- This topic has 21 replies, 2 voices, and was last updated 7 years, 2 months ago by
.
-
Posts
-
Hi Florian
I would like to CREATE a color profile using Dispcal while taking into account an existing calibration done with Calman. To make a new .icc file containing both. (Profile based on a Calman calibration.)
So the result will be an .icc profile containing the existing calibration from Calman and the color profile from Dispcal (which is created based on the calibration from Calman.)
Could you please tell me how to accomplish this. ie. What settings to choose?
Thank you very much. 🙂
Hi,
I would like to CREATE a color profile using Dispcal while taking into account an existing calibration done with Calman.
Make sure the calibration is loaded, then in DisplayCAL, set tone response on the “Calibration” tab to “As measured” and disable interactive display adjustment. The button at the bottom will change to “Profile only”. Confirm the dialog message that the current calibration will be used when you start profiling.
Thanks!
Quick question… What would be the best (Least expensive) 3D LUT box available that would work best with DisplayCAL for an HDTV?
I think I’ve seen one around $300 somewhere, but I can’t remember it’s name. edit: (I think It’s called ‘eecolor’)
Can you point me (and others) in the most compatible direction? … I’d really like to benefit from one. 🙂
Thank you very much
Yes, the eeColor is the least expensive 3D LUT box available, but keep in mind that it’s being phased out by the company that developed it, and only supports up to 1080p60 (and no frame packed 3D). This limitation may not affect you if 1080p is all you need (and you have no interest in 3D).
Sounds good. Just might order one! (Looking at about $500 Canadian though.) Kinda hard to swallow considering Americans only pay $300. Rrr…
So let me ask you, I’ve noticed that my HDTV’s color gamut is weak in certain color areas. Namely skin-tones, cyan’s and yellows (of course, lol) so I keep increasing the characterization patch-set to correct these areas. But this is overkill to all the other areas of the gamut. (Needs 11,000 patches to correct skin-tones!) … So, is there anyway to determine then generate a patch-set in DisplayCAL that will emphasize these weak areas? … Where can I find detailed instructions that aren’t too difficult to understand.
Thanks so much!
You can use the testchart editor to create a base patch set with a few hundred to around a thousand patches and add (e.g.) patches from image files via drag-n-drop (TIFF / JPEG / PNG).
Ok, this is exciting! … So please help me out then…
My understanding is that I open the testchart editor which will open with the current profile’s patches used…Then I can add reference patches to that set by dragging and dropping image files into the bottom box on the editor page. Patches of the colors from those image will be added to the set. I can drag several images to create more patches Up to 1000…. Does this mean that I need to run the entire set again including the updated patches from the images, or can I update the existing profile with just the new patches from the images?
So let’s say I want to include images with lots of varying skin-tones…Could you please take me through the process. I’d like to use the maximum allowed as the yellow, brown, orange area of my monitor’s gamut suffers a bit.
Just need to know what the steps are to accomplish this wonderful tweak!
Thank you so much. (Can’t wait to finally have a solution to my gamut imbalance)
Edit: I see that there are many options in the editor…Perhaps I should attack my gamut problem in another way. Adaption? distribution? pre-conditioning? … Not entirely sure, but these items sound like possibilities…
Ideas? … And how to set them up?
Thank you once again Florian. 🙂
Set white and black patches to 4, single channel to 17, neutral to 50 and multidimensional centered to 5. Increase iterative patches (optimized farthest point sampling) until you hit around 1000 patches total. Use that as your baseline, you can increase these values if need be.
Leave gamma at 1, set neutral axis emphasis to 50% and dark region emphasis to around 20% respectively. Set preconditioning profile to a previous profile of the TV.
Then create the testchart.
Now you can add patches from images. If an image does not contain an embedded profile, the preconditioning profile will be used to determine the colorimetry of the image pixels. Each image will on average add around 50-100 patches depending on how varied the color information in the image is.Than you for the instructions. I really appreciate it!
I gave it a try and ended up with an xxxl verification that was less accurate than the original profile of 11000 patches. (But seemingly more accurate in the problem areas due to the 10 images I imputed).
So please help me to understand whats going on here… When you tell me to input 1000 patches, what exactly does this do? Does it take a subset of the original profile to re-profile or somehow integrate with the new patch-set from the imputed images? … Shouldn’t I be using the full 11000 patchset plus the imputted colors from the images?
Just need some understanding if you don’t mind.
I’m very excited to now be able to go beyond (improve) over the default profiling procedure!
Thanks Florian for any explanation of the above you can give. 🙂
Edit: I just realized that I should clarify… In the original profile (Done before patch editing) I used 11,000 patches in the ‘Auto optimize’ mode. I need to do that to resolve skin-tone gamut issues. (Inexpensive HDTV, but it gets close.)
I realize that other parts of the gamut don’t require so many patches, which is why I’m trying to target the problem areas with the patch editor by way of imputing images in which the skin-tones don’t look natural. (To profile those actual colors)
So I guess that that is why you say to use 1000 patches as a baseline to see if the rest of the gamut gets resolved. (Along with the use of my inputted images.) … Then I increase the patches until it does.
Is this what is meant by being more efficient?
Still want to know what\where the 1000 patches is coming from.
Thanks a lot!
As I continue I have a question:
I’m assuming that I can raise the Iterative patches (As I do the slider in the profiling tab) to increase the sampling… But what about those other settings you gave me. Do they also need to be raised in some proportional way? … How do I know how much to raise them if say I want to use the full 9999 patches? … Could you give me the settings for such a patch-set. (Or any consideration I might be missing here.)
Thanks for all your help!
OK, so just for fun I tried doing a profile edit using 3000 patches plus another 3000 patches collected from images displaying incorrect flesh-tones… The problem I keep running into is in the xxxL verification… There is always one or two colors that are in the red. (Cyan) but the rest of the report is very good… The thing is that I have never received an ‘in the red’ reading when I profiled the usual way. (Auto optimized).
Any idea why this happens? … And am I able to feed that exact color(s) back in to fix it?
Is there anyway for me to scan the gamut first then automatically have patches generated to fix the actual weak spots? … Why do I need to guess using photos containing approximate color errors… There must be a more pin-pointed way to read the weak spots and make exact patches to correct those areas.
I know I wrote a lot. I’m just very excited about the possibilities, but at the same time frustrated with the learning curve. I just want to know what I’m doing so I can get creative and really dial things in. I’ve been calibrating for years the automatic way, but now I am beginning to see that much more is possible. Are there any dummied down instructions around explaining whats going on with editing testcharts?….. Thank you for your patience. 🙂
OK, so just for fun I tried doing a profile edit using 3000 patches plus another 3000 patches collected from images displaying incorrect flesh-tones… The problem I keep running into is in the xxxL verification… There is always one or two colors that are in the red. (Cyan) but the rest of the report is very good… The thing is that I have never received an ‘in the red’ reading when I profiled the usual way. (Auto optimized).
Yep, it’s hard to beat the auto-optimized chart creation.
Is there anyway for me to scan the gamut first then automatically have patches generated to fix the actual weak spots?
There’s nothing stopping you from creating a text file containing the exact RGB values of the problem areas by hand (if you know them) and adding that to your existing measurement set (same as the images, via drag’n’drop) – but it is usually far easier to just increase the total amount of measured patches.
