curve would fall back to SDR. But the output looked modified in someway. I couldn't figure if the BluRay player was making the adjustment or the non HDR panel was responding to the signal.

 

[Wayne Morellini]

M, it could merely be the material is graded to express HDR, and when squeezing it down it maintains some of the look.

 

[Wayne Morellini]

I don't think you can get anything even vaguely approaching "true" Rec2020 or Rec2100 performance out of a $600 computer display. It's tough to get it with a $3000 display, and even the $30,000 Sony X300 can only hit about 90% of all the parameters. All this stuff is much, much harder to do than a lot of people grasp.

HDR is far more difficult because of voltage regulation and issues with ABL. The set might be able to handle (say) 1000 nits in a small square area, but no way can it handle it across the entire face of the display. This has been heavily documented all over the web. It also makes calibration really difficult, because it boils down to where do you measure the display for brightness parameters? If it's different in the center than in the corners, how do you average it? Then when you weigh the relative lack of panel consistency into it, you wind up with a monitor that's a little darker and bluer on the left side than the right (or worse).

My point, it's hardly a comparison, and not what I'm looking for, so it doesn't matter to me that it does P3 at $600, because if that is all it does, it is not all I'm going to be doing or reviewing in the future. I do review camera development footage, so the latest standards allways factor in for reviewing at least. Now I have high speed I can afford to down load uncompressed footage. But nobody is left awake I fear.

In reality, some of the issues you mention are not so important, you either know it's right somewhere in the screen and can ignore the other areas that should look the same as imaginary, or you know its close (if its not, you should be doing something about it, like opting firm another monitor). When you are dealing with fractions of a percent difference, it is probably not going to make much of a difference to the get neral audience, though you and me can see it and it annoys us).

How do you calibrate, 16 bit pixel precision. That is why I keep mentioning it. You can develop a multi level mask product to map to every pixel. But the reality is if it is a fraction of a percent put from worse to vest part of the screen,look at the best part and assume in your mind the rest looks like that in your minds map. Howevery, if the colors/scene is different a lot from one point to another, and you vsnt mentally map the change in white point accriss the screenbeithet, you are stuffed. You won't know what true color difference between the bluish thing in the optimal zone and the greenish thing to one side of the screen is.

I don't know, if I saw reference to an oled backlit quantum dot color filter. However, this and other microked arrays are the way to go short term.

They are targeting true 90% rec2020 this year. These things don't have to stay in the commercial realm.

 

[Marc Wielage]

I've seen the image and it's great but the reliable part is something I'm still considering. OLED has burn-in issues and we typically don't grade just 16/9 material. And projects take time. It has a system built-in to deal with burn-in in typical scenarios but this is not a typical usage scenario.

The burn-in issues are no worse than plasma, and they do go away after a few minutes. Just don't sit on a super high-contrast white-on-black title for ten minutes. Dolby's $100K HDR displays are no different.

As a standard procedure, I'll use the plasma monitor's "wipe/deburn" mode after I do a longterm letterbox job to eliminate or at least reduce any apparent burn pattern. I still see what I would call a "reasonable" full-field gray pattern, so whatever the burn-in is, it's not enough to bother me. Hell, we had the same problem with CRTs, so this is nothing new.

In reality, some of the issues you mention are not so important, you either know it's right somewhere in the screen and can ignore the other areas that should look the same as imaginary, or you know its close (if its not, you should be doing something about it, like opting firm another monitor). When you are dealing with fractions of a percent difference, it is probably not going to make much of a difference to the get neral audience, though you and me can see it and it annoys us).

I think your reality is not my reality. There is a quality point below which I won't go. Even if the client tells me, "this is just for the internet," I'm still going to clip it and watch the levels because it's what I do. You never know -- I know of cases where projects start off like this, then they get "repurposed" later on, and somebody kicks it back from QC because it violated their specs. For this reason, there are compromises I just won't make, because my clients deserve better.

Not adhering to known industry standards is a very slippery slope. The moment you say "it's close enough" or "this monitor is just as good as more-expensive Monitor X," the truth is that neither actually work in the real world. At least, that's my experience. Even if it's a home movie for YouTube, I'll at least glance at the scopes so there's nothing egregiously wrong, and I know the titles aren't blazing and clipping.

We are truly in the Wild West when it comes to monitor standards. I've lost count of how many varieties of HDR exist right now. And the implementation of these standards into consumer sets... it's not great. None of the people I know who are doing actual HDR deliveries are using any consumer sets at all for that purpose. You can get away with it to a point in SDR with high-end consumer displays that have been calibrated, and they are within 95% of what I would call "reasonable."

 

[Wayne Morellini]

Im talking artistic choice. I'm talking within 1% of perfect. The testing was 128 leveks, but inreality they atent linear, so we use 256 levels. Now, the reality is that most of your audience are not going to care between 1 level difference (though pitch black is pitch black). So, back to 1%. Now, from point to point its can be a completely different 8 bits, also. But you get the drift. I am not talking about accumulatuve errirs here. So, I am talking much more restructive than 95% there.