If I have an image and it is an equal number of pixels for pure black and pure white so 0 and 255 and I then draw an ROI with the color picker setting it to average and the histogram profile set to srgb, what should the resulting value be??
127/128 or 186/187 ish
I got to reading these two posts when I was fumbling around looking up something related to the most recent post on color harmonies. I always come back to this because we from time to time have these discussion’s and we are often showing color patches and citing picker values but are they right and is DT handling scaling correctly or could there be some issues.
The material that I have read as of late can be found here and here wrt how samples are often handled the gamma encoding of data and perform calculations…
A couple of similar threads were started but I don’t think it has gone anywhere. I likely do need to go back and review that… but it would seem that DT might do the same??
This one and the one referred to within…
In any case DT reports 128…
Just curious as to what people think and or any comments to further clarify things…
Well, if you just average pixel values, then around 50% sounds reasonable (“around”, as you can’t be sure you select exactly the same number of bmack and white pixels).
It also depends on what you want as a basis: perceptual value or light energy. Unless you specify that, hard to say whether the colour picker is correct.
As we have pure black and pure white here, I’m not sure gamma changes anything. When you compare other values, the effect of gamma depends on whether the values used are all within the linear part of the curve or not.
I used black and white to keep it simple… for color the same issue you can see it was raised on github… I guess what needs clarification if what value is reported and is the way it is calculated correct, ie should it account for gamma or not… and how are the values handled when scaling, zooming previewing happens as there can be hue shift…
No sure the needle moved on this one… the impact of the correct module order for jpg was noted but still the calculations are not clear … at least to me…
Bottom line is take this image Not sure how it will look in your browser but it looks pink/red until you click on it and then the preview brought up in chrome is the yellow green… again maybe crazy scaling in chrome as well…
Load it in GIMP… activate the picker…see what you get for a color and scale it up and down… (sadly again in chrome it looks funny but in gimp you will see it looks like the selected patch shown in the picker which is captured and not displayed wrong)
Preview is consistent as is I believe the value reported by the picker and I think it would be correct…
Do the same in DT… any zoom that is not 100 or 200 is shown as magenta and the picker value is the math without gamma… At 100 and 200 the image looks more yellow green and the picker does not change value so there is a large disconnect between the displayed preview and the picker… This would lead me to think that things are okay in gimp but DT with its preview pipelines at the very least could be better when scaling… and the picker calculations could also be clarified and or examined…
