colour difference / colour space

p.pete

Apologies if in the wrong site area...

I'm trying to create some colour stimuli for a psychology project (I'll give a little explanation of the experiment toward end, and what I'm trying to achieve with colours towards the top). I'm sadly lacking in knowledge about colours, but trying to use colors that increment by a reliable amount. I'm trying to find the most variation within one colour also, eg a variation across blues but you'd definitely recognise them all as blue.

I hope I'm not shot down for this, but I'm using MS paint at the moment - I'm sure there's better options, so suggestions welcome but my funding is none (bummer). Originally I choose red, went into the colour editor and just varied it from 0 to 255, nice simple maths. My project supervisor pointed out that some of the colours he would consider to be more black than red, fair point, we don't want a dichotomy of colour for experimental reasons.

Second attempt, I choose green (not too much to the choice), and varied the luminocity this time. I decided cut-offs where it becomes more white than green, and black than green (think I went for 35 and 205 on scale to 240), and then used my variation between these points.
-this works reasonably well but increments don't quite seem to do the same things in the two halves (35-120 and 120-205). Going from green to black the increments seem slightly more noticeable than green toward white, not sure what I'm basing this on...

I'll try organise what I'm trying to get at here:
-does luminocity increase incrementally in any kind of sensible way? (be it mathematically, physiologically or other)
-is there better ways of getting equal increments of colour?
-is there any colour that will give a particularly large colour space, or variety of colours within the one "colour"? (not sure if that question even makes sense)
-is there better free packages for playing around with colour than MS paint? (given I'm familiar with it and hoping to just use it for something basic).

I'll start typing a bit about the study as a second post (not sure anyone's interested!:p)

p.pete

and now a bit about the research...

Sadly not the most exciting study - it's basically looking at categorisation. I'm using circles that vary in diameter and colour as my stimuli. Basically I set up a 2-D space with colour and diameter as the axes - draw a 45 degree line across the space and then create a load of colour stimuli based on points clustered around an area on either side of that line. The line divides between category A and B, so say A is to the left of the line and will typically have larger darker circles, B to the right of the line will have smaller brighter circles (these factors get switched around equally for different participants so there's no actual effect purely due to them being larger or brighter or anything like that).

Participants are presented with circle after circle (lots of fun for them, not) and they have to guess whether it belongs to A or B, and get feedback as to whether they're right or wrong. So they eventually learn what's an A circle and what's a B circle - with the green colours I have at the moment I've tested one or two people already, so this bit's working .

A is going to be the control category for explanation purposes. The study is interested in the effects of diversity. So the example circles we give participants will always be similarly centered around a point in the left of the space that's been created. Some participants in a clustered condition, will have examples from the B area which are more tightly bunched around the one area of the B side of space. Other participants, in the diverse condition, will get examples to learn which are more spread out across that area. Previous research has shown the clustered condition should be the easier one to learn.

After participants have learned the differences between A and B comes the test phase. Here we are interested in generalisation. We show participants some stimuli that they will have seen before, but also some that they won't - these will be from areas of the 2-D space not already shown to them. The participant has to decide if they belong to A, B, or new option Neither. There isn't actually a right or wrong answer, but the theory is that depending on whether participants learned the categories in a clustered or divers condition, they will be more or less likely to generalise - say the new circle belongs to one of the groups rather than saying it's neither.

Told you it was exciting, I've to spend months on this!
Any help on the first post would be greatly appreciated...