Hi there,
I probably found a general misbehavior in qlc+, but I'm not quite sure if it's a bug or a feature.
First some theory about light color mixing; If you mix 1 part red and 1 part green, you'll get yellow. If you mix 1 green and 1 blue you'll get cyan. That was easy.
Now we mix 1 yellow and 1 cyan. We now have 1 red, 2 green and 1 blue part and get pale green. We can also archieve pale green by mixing a fresh green (1 red, 2 green) with 1 blue part.
And here comes qlc+ which does not work like this. If you create some scenes to represent those settings (I used 128 = 1 part) and mix them with a fader, you'll get a dark white (128,128,128) or (1,1,1) Instead of pale green (1,2,1) or (128,255,128). But if you mix fresh green and dark blue, you get whats right.
The problem seems to be the way qlc+ processes multiple channel values. Instead of summing them up it just applies the highest value to the channel, which is mostly not enough.
Take for example the new amazing crossfaders. I like them really, but they have one tiny problem which is related to this. If you have a chaser with two identical scenes and crossfade (linked) from one to another, the value won't stay constant. Instead, it gets reduced to about 1/2 of the original value which is wrong for crossfadings.
That is because of the maximum function (abrv. max). The max sees two values: "1/2 of scene1 and 1/2 of scene2" The max now just takes the highest value - which is 1/2 of the actual correct value.
So my Idea would be to sum the single channel faders/inputs up and scale them to 256 by dividing the result trough the number of faders/inputs. This would solve the problem of color mixing and crossfading, I think.
Any ideas about this being intended to work this way or even if my idea would work?
Edit:
Working with QLC+ 4.6.1 on Windows. But I recognized this problem already about a year ago (QLC+ 4.3.1)
Wrong value mixing behavior
-
Michael Clements
Hi CShark,
I agree that the behaviour with the crossfaders in the cue list is wrong – there is a dip between most pairs of channel values, not just between identical values – it looks like the HTP rule is being applied to the 2 crossfaders. I hadn't noticed this before because I always run projects from a sequence of buttons and not the cue list. I may have missed something in the forum, but it seems strange that there aren't more users of the cue list that have a problem with this.
When crossfading normally between 2 scenes the value of any individual channel should change smoothly towards its new value in the incoming scene at a rate dependent on the crossfade time - which is what qlc+ does when operating the scenes from buttons and in chasers.
I hope the following note on colour mixing helps.
If you have 2 LED fixtures focussed on the same area, one set to yellow (128, 128, 0) and and the other to cyan (0, 128, 128), you will then get the correct double dose of green in the mixed colour. This is the same when mixing the colours from two tungsten filament fixtures fitted with differently coloured gels (additive mixing).
Normally, colour mixing with yellow, cyan and magenta filters is subtractive, with the 3 variable density filters one behind the other in front of a white light source.
Personally I think the simplest way of controlling colour on LED fixtures is to mix the colours you want with RGB (and A and W if you have them) and assign each colour to a different scene and VC control.
A possible strategy for achieving the sort of thing you are describing might be to have controls assigned to (255, 0, 0) (0, 255, 0) and (0, 0, 255) plus (128, 128, 0) (128, 0, 128) and (0, 128, 128).
Then you can the same effect as adding yellow to cyan by combining green (0, 255, 0) with magenta (128, 0, 128). Obviously the whole thing can be refined by adjusting the various values.
I have had lighting desks which will sum (pile add) the values contained in 2 or more scenes as an alternative to HTP, but I have never found a use for it!
I agree that the behaviour with the crossfaders in the cue list is wrong – there is a dip between most pairs of channel values, not just between identical values – it looks like the HTP rule is being applied to the 2 crossfaders. I hadn't noticed this before because I always run projects from a sequence of buttons and not the cue list. I may have missed something in the forum, but it seems strange that there aren't more users of the cue list that have a problem with this.
When crossfading normally between 2 scenes the value of any individual channel should change smoothly towards its new value in the incoming scene at a rate dependent on the crossfade time - which is what qlc+ does when operating the scenes from buttons and in chasers.
I hope the following note on colour mixing helps.
If you have 2 LED fixtures focussed on the same area, one set to yellow (128, 128, 0) and and the other to cyan (0, 128, 128), you will then get the correct double dose of green in the mixed colour. This is the same when mixing the colours from two tungsten filament fixtures fitted with differently coloured gels (additive mixing).
Normally, colour mixing with yellow, cyan and magenta filters is subtractive, with the 3 variable density filters one behind the other in front of a white light source.
Personally I think the simplest way of controlling colour on LED fixtures is to mix the colours you want with RGB (and A and W if you have them) and assign each colour to a different scene and VC control.
A possible strategy for achieving the sort of thing you are describing might be to have controls assigned to (255, 0, 0) (0, 255, 0) and (0, 0, 255) plus (128, 128, 0) (128, 0, 128) and (0, 128, 128).
Then you can the same effect as adding yellow to cyan by combining green (0, 255, 0) with magenta (128, 0, 128). Obviously the whole thing can be refined by adjusting the various values.
I have had lighting desks which will sum (pile add) the values contained in 2 or more scenes as an alternative to HTP, but I have never found a use for it!