- This topic has 116 replies, 12 voices, and was last updated 1 year, 3 months ago by
.
-
Posts
-
As explained above a matrix matches RGB primaries and whitre from one device to another. if uncorrected i1d3 are consistent (under some dE for repeatability) for several display types, the source of error comes from:
a) display itseft, not consistent over time (tremor/shivering in light emited)
b) munki readings are not consistent
If a) uncorrected i1d3 should show the same tremor.So just to make sure: You also think that the difference in those three matrices shows that there must be a problem somewhere?
If yes, how can I find out which component gives me bad results?
It is explained point by point in your quote…
if severeal uncorrected i1d3 shows the same 100% full saturation patches of primaries for the same display (under some low dE)… the culprit of those different matrices is the poor repeatability of your spectrophotometer.
If i1d3 readings show “big” tremor reading 100% full saturatuon primaries (and do not happen for other displays) the culprit is your OLED display which is not stable.And all of this is not related at all with black clippiping in red channel.
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.What correction are you using btw? I returned my Oled a while a go because of bad side bleed so I can’t check my old calibration if it has the same problem.
Do you have the same side bleeding? I saw it very clearly at almost black dark gray areas near the side of the monitor..
For now I decided to go with the BT709 setting within the monitor itself. I talked to a professional calibration guy here, and he told me that he had tested many of the 31LGEP950 and that they come quite properly calibrated. This means I will send back the i1d3. Also I will send back the decklink card and the bi converter which I wanted to use as an external LUT holder, and rather invest the money in a professional calibration. I spent too much time with this and if I can’t trust the result, I will be better off with the factory calibration. According to him, the only thing that should be corrected is the white point due to the colour metamerism.
The only thing that concerns me now is, which ICC-Profile I have to set in macOS (11.6.8 Big Sur) for the Radeon to output the correct signal. As far as I understood I should set the display profile in system preferences to Rec709. At least this gives me the identical image when I compare the HDMI out of the decklink card with the Clean Video feed via Displayport in Resolve.
Or is there a way to set a ICC-profile that doesn’t do any colormanagement?
Before returning i1d3, measure factory calibration TRC(gamma) in OSD Rec709 mode. (or even better, validate it: measurement report, simulation profile + use simulation profile as display profile)
Whatever profile you set as default profile (any OS) must match display behavior of display in its current OSD preset/setting. If std profile in macOS settinsg has a TRC that does not match your OLED display in Rec709 mode.. then it is useless to use such profile.
If you want o avoid a custom profile, you can use DisplayCAL synthetic profile editor (app in DisplayCAL folder). Choose Rec709 RGB primaries (or actual ones measured), and the closest TRC to actual dispay TRC in OSD Rec709 preset.It is silly to have and i1d3 and don’t do this.
Before returning i1d3, measure factory calibration TRC(gamma) in OSD Rec709 mode. (or even better, validate it: measurement report, simulation profile + use simulation profile as display profile)
OK, I will do that. Thats a good idea. <I hope I will figure out the correct settings.
Whatever profile you set as default profile (any OS) must match display behavior of display in its current OSD preset/setting. If std profile in macOS settinsg has a TRC that does not match your OLED display in Rec709 mode.. then it is useless to use such profile.
Well how does the video out card (decklink mini monitor 4k) do it? The decklink card doesn’t know the TRC of my Display either. I just need the video card to output an unaltered signal.
The only thing that concerns me now is, which ICC-Profile I have to set in macOS (11.6.8 Big Sur) for the Radeon to output the correct signal. As far as I understood I should set the display profile in system preferences to Rec709. At least this gives me the identical image when I compare the HDMI out of the decklink card with the Clean Video feed via Displayport in Resolve.
Unfortunately I cannot edit this post anymore. Don’t choose the Rec709 profile as display profile in system preferences! The Clean video feed doesn’t look the same as the decklink out. The standard profile “Ultra Fine Pro” seems to be correct. But I will do the testing as Vincent suggested.
Before returning i1d3, measure factory calibration TRC(gamma) in OSD Rec709 mode. (or even better, validate it: measurement report, simulation profile + use simulation profile as display profile)
OK, I will do that. Thats a good idea. <I hope I will figure out the correct settings.
Whatever profile you set as default profile (any OS) must match display behavior of display in its current OSD preset/setting. If std profile in macOS settinsg has a TRC that does not match your OLED display in Rec709 mode.. then it is useless to use such profile.
Well how does the video out card (decklink mini monitor 4k) do it? The decklink card doesn’t know the TRC of my Display either. I just need the video card to output an unaltered signal.
No decklink, since you are going to return it, hence you are not going to use it. So you need to validate & measure TRC of the display in the configuration you are going to use:
Also I will send back the decklink card and the bi converter which I wanted to use as an external LUT holder, and rather invest the money in a professional calibration
No decklink, since you are going to return it, hence you are not going to use it. So you need to validate & measure TRC of the display in the configuration you are going to use:
I know. But still, how does the decklink card know about the gamma? Besides, if i can’t find a plausible and understandable explanation, then maybe i’ll keep it. basically, what i want is to at least have a reference monitor in davinci resolve. and an i/o card is probably standard. it just annoys me that the decklink card makes my system slower.
The display is warmed up, I will do the testing now. BTW, is there a possibility to test/validate the display within resolve? After all, I need correct rec709 inside Resolve.
Before returning i1d3, measure factory calibration TRC(gamma) in OSD Rec709 mode. (or even better, validate it: measurement report, simulation profile + use simulation profile as display profile)
Do I choose Tone Curve “unmodified” or “Custom Gamma”? Unmodified gives me a warning.
