I searched on this eyedropper situation and found nothing ! My other thread I didnt reply to as I was trying its answer out, and I then decided to post a question first before answering one, as that thread has nothing in it about this problem I pose so it had no bearing on me posting this thread !! Having then gone to the other thread I have given feedback.
Problems in both threads are similar.
Here, case 1: if the black color is the same - the size of the eyedropper tool does not matter (the size only should be smaller and completly within the black area)
case 2: if the black color is slightly different - use the average (eg on a duplicate of your layer) and than the size of the eyedropper tool does not matter (the size only should be smaller and completly within the black area)
I need to restate what I am trying to do and explain this better..
I need the eyedropper to sample a 31x31 pixel area ON THE IMAGE, so whatever magnification I am at, it is always grabbing 31x31pixels from the image.
Thats it, how do I set it to sample image and not screen ?
many folk reading this will be thinking ...but doesnt it sample the image and not the screen anyway...answer NO, ask adobe why not !
My black square was simply to display and provide evidence of this failing in its design.
Place a piece of transparent acetate over the monitor, mine is 1920 pixels width, 517mm wide , maths 31/1920 x 517mm = 14mm.
so draw a square on that acetate of 14mm x 14mm, thats the area that gets colour sampled, thus there is one magnification value where that square would fit 31pixels on the image, however we all sample away at varying magnifications and we are in fact grabbing anything but 31pixels OF THE IMAGE.
Example 1. I have little patches of light blue amidst slightly darker colours in my image, I can see they are 35 or so pixels in size so eyedropper set to 31pixels will sample them and give me a true average of their pixel shades make up, as they are made of a few differing shades of light blue, some a bit greener some a bit darker. I want to sample these patches in a few areas, each sample will average out the pixels sampled, and I wish then to dump these colours into decent sized squares on a blank page and see if they differ much and get an idea of the average colour between them.
case 1: if the black color is the same - the size of the eyedropper tool does not matter (the size only should be smaller and completly within the black area)
I am not sure what you mean with 'if black colour is the same'...same as what ?
I have when sampling my real life image no goal colour to see if sample different to, the sampling IS MY GOAL.
There are three things that will happen when sampling.
1) If I zoom in too much my 14mmx14mm drawn square will be covering lets say 7 pixels, so sampling at that mag is grabbing only 7 pixels, not the 31 I am after, and missing out on some of the variations in shade.
On the other hand if I zoomed out more, that 14mmx14mm will be covering lets say 110pixels so now I am sampling more than I wish, and maybe pulling in surrounding unwanted colours. I cant tell when I click the eyedropper what its capturing. If it grabs a few slightly darker pixels, I wont know, save for the fact that the sampled colour will be a tad darker or greyer but that might be the light blue anyway correctly sampled. I have no answer to the question I can look at to see if I got the answer right.
case 2 if the black colour is slightly different....different than what ?
I have no target colour to see if my sample is different from, so as to then zoom in or out until I am truly sampling 31pixels.
Why have an eyedropper than ignores the image and samples always the colours that lie within that drawn 14mm box on the acetate ?
I had photographed a colour swatches chart and spent hours sampling all the colours on 51x51 as I could see they were lets say 100pxels square, but couldnt understand why a green was coming out much paler, then zooming in on it and sampling on 51x51 it was as expected, sampled it again at my better overview of the chart and it went pale, twigged that it was capturing the white cardboard between the colours. Had to redo the entire days work. Now as the colours were pure even colours painted out I could zoom fully in and sample, and I had an idea what I was expecting, but in my current task I am after truly capturing 31 x 31 as beyond 35 it gets 'polluted' a bit.
I use the eyedropper a lot and now discovering this I dont know what magnification to be at so that its sampling 31x 31px of the actual image.
I wish to see a sampling ring on the image, and this ring would be really small if I was zoomed out, and would get bigger and bigger as I zoomed in, it is relating to the pixels of the image as such, so I get to see exactly what its sampling.
There's simply no way the eyedropper can sample screen pixels and not image pixels: if it did, you'd be getting Monitor RGB values and not original document RGB values. You're not getting that. You're getting document.
Maybe your cursor is slightly misaligned for some reason, like a driver bug?
the test I did in #1 is proof it is sampling screen as if I start off within the black square i get black, as I zoom out still sampling the centre of the black square I get more and more green in the black as it sees more ane more the surroundings.
Try the test,rgb document. pipette set to sample all layers, black square 51pixels square (preferences >units>pixels) on black layer over background layer white, enables guides to snap to box centre if snaps are on, caplock to give gunsight for pipette, we can now flatten image if wish and set pipette to current layer else keep it on all layers to mimmick sampling a photo.
is sample 51x51 always black at any magnification or going grey more as you zoom out and out ? If that black square is 3mm wide on screen is it black in the result of the sample or grey ?
Erratum by the way in last post 51/1920 x 517mm = 14mm not 31/1920.
I get rgb 222 222 222 if the 51pixel x 51 pixel BLACK square on a white background (A4 300dpi) is 5mm wide on my 1920pixel wide monitor, so as said its sampling the white as well as black. magnification is 36.6%. Its seeing 2.5mm of white all around the black and as such one gets light grey !
making it 3.5mm wide at mag 25% I get 239 239 239 we are lessening the white it sees as we zoom in. the black though is always 51pixels wide in the image.
94.94% mag gives me 13mm wide and I get 36 36 36, just 1nn of white now being included.
104.43% mag gives me 14mm wide (thats the 51pixels screen as mentioned) AND BEHOLD I GET 0 0 0.
Any further zoom in will now give 0 0 0. as black test square is now bigger then the screens 51pixels.
I am using gunsight (caps lock toggle) exact on those cross hairs of guides set to box centre by the way.
I am using mouse centre wheel for magnification hence the odd percentages.