AVS Forum banner

7701 - 7720 of 7951 Posts

·
Registered
Joined
·
4,092 Posts
Build number went from ....148 for RC to ....149, so probably not much difference. Honey, I still shrunk the screens :)
What is this shrinking I have seen a few posts about? I do not see any issues with the UI myself. Can you post a pic of what you see so I can better understand?
 

·
Registered
Joined
·
4,296 Posts
Discussion Starter #7,702
What is this shrinking I have seen a few posts about? I do not see any issues with the UI myself. Can you post a pic of what you see so I can better understand?

It is something that only happens in design mode which is only in Calman ultimate or legacy enthusiast versions.

The bug will be fixed in the next release (R2)


Sent from my iPhone using Tapatalk
 

·
Registered
Joined
·
2,755 Posts
What is this shrinking I have seen a few posts about? I do not see any issues with the UI myself. Can you post a pic of what you see so I can better understand?
Actually it happens in normal mode too, not just design, to wit:
 

Attachments

·
Registered
Joined
·
2,755 Posts

·
Registered
Joined
·
9,198 Posts

·
Registered
Joined
·
1,095 Posts
A while back there was a discussion about meter profiling, and how RawXYZ should always be used. Yet I noticed tonight that PD's Guide says: "Select the display type closest to the target display. This will give the new the new [sic] profile a baseline to be adjusted from." I gather the Guide should be ignored in that respect?

Also, the on-screen text in the workflow doesn't seem to say how to configure the pattern generator. I assume the same SDR settings that would be used for normal calibration?

I tried meter profiling for the first time tonight, C6 HDR2000 and i1Pro 3. Can someone tell me if these results look good, bad, or ugly? I did it twice (contact multi-pass using RawXYZ), and got identical results.
 

Attachments

·
Registered
Joined
·
11 Posts
What does this feature really mean?

  • Added 3D LUT Unity Grayscale option for 2019 and 2020 LGs
I'm guessing that it bypasses the 1D grayscale LUT and let the 3D LUT handle the grayscale as well, is this correct?

Has anyone given it a try?
 

·
Registered
Joined
·
9,198 Posts
You were put on the naughty list. :p LOL!
I'm sure it's an oversight! Even if I'm naughty, which I don't think I am :D, I'm naughty customer. ;)
 

·
Registered
Joined
·
9,198 Posts
What does this feature really mean?

  • Added 3D LUT Unity Grayscale option for 2019 and 2020 LGs
I'm guessing that it bypasses the 1D grayscale LUT and let the 3D LUT handle the grayscale as well, is this correct?
No, sorry not correct. The 1D LUT is not changed or bypassed in any way by this option. If you wish to bypass the 1D LUT completely, you have to specifically set the 1DLUT to unity (which happens if you press the "DDC reset" button or the 1D LUT "Autocal" button, and is why you see an immediate visible difference at that moment). That's a separate option, and you've always been able to do that. (*)

This option, on the other hand, is a tweak to the 3D LUT, which you can do (or not do as you see fit) when you are configuring your 3D LUT, if you've done the "normal method" of using a user 1D LUT with autocal. Please see >> THIS POST
 

·
Registered
Joined
·
9,198 Posts
A while back there was a discussion about meter profiling, and how RawXYZ should always be used. Yet I noticed tonight that PD's Guide says: "Select the display type closest to the target display. This will give the new the new [sic] profile a baseline to be adjusted from." I gather the Guide should be ignored in that respect?
I would personally, because I want to follow the science behind what a meter profile actually is. It's not supposed to be layered on top of an EDR because you would then have two separate corrections going on at the same time, and that's completely and utterly unpredictable! It would also mean you are not following either the Four Colour Matrix Method or the Bodner Method - you're doing the "thoth method" :D. You might then say to me, "ah mrtickleuk, but I think I get better results with the thoth method!". I would respectfully, yet vehemently, disagree with such an approach. How do you know? You don't! On the other hand, the Four Colour Matrix Method or the Bodner Method came about after extensive research with reference equipment and are properly specified and peer-reviewed scientific methods. So I would say in response, I bet you haven't done the same level of research and published your own Paper advocating using an EDR as well, have you? ;) Hence: I would never ever use an EDR with a profile. At a high level, these profiling methods are not matters of opinion, or subjectivity, or "what works best for you", or debate or anything like that. The profiling methods are published matters of fact. But I've said my piece on this before, this is only re-hashing the same argument. There are dedicated profiling threads where it's more on-topic to discuss it more if required.

