Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Color Picker x Sample Point
#1
Photo 
Color Picker x Sample Point
  • Why are the values collected by Color Picker not the same as displayed by Sample point?  Huh
I was using Color Picker to find two points; each on a different image that have the same Value. (in this case I was looking for points with 85%).

But when creating the Sample points at the specific coordinates found in the two images to 85% of value, the values displayed in the Sample Point window are different from 85%. (83.9 point 01 and 86.3 point 02)  Dodgy

At first I thought it might be due to the value used for the Radius - Sample average, in the color Picker. Initially I was using the value of 4 for the radius.
But I tested it with other radius values for the same coordinates and they also didn't match the values displayed in Samples Points.

I finally gave up on Color Picker and used Pointer, which finds and records the same values as those displayed in Sample Point.
  • Am I using Color Picker the wrong way? Any adjustments I should have made? A bug? An outdated tool?
Thx!
                               .....
Samj PortableGimp 2.10.28 - Win-10 /64.
Reply
#2
I can not really reproduce that.    Two different images, same co-ordinates, same (nearly ) Value percentage.

   

Color picker or pointer the reported values match.

   

I would set the picker radius to 1. Zoom right in. Even then that covers adjacent pixels, see the box,  which will average a value. As with most graphics, it depends on the image.  Pointer works for your images, use it.
Reply
#3
No time to check but likely a difference between a linear value and a gamma-corrected one. There are several display options in Sample Points, some are linear, some are gamma corrected...
Reply
#4
Hi rich2005!

(11-23-2021, 09:51 AM)rich2005 Wrote: I can not really reproduce that.    Two different images, same co-ordinates, same (nearly ) Value percentage.

Wow! How does your Color Picker achieve this?

Apparently there is something wrong with my Color Picker; or how I use it.

(11-23-2021, 09:51 AM)rich2005 Wrote: Color picker or pointer the reported values match.

This is the situation I referred to:

[Image: 7ZFW8np.png]

(11-23-2021, 09:51 AM)rich2005 Wrote: I would set the picker radius to 1. Zoom right in.  Even then that covers adjacent pixels, see the box,  which will average a value. As with most graphics, it depends on the image.  Pointer works for your images, use it.

I tried with different values for the radius of the Color Picker, but I didn't get the values displayed in the Sample point window:

Sample Average - Value %
0 - 83.9%
1 - 84.6%
2 - 84.8%
3 - 84.9%
4 - 85% (my initial value for searching the points)

As I said, I abandoned Color Picker and started using Pointer. All results were identical (Pointer and SamplePoint)

[Image: daOI1U2.png]

BTW, is there any way to adjust the sample average used by Pointer in the same way as it is possible for Color Picker?


rich2005, thanks for assessing the situation and trying to elucidate it, Thx!
                               .....
Samj PortableGimp 2.10.28 - Win-10 /64.
Reply
#5
(11-23-2021, 10:37 AM)Ofnuts Wrote: No time to check but likely a difference between a linear value and a gamma-corrected one. There are several display options in Sample Points, some are linear, some are gamma corrected...

Ofnuts,

Somehow I didn't see (realized / noticed) your post today when I posted my previous post. Désolé!  Blush

"likely the difference between a linear value and a gamma-corrected one"

Hmmm...I have no idea what that would be.  Confused

I attach the xcf (41.95 MB) of the images with the marked samples points:

https://www.sendspace.com/file/9u8l7e

Thank you!
                               .....
Samj PortableGimp 2.10.28 - Win-10 /64.
Reply
#6
Same as previous, except more complicated colours so now a definite zoom in to pixel level before comparing values. Keep options the same for all, sample average off, sample merged on

   

   
Reply
#7
(11-24-2021, 08:48 AM)rich2005 Wrote: Same as previous, except more complicated colours so now a definite zoom in to pixel level before comparing values. Keep options the same for all, sample average off, sample merged on

rich2005,

I needed to analyze your images from the previous post to understand why the whole thing works in your examples while I got different results.

After studying and understanding the issue better, everything seems very simple.

My problem was a misconception. I had a total misunderstanding of how the Sample Average Radius for the Color Picker tool worked.

I thought:
- With a radius of 1, 1 pixel would be selected;
- With a radius of 2, 4 pixels would be selected;
- With a radius of 3, 9 pixels etc would be selected.

Ehhh I know... ridiculous, noob thing. (what would it be without the Sample Average option checked being selected?)  Blush

I realized that actually:
- Without the Sample Avarage option checked - 1px x 1px;
- Radius 1 --> 3px x3 px;
- Radius 2 --> 5px x 5px;
- Radius 3 --> 7px x 7 px, etc.
   
In other words, Pointer works the same way as Color Picker with sample average off. (This is now so clear!)

BTW:

Some people use a technique to blur the image when using Sample Points. Thus, they collect a more homogeneous sample from the chosen point.

Pat David in one of his many excellent articles wrote:
"It would be nice if there were an option for Sample Points to allow an adjustable sample radius (if there is an option I haven't found it yet)."

https://pixls.us/articles/color-curves-m...e%20Points

Perhaps using Color Picker (with Sample Average ON) is the option for these situations.

rich2005, thank you so much for the super clarification!
                               .....
Samj PortableGimp 2.10.28 - Win-10 /64.
Reply


Forum Jump: