madVR - high quality video renderer (GPU assisted)

[tp4tissue]

Originally Posted by VBB

The funny thing is that the rest of us at least have OLEDs, which makes the most sense, if you really care about picture quality. That's basically step one right there. Then comes the i1D3, but without a spectro to match, you still have no idea how accurate your calibration really is.

tp4tissue, do yourself a favor and go over to AVS and read some of the recent posts by liberator72, jrref, mrtickleuk, and even our own SamuriHL. Talk about passion...

I owned an oled for a short amount of time. I really like the picture too.

What I couldn't accept is how fragile they are, there was clearly bad retention for me in the month i tested it out. It made no sense to stay on oled, when LMCL is right around the corner, and makes oled obsolete.

It doesn't come down to price, even a $3000 tv, over the course of 5 years, that's really nothing.

But Every waking moment at the PC, I gotta add an extra few steps to my usage routine just to not burn in the menu, or take time to consider if a UI would affect the uniformity, That's truly insane.

Picture quality is the most important thing, but OLED is not the solution, it never was, and probably won't be for 10-15 years.

Meanwhile, LMCL this year, all day panel.

[tp4tissue]

Originally Posted by SamuriHL

I go back to my point that if you care that much about calibration then you should NOT accept the results of a 250 dollar meter. Because that meter is not accurate enough to calibrate to the professional standards you claim to care about. Then you say it's not worth it on a consumer display because they are also not able to represent the accuracy of a professional studio quality screen. Yea. No kidding. That's kind of the point we're making here. That for all your talk of calibration, you're still not 100% accurate and never will be with the equipment you're using. But yet you tell others that if you don't calibrate they're wrong. Calibrating with the wrong tools isn't exactly right.

@vbb yup lol

You're wrong, i said it's not worth it to buy $10000 meters on a consumer screens because they're not able to HOLD the accuracy of a studio screen.

There's a difference, YES you can make them more accurate with better measurement tools, but they can't hold that accuracy for long enough to make usage practical.

I've said you need the right tools for the right job.

For any consumer panel, va, oled, any of um', i1d3 is the right tool. And it works within the usage tolerance of these panels.

[SamuriHL]

Originally Posted by tp4tissue

For any consumer panel, va, oled, any of um', i1d3 is the right tool. And it works within the usage tolerance of these panels.

I'm sorry, but, it's nonsense. You absolutely CAN NOT know that you've improved things with an i1d3 that has not been properly profiled. Period. Full stop. You can absolutely make the picture FAR worse with a bad calibration than if you just left it alone. This is a fact. And without profiling your meter against a spectro, you don't know for sure whether you've made it better or worse. So you can sit here and talk about the magic of calibration and accuracy all you want, but, the tool you're using doesn't allow you to know for SURE that you're more accurate than it was before. This is one of the topics we've been discussing in depth in the Calman thread.

[tp4tissue]

Originally Posted by SamuriHL

I'm sorry, but, it's nonsense. You absolutely CAN NOT know that you've improved things with an i1d3 that has not been properly profiled. Period. Full stop. You can absolutely make the picture FAR worse with a bad calibration than if you just left it alone. This is a fact. And without profiling your meter against a spectro, you don't know for sure whether you've made it better or worse. So you can sit here and talk about the magic of calibration and accuracy all you want, but, the tool you're using doesn't allow you to know for SURE that you're more accurate than it was before. This is one of the topics we've been discussing in depth in the Calman thread.

It comes down to what you're trying to improve. The goal of calibration is not to acheive lab accuracy. It's to replicate the artistic intent of the deliverable, (if possible make it look nice).

You fundamentally misunderstand the application of colorimetry if your goal is absolute accuracy.

It'd be nice, but that's not the point of including calibration in the loop.

Having calibration, as i've said in the beginning, is to at the very least remove the blindness of the output devices.

And your statements are blatently false. You can not make a panel worse, unless the spec correction is for the wrong panel.