It would be helpful and nice if that Calman guide was clarified at some point, sure. But for me, I'm just pleased that Tyler's later videos are clearer on this subject (and IIRC he also posted a link to the EDR Patent in another thread recently which helped clarify this) and I'll happily take solace from those most recent things, and leave it there.

Also, the on-screen text in the workflow doesn't seem to say how to configure the pattern generator. I assume the same SDR settings that would be used for normal calibration?
There are several "gotchas" here which can lead you to unfortunately make an invalid profile - which may look ok when only validated with just 4 points, and may mislead you into thinking you have a valid profile, but won't work properly when reading the other several million possible points when you use it! (There's definitely a case for validating a profile with more than 4 points; but obviously it takes a lot longer. The LS/CS guys often do 1,000 point validations without even thinking about it!)

Anyway, I make various notes from posts all over the place so I don't have references sorry (and improve my own notes over time), but at present my notes contain these key things:
  • You must profile with the TV set to its fully wide Native gamut. So either set the Gamut to Wide in the menus, or reset the 3DLUT if that Picture Mode has one loaded. (This is critical! If you were to profile in Rec709, it'll never be able to correct accurately anything outside of Rec709. The 3D LUTs are profiled in almost full DCI P3 gamut. So colours would clip and crush!)
  • Do the profile at 100 nits, even if you are going to calibrate to a higher target nits later.
  • Use the correct meter mode (see above).
I tried meter profiling for the first time tonight, C6 HDR2000 and i1Pro 3. Can someone tell me if these results look good, bad, or ugly? I did it twice (contact multi-pass using RawXYZ), and got identical results.
To validate the profile, ie to see if it gives values within the NIST tolerances, I liked the sheet in >> this post
 

·
Registered
Joined
·
1,095 Posts
Thank you for all the information!

It would be helpful and nice if that Calman guide was clarified at some point, sure. But for me, I'm just pleased that Tyler's later videos are clearer on this subject (and IIRC he also posted a link to the EDR Patent in another thread recently which helped clarify this) and I'll happily take solace from those most recent things, and leave it there.
Can you point at one of those later videos? His 2015 video on meter profiling does not use RawXYZ.

You must profile with the TV set to its fully wide Native gamut. So either set the Gamut to Wide in the menus, or reset the 3DLUT if that Picture Mode has one loaded. (This is critical! If you were to profile in Rec709, it'll never be able to correct accurately anything outside of Rec709. The 3D LUTs are profiled in almost full DCI P3 gamut. So colours would clip and crush!)
Yowza, haven't see that mentioned anywhere. The pattern generator itself is still configured as for normal SDR Rec.709 calibration?

At present, you have to do those 12 calculations by hand. Calman does not (yet) do these calculations for you in the workflow, but in the meantime, you can sign up for @Iron Mike's excellent free tool here: https://www.displaycalibrationtools.com/display-profiling-calibration-tools/meter-profile-comparison/ to do the calculations for you.
Disappointing, but glad to know there's a tool available!
 

·
Registered
Joined
·
9,198 Posts
Thank you for all the information!

Can you point at one of those later videos? His 2015 video on meter profiling does not use RawXYZ.
Certainly, the newest is the Bodner method one. Note that this is just (very simplistically, don't jump on me for this next bit) multiple FCMM matrices used together so the exact same underlying principles wrt "not using two different forms of correction" apply as for the simple Four Colour Matrix Method.
Aside: @WiFi-Spy, this was an "unlisted" YouTube video taken from the email I had on the 28th of April about the first beta for Calman 2020. Now that it's been fully released, perhaps you could switch the flag for these videos on YouTube to unhide them? Or, let me know if it's any issue and I'll edit this post to remove it.
edit to add: That was fast! :D

