- This topic has 7 replies, 2 voices, and was last updated 9 years, 4 months ago by
.
-
Posts
-
Hi Florian/everyone,
I have been into calibrating for 6 months or so, so still a relative newbie. I successfully calibrated my previous full-array (no LD) LCD TV to my liking using your fantastic tool(s). Now, with my recently purchased FALD (65PUS7601), I am having some real difficulties.
The problem (to my eyes) is that after creating a 3DLUT shadow detail looks.. gray. It looks like it comes out of black far too quickly. I have tried creating different 3DLUTs with different gamma settings from different measurement sets, and they always seem to exhibit this near-black issue. Percevied contrast is reduced due to this, making for a washed out image. FWIW I set madVR to PC Levels, NVIDIA drivers to RGB Full, and I don’t have any HDMI Black Level on this TV as I did on my last.
So, I am using a ColorMunki Display. I use madTPG for the test patterns and have tried 1%/100 APL, full field, and 18%/22%. Part of the problem is that I don’t know what the objective with window patterns is: are we trying to defeat local dimming or deliberately contaminate the picture? Why would we want a BT1886 gamma curve when the measurement data thinks black is 0.040 cd/m2, when in reality (with local dimming) it measures 0.000 cd/m2? I am not sure if measurement windows are my problem.
I have been using HCFR to pre-calibrate as close as possible to my desired gamma/whitepoint, however, my TV generally has a poor gamma response. It has only a 2-point control and is not possible to have 5-10% stimulas at a higher gamma than the midrange. Example: 10% is 2.4, 30% is 2.2, 10% is 2.6, 30% is 2.4. Is it better to pre-calibate even? If so, is it better to be below target gamma for 3DLUT or above? Does it even matter?
A final question. Is it possible for a 3DLUT to be measured ‘perfectly’ in HCFR at a flat 2.4 gamma and still be completely wrong? Does the 3DLUT/calibration data just mask the issue?
Thanks for your time.
Calibrite Display SL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.Hi,
Why would we want a BT1886 gamma curve when the measurement data thinks black is 0.040 cd/m2, when in reality (with local dimming) it measures 0.000 cd/m2?
In that case, you’ll have to make some adjustments to compensate. Try setting gamma to “Relative” as well as output offset to around 90+%. Increase output offset if near-blacks still looks too flat. Decrease if near-blacks get too dark.
I have been using HCFR to pre-calibrate as close as possible to my desired gamma/whitepoint, however, my TV generally has a poor gamma response. It has only a 2-point control and is not possible to have 5-10% stimulas at a higher gamma than the midrange. Example: 10% is 2.4, 30% is 2.2, 10% is 2.6, 30% is 2.4. Is it better to pre-calibate even? If so, is it better to be below target gamma for 3DLUT or above? Does it even matter?
I think for a 3D LUT it may be better if gamma is too low than too high.
A final question. Is it possible for a 3DLUT to be measured ‘perfectly’ in HCFR at a flat 2.4 gamma and still be completely wrong?
No.
In that case, you’ll have to make some adjustments to compensate. Try setting gamma to “Relative” as well as output offset to around 90+%. Increase output offset if near-blacks still looks too flat. Decrease if near-blacks get too dark.
I will try that. What is the advantage of this method compared to using full fields with local dimming enabled? Why not measure ‘native’ black level at 0.000 cd/m2 instead of compensating after-the-fact?
I think for a 3D LUT it may be better if gamma is too low than too high.
Thanks, I definitely won’t calibrate gamma too high beforehand.
No.
I think in that case my calibration is probably quite good and it is simply the volume of shadow detail (even at 2.4) that is revealed, which I am not used to. Perhaps obvious, but pre-calibrating to 2.4 on its own definitely crushes black in comparison to then using a 2.4 3DLUT.
Why not measure ‘native’ black level at 0.000 cd/m2 instead of compensating after-the-fact?
You can try that, but any dynamic dimming feature has the potential to interfere with the measurements.
Why not measure ‘native’ black level at 0.000 cd/m2 instead of compensating after-the-fact?
You can try that, but any dynamic dimming feature has the potential to interfere with the measurements.
Well, I guess you were right. I am still not sure calibrating with local dimming off/defeated is correct, as enabling it and re-measuring completely skews the calibration. I think I will settle with my latest attempt.
Would I be able to use the profile measurements taken during 3DLUT creation to create a profile for Windows? I tried installing the main .icm file from the folder into ‘Profile Loader’ but desktop colours don’t change and images in Photo Viewer (example) appear inversed or such (very green, neon).. noob mistake no doubt. I don’t mind measuring again but I wonder if I did something wrong.
Would I be able to use the profile measurements taken during 3DLUT creation to create a profile for Windows?
Yes.
I tried installing the main .icm file from the folder into ‘Profile Loader’
Not the right way to do it. Open DisplaycAL, select the profile under settings, on the “Display & instrument” tab, select your actual display device, then click “Install profile”, confirm the creation of high quality PCS to device tables.
but desktop colours don’t change
The 3D LUT presets don’t do 1D calibration by default.
images in Photo Viewer (example) appear inversed or such (very green, neon)
Yes.
For desktop colour correction?
Not the right way to do it. Open DisplaycAL, select the profile under settings, on the “Display & instrument” tab, select your actual display device, then click “Install profile”, confirm the creation of high quality PCS to device tables.
This resulted in the same outcome as installing the profile manually. ‘Wacky’ colours in Photo Viewer and no change to the desktop.
The 3D LUT presets don’t do 1D calibration by default.
This is where I am getting caught up. 3D LUT presets don’t do the calibration by default, but how can I create a desktop correction .ICC for Windows using the measurements taken during profiling?
Thanks, that explained things somewhat. I tried creating a XYZ + matrix profile from the measurement data, still no dice (no wacky colours this time..). I think “Create profile from measurement data…” is in the File menu and not “Options”.
I hate bothering you with my noob questions. Thanks for your continued patience; you are a bonafide wizard 🙂