Maybe back when we had CCFLs and very unstable phosphors that was the case, but in the case of the LEDs, and the Newer generation quantum films, they're very very stable over time, in fact probably OLED is the least stable among all the panel techs out..

Whether or not we know it's accurate is irrelevant. This is the limit of what is possible with consumer devices, these are the product options, and that's that, we do the best with what we've got.

[SamuriHL]

Oh you definitely CAN make it worse. These meters are all over the place in terms of accuracy. So if you think your 250 dollar meter is as accurate as a 7000 Klein you're simply wrong. There's a reason they sell the Spectracal C6 for 700 bucks even though it's literally a rebranded i1d3. It's because they've certified the accuracy of it. When you buy a retail i1d3, until you profile it, you simply don't know how accurate it is. If you calibrate with a meter that is not profiled, you absolutely can make a TV worse. We have tons and tons of posts in the Calman thread to prove that point. This is why most people who care about accuracy get their meter profiled against their tv with a spectro. I have not, so, even I can't trust my calibration results. To me it definitely looks better but I've no way to validate that other than my own eyeballs. There's another whole discussion about whether to use the Lightspace generated EDR for the i1d3 with Calman or to use raw. Because the EDR was profiled against a WOLED panel that is identical to the C8. That's what I use, and I definitely have better results using it. But, what would be better is to profile the meter with a spectro against my C8 directly. Then I'd know for sure that the results I get are accurate. Otherwise, we're just hoping the meter is "close enough". And the retail version of this meter is another lottery. Some are quite accurate. Some aren't.

[tp4tissue]

You're projecting your own poor comprehension. I never said a 250 meter is as accurate as a Klein.

You absolutely CAN NOT make a TV worse as long as the right spectro correction is used, even with a $150 meter.

Even in the cases where the spectral distribution has drifted or is inaccurate, using only a colorimeter will still dramtically decrease the Average error.

We don't NEED TO KNOW how accurate it is. This is not possible or necessary, I don't know why you're stuck on this point, because I never claimed otherwise.

I never said anything about validation, because we don't have the tools or the panels to justify it.

I've only ever said, we need calibration in the loop. Without which it's a mess due to factory tuning. And along the way, I've corrected some of your ramblings.

Nothing you've been harping on conflicts with what I've said. You've created some invisible enemy in your mind, I have no idea who you're arguing with.

[SamuriHL]

Maybe I'm "stuck on that point" because I've been working on calibrating my C8 with people who are considered very experienced in doing it for the past several months. But hey, what do I know. What exactly does calibration mean without knowing what the results are? Let me give you just a small example of what I mean....this meter is not very good at handling low light level situations, which really SUCK for calibrating an OLED because there's no backlight to measure again. So if you just blow through an autocal like they tell you to do, you're going to have very messed up levels in the lower end of the grayscale. And when the 3DLUT is calculated based on that screwed up grayscale measurement and adjustment, guess what happens to the picture quality? My only point that I'm making is that not all "calibrations" are equal. Just like you're not accepting someone manually adjusting a TV with their eyes because that's error prone, I'm saying using a meter that isn't profiled against reference equipment is also error prone. But if you don't want to believe and accept that, it's perfectly fine.

[tp4tissue]

Originally Posted by SamuriHL

I'm saying using a meter that isn't profiled against reference equipment is also error prone. But if you don't want to believe and accept that, it's perfectly fine.

All meters are profiled against referenced equipment, if they weren't they wouldn't be meters.

The spectral correction from a different meter, is to tailor the colorimeter even closer to the true value.

What exactly are you saying I don't believe, I've gone through the same work flow and talk to the same people you have.


I'm sorry to hear autocal didn't work out for you, it seems like you're mad at me, for LG's incompetance.

I never said all calibrations are equal. Why are you saying that I've made all these arguments I haven't made.