Yowza, haven't see that mentioned anywhere. The pattern generator itself is still configured as for normal SDR Rec.709 calibration?
Yes, the patterns are still the same RGB triplets eg (255,0,0 = "full red" etc). The colours that the TV will produce when sent those patterns depend on the settings at the TV end. (This is why you have to manually intervene to set the TV's gamut as normally you'd have the TV on Auto) Therefore when you think about it, you want your profile to be made of the TV's full native gamut (the reddest red, the greenest green and the bluest blue!) not any smaller sub-set gamut like Rec. 709 or Rec. 601. Otherwise you are limiting the profile to only a narrow gamut and not properly profiling the capabilities of the display. The reason it's more critical for LGs is that the 3DLUT is done in panel native gamut so you're doomed before you start if you've created the profile in less than native gamut.
 

·
Registered
Joined
·
3,764 Posts
No, sorry not correct. The 1D LUT is not changed or bypassed in any way by this option. If you wish to bypass the 1D LUT completely, you have to specifically set the 1DLUT to unity (which happens if you press the "DDC reset" button or the 1D LUT "Autocal" button, and is why you see an immediate visible difference at that moment). That's a separate option, and you've always been able to do that. (*)

.
This still bothers me as my first autocals would visibly change the gray slide very bluish on DDC reset but anymore, nothing changes visibly?
 

·
Registered
Joined
·
9,198 Posts
I just read that I need to profile the tv on wide. Would I need to do this on a Sony TV as well?
We were discussing "creating a meter profile using a spectro", and yes it would apply to any time you do this. This means, using the meter profiling workflow, and using a spectro. This is not something many users do.
EDIT: Tyler has confirmed this to be correct, in his post below.
Obviously, the names on the menu vary between models, ie it may not be called "Wide" on a Sony.

But "profile the TV" can also mean taking thousands of measurements (eg to create a 3DLUT and perform a calibration).

You need to be much clearer as to which one of these you are talking about.
 

·
Registered
Joined
·
1,588 Posts
Certainly, the newest is the Bodner method one. Note that this is just (very simplistically, don't jump on me for this next bit) multiple FCMM matrices used together so the exact same underlying principles wrt "not using two different forms of correction" apply as for the simple Four Colour Matrix Method.
https://www.youtube.com/watch?v=IJbKpSnPFNI
Aside: @WiFi-Spy, this was an "unlisted" YouTube video taken from the email I had on the 28th of April about the first beta for Calman 2020. Now that it's been fully released, perhaps you could switch the flag for these videos on YouTube to unhide them? Or, let me know if it's any issue and I'll edit this post to remove it.
edit to add: That was fast! :D



Yes, the patterns are still the same RGB triplets eg (255,0,0 = "full red" etc). The colours that the TV will produce when sent those patterns depend on the settings at the TV end. (This is why you have to manually intervene to set the TV's gamut as normally you'd have the TV on Auto) Therefore when you think about it, you want your profile to be made of the TV's full native gamut (the reddest red, the greenest green and the bluest blue!) not any smaller sub-set gamut like Rec. 709 or Rec. 601.

@mrtickleuk Has anyone done a full comparison on our OLEDs of the Calman Bodner method versus the volumetric profiling method of Lightspace/Colorspace?
 

·
Registered
Joined
·
4,296 Posts
Discussion Starter #7,719
This still bothers me as my first autocals would visibly change the gray slide very bluish on DDC reset but anymore, nothing changes visibly?

A full reset should always make it bluish. Since that is what the panel native white point is.


Sent from my iPhone using Tapatalk
 

·
Registered
Joined
·
4,908 Posts
@mrtickleuk Has anyone done a full comparison on our OLEDs of the Calman Bodner method versus the volumetric profiling method of Lightspace/Colorspace?
I'm not #mrtickleuk , but I don't believe so. Whoever does it would have to own both CalMAN and LightSpace/Colourspace at license levels that allow use of a pro-grade spectro, plus the spectro itself. An I1Pro/2/3 would probably be good enough for the profiling but not for the verification. It would likely have to be someone like Ted, D-Nice, or Iron Mike with the proper equipment and no skin in the game.
 
7701 - 7720 of 7951 Posts
Top