EDIT: OK got it. “Custom gamma” doesn’t let me take the measurement, so it must be unmodified.
No decklink, since you are going to return it, hence you are not going to use it. So you need to validate & measure TRC of the display in the configuration you are going to use:
I know. But still, how does the decklink card know about the gamma?
It does not know.
Besides, if i can’t find a plausible and understandable explanation, then maybe i’ll keep it. basically, what i want is to at least have a reference monitor in davinci resolve. and an i/o card is probably standard. it just annoys me that the decklink card makes my system slower.
The display is warmed up, I will do the testing now. BTW, is there a possibility to test/validate the display within resolve?
Choose Resolve output, simulation profile + use simulation profile as display profile (this has nothing to do with OS settings, it’s just setting the reference for display), no device link (no LUT3D) since you are verifying out of the box behavior.
This (whole setup) is equal even un common GPU to output raw patches with no correction or transformation (other than GPU driver itself), measure them and compare measured values to simulation profile.
Bold letters configuration is the way to test factory or HW calibrations. Display output (1st tab) just sets patch destination.
If that does not output patches through resolve + deckling use RGBCMY saturation patches and IRE grayscale in sample MP4 files encoded in Rec709. AVSForum has a free catalog. You can test display behavior with ArgyllCMS equivalent to calman “HCFR” (Avsforum too, win app) or Calman or whatever you like, even raw command line measurements with spotread. HCFR is easier to use.
No decklink, since you are going to return it, hence you are not going to use it. So you need to validate & measure TRC of the display in the configuration you are going to use:
I know. But still, how does the decklink card know about the gamma?
It does not know.
Obviously. Let me rephrase it: how can the decklink card output correct values for the display’s gamma then?
No decklink, since you are going to return it, hence you are not going to use it. So you need to validate & measure TRC of the display in the configuration you are going to use:
I know. But still, how does the decklink card know about the gamma?
It does not know.
Obviously. Let me rephrase it: how can the decklink card output correct values for the display’s gamma then?
Through a LUT3D run by Resolve (or any other tool). Did you read the faq/howto/purpose of all of this?
Content IN -> LUT3D -> Content (same CIE xy colors) reecoded in display colorspace. That’s why LUT3D creation needs 2 sources, the content colorspace and the display colorspace, obtained though profilling (or if you trust whatever factory or HW calibration, the colospace display -aka generic profile- “should be calibrated to”).
Through a LUT3D run by Resolve (or any other tool). Did you read the faq/howto/purpose of all of this?
Of course I did.
Content IN -> LUT3D -> Content (same CIE xy colors) reecoded in display colorspace. That’s why LUT3D creation needs 2 sources, the content colorspace and the display colorspace, obtained though profilling (or if you trust whatever factory or HW calibration, the colospace display -aka generic profile- “should be calibrated to”).
I thought that it’s the purpose of a video out card like a decklink card to have standards. And if I buy a flanders reference, that comes perfectly calibrated, that I can just connect it and the colors are 100%. Well, Flanders even recalibrates the display for free, if you send it back.
So what you are saying is that you always have to go through a 3DLUT, no matter what perfectly calibrated reference display you have?
Through a LUT3D run by Resolve (or any other tool). Did you read the faq/howto/purpose of all of this?
Of course I did.
Content IN -> LUT3D -> Content (same CIE xy colors) reecoded in display colorspace. That’s why LUT3D creation needs 2 sources, the content colorspace and the display colorspace, obtained though profilling (or if you trust whatever factory or HW calibration, the colospace display -aka generic profile- “should be calibrated to”).
I thought that it’s the purpose of a video out card like a decklink card to have standards.
Not in the way you think.
And if I buy a flanders reference, that comes perfectly calibrated, that I can just connect it and the colors are 100%. Well, Flanders even recalibrates the display for free, if you send it back.
So what you are saying is that you always have to go through a 3DLUT, no matter what perfectly calibrated reference display you have?
Not at all. I have not said such things in my messages.
I think you should get a course on calibration/color management and this stuff before buying more expensive things.
If display matches content (Display is perfectly caibrated to accept some content encoded in some colorspace and display it), that LUT3D is an identity, no transformation at all. = No lut3d is needed. Hence decklink does not need to know gamma or alter signal. Decklink does “nothing” other than ensure whatever range & signal encodiging you configured it.
It expects display to behave as it should (factory/HW callibration) or expect content to have been re-encoded (by Resolve or whatever app) in display colorspace (LUT3D). ***It does nothing to your colors or gamma.***
Not at all. I have not said such things in my messages.
Sorry for the missunderstanding.
I ran a verification with the BT.709 HW/standard settings of the 32EP950 through the Video Card out via Displayport and the Decklink card. Both failed immensely.
At the moment I am doing a calibrating/profiling/3DLut generation as described in the WIKI . I will be happy to post the verification results here once it’s done.
Not at all. I have not said such things in my messages.
Sorry for the missunderstanding.
I ran a verification with the BT.709 HW/standard settings of the 32EP950 through the Video Card out via Displayport and the Decklink card. Both failed immensely.
Share, otherwise we do not know what is wrong. Sometimes it’s user’s fault because a misconfiguration or it can be factory calibration fault: It may be a range issue (16-235 vs 0-255), it can be a whitepoint issue, it can be gray range issue (colors in grey), it could be a gamma issue, it could be an user fault in preset (native RGB oled gamut is show)… all this can be spotted is we have HTML report.