[SamuriHL]

What I'm saying is that you keep beating on people that calibration is the end all be all of everything video. And what I'm saying to you is that it's NOT that simple. You can screw things up with a bad calibration, whether you want to believe it or not.

[tp4tissue]

Nothing is screwed up by the calibration.

The panel itself, its software, and its black roll-offs are at fault. Not the meter or the calibration process in it of itself.

And at the end of the day, if the calibration doesn't work, Nothing is worse, it's still as good or bad as you've bought it. settings could be changed back.

So again, you are incorrect in insinuating that a calibration can make a panel worse, it clearly has not and CAN NOT.

And a spectral correction also would not help with your issue. because those corrections don't affect the black roll offs. LG is going to implement the new 1-3% gamma adjust to address this exact issue.

[SamuriHL]

It's clear you have no idea what I'm talking about so I'm just going to leave at that. Of course you can factory reset it back to defaults. I never said you couldn't. I didn't say the TV is permanently screwed up by a bad calibration. However, since us LG owners upload the LUTS directly to the TV, if you DON'T factory reset, you are stuck with a bad calibration until you do. What I SAID was that a bad calibration makes the picture quality worse. And the point of a calibration SHOULD be to get the picture as close to reference as possible. Anyway, you're just not getting what I'm saying so we'll leave it at that.

[Manni]

Please could you take this off topic generic calibration discussion in another thread? Thanks

[SamuriHL]

Don't need to, I'm done.

[SamuriHL]

I have what is very likely a stupid question. But you know me, never afraid to ask! Windows has the idea of scaling when you set a fairly high resolution. For 1080p it tends to go to 150% and for 2160p it defaults to 300% on the desktop. When playing video full screen windowed mode, does that scaling apply to the video at all? Or does madvr's windowed mode "override" the scaling set on the desktop? As I said, probably a ridiculously stupid question on my part but I am curious. Thanks!

[Manni]

Originally Posted by SamuriHL

I have what is very likely a stupid question. But you know me, never afraid to ask! Windows has the idea of scaling when you set a fairly high resolution. For 1080p it tends to go to 150% and for 2160p it defaults to 300% on the desktop. When playing video full screen windowed mode, does that scaling apply to the video at all? Or does madvr's windowed mode "override" the scaling set on the desktop? As I said, probably a ridiculously stupid question on my part but I am curious. Thanks!

The scaling is only for the GUI. I use 200% and that doesn't impact on video content.

[SamuriHL]

Thanks, Manni. I figured that was probably the case but thought I'd ask cause I didn't know for sure.

[Siso]

Guys, which scenario is right - NGU Sharp low+SSIM 1D 100 AR LL, or NGU Sharp medium+Bicubic 150 AR LL?

[ryrynz]

There is no "right" but I would think most here would go with NGU Sharp Medium + Bicubic 150 AR LL for the higher quality upscaling and sharper resulting downscaled image but at the end
nobody knows your panel sharpness or your viewing distance and preferences so just pick what you like.

[nevcairiel]

Originally Posted by SamuriHL

Thanks, Manni. I figured that was probably the case but thought I'd ask cause I didn't know for sure.

Actually, that depends on the player. If a player entirely ignores any Windows DPI handling, then the desktop scaling would in fact affect the video.

An application (ie. the player) would have to tell windows to not scale it, since it does its own scaling. Windows otherwise cannot scale selectively. It either scales the entire player, or nothing of it.

[WuNgUn]

Originally Posted by nevcairiel

Actually, that depends on the player. If a player entirely ignores any Windows DPI handling, then the desktop scaling would in fact affect the video.



An application (ie. the player) would have to tell windows to not scale it, since it does its own scaling. Windows otherwise cannot scale selectively. It either scales the entire player, or nothing of it.

Can someone help with my scaling...
Whenever I'm in 23Hz film mode, my TV overscans. The TV doesn't give me an option to resize and scaling from the nVidia CP just reduces the resolution to make the image fit...
What to do!?