Click here to Skip to main content
15,070,379 members

Comments by Eric Ouellet (Top 20 by date)

Eric Ouellet 15-Jan-18 11:17am View
I understand that you will not use it now. But it could be useful for anyone with the same need in the future :-) !
Eric Ouellet 2-Dec-14 14:18pm View
Thanks! I didn't know.
Eric Ouellet 2-Dec-14 13:49pm View
I expect that the person who correct what should be corrected will also do whatever should be done. He should know better than me, what should be done.
Eric Ouellet 2-Dec-14 13:37pm View
Done! I think it is also solved -> Article removed... I think.
Eric Ouellet 2-Dec-14 12:49pm View
Thanks. I accepted the 3rd one because it appears to me more complete but I don't know what I should do "Ethically" when a previous answer was also excellent?
Eric Ouellet 2-Dec-14 12:49pm View
Thanks. I accepted the 3rd one because it appears to me more complete but I don't know what I should do "Ethically" when a previous answer was also excellent?
Eric Ouellet 2-Dec-14 12:47pm View
Eric Ouellet 9-Sep-14 11:20am View
Reason for my vote of 5 \n Thanks a lot. Not perfect. But offer a very nice place to start!!!
Eric Ouellet 19-Oct-12 11:50am View
Reason for my vote of 5
Excellent, Thanks.
Very good start...
Eric Ouellet 11-Oct-11 9:06am View
Thanks a lot !

That's exactly what I was looking for.
I knew it was existing but it is hard to find on the net with so many people asking so many already answered questions...

thanks !
Eric Ouellet 4-Oct-11 15:34pm View
Reason for my vote of 5
Rob seems to really know the subject and give good corrections !
Eric Ouellet 4-Oct-11 15:33pm View
Thanks a lot !
I will post my new code in few minutes. (Sean Ewington from Code Project help me, it could takes few hours.)
I let D50 code commented out as reference. I suspect that most people will want to work with D65 white ref.

I'm very happy not to be a chicken in your hands... ;-)
Eric Ouellet 4-Oct-11 15:03pm View
Hello Rob,

This is what I did and would appreciate your comments before posting it.

I added you first code change (the new const).

I modified the code to add the next 4 corrections (D50 ref) but commented out the code after (just as reference for anybody). I added the big last bloc as is.

I tested the result with my little sample and it seems to work fine.

Do you think I made corrections ok ?

"Correction for sRGB with D65 reference white"
Eric Ouellet 4-Oct-11 14:19pm View
Thanks Rob,

To be honnest, your alternative almost gave me an heart attack...
I know the world of mathematics and color like a chicken know how to fly.

But I took a look and try to understand.

There is one thing that I fail to understand. You wrote "Adobe uses RGB Bradford adapted with D65 reference white." and "If you want to use RGB Bradford adapted CIELab conversion do the following:" and ask to change from D65 to D50 at 2 places. I'm a bit confused. You want to turn that from D65 to D50 reference white or the opposite ?


Note: Writing a tips and tricks is a real adventure but I like it :-)
Eric Ouellet 29-Sep-11 12:10pm View
I'm happy to know. I hope it will be as expected !
Eric Ouellet 22-Sep-11 7:58am View
It works as a color generator.
But it is the exact same solution as alternative 1 and it does not meet all the criteria (needs) as explain in alternative 1.
Eric Ouellet 21-Sep-11 11:27am View
Thanks a lot.
I didn't write an article because I thought that the most usefull information is already in Guillaume Leparmentier article in code project (link included in the tips). Majors concepts and information is very well documented there.
I think that my algorithm could be better written by anybody who better master color area. I mainly wanted to share my code with anybody who would need something like that and also share the GetColorDistanceCIELab function that I didn't find in c# anywhere on the web.
But probably it would have been better to explain more how it works. I will try to find some time to do that (but give a few weeks).

Eric Ouellet 20-Sep-11 9:30am View
Thanks for your vote !
I agree the picture is a kind of hook but it also give a chance to see results (included as a demo in the code).
My algorithm is far from being perfect. It is also very inefficient to find very distant colors. But as far as what I have found on the net, it is the best (or only) random generators for the purpose (needs) I had.
Also, I think that it is the only lib with a GetColorDistanceCIELab in .net, which could be usefull in certain specific cases.

Eric Ouellet 15-Sep-11 16:35pm View
Hello Flector,

I think you should not create alternative for you comments. It would be better using "Have a question or comment?" hyperlink into your alternative.
Eric Ouellet 15-Sep-11 16:10pm View
Thanks, you are right.
This is the easiest method. But it did not correspond to my needs...
I think I forgot to add some informations (my needs) in my tips. Thanks.

- Never the same color twice
- Each new color should be easily distinguishable from any previous one
- Be able to set inital colors to reject
Nice to have:
- Add a distance factor to any inital colors to reject in order to stay very far from them. For example a generator to get very different colors from a background color.