Page 1 of 2 12 LastLast
Results 1 to 10 of 12

Thread: color balance or Colour balance

  1. #1
    Join Date
    Nov 2008
    Location
    Cleveland Ohio
    Posts
    2,599

    Default color balance or Colour balance

    https://www.kvantlasers.co.uk/latest...e-display-mode

    Never really thought of this but it is almost obvious. Great idea. I'd like to hear how it works from someone who gets one.
    Pretty darn neat.

    Does anyone do this already in software? It would make the most difference for the dim colors. But it is the extremes that make the most emotional impact.

    Just when you thought there was nothing left to innovate.

  2. #2
    Join Date
    Jun 2010
    Location
    Blacksburg VA
    Posts
    137

  3. #3
    Join Date
    Nov 2015
    Location
    DC area
    Posts
    53

    Default

    I've put a fair amount of time into setting up this same kind of correction on my Lightspace projectors using Beyond's advanced color settings, and it has made a massive difference. I'll take some pics once I have everything unpacked and set up again (and some of you may have seen the difference in person at SELEM - I was set up middle right in the gym with 5 3W projectors: 1 Venus II and 2 of the new Starbursts, both with correction, and 2 Plutos with no correction). I use the included grayscale test pattern (two bars with a bunch of circles in between) to tune them - with the default settings, there is usually one missing circle, a slight brightness gradient across the next two, and the remaining ones all pretty much identical at full brightness. After correction, there is a gradual brightness gradient across all the circles, progressing from very dim to full brightness.

    It took a long time to get everything right (figuring out the proper mixes of red/green, green/blue, and blue/red for yellow/cyan/magenta; matching the brightness curves for red/green/blue so that yellow/cyan/magenta also fade in smoothly, and then matching the 3 projectors to each other), but to me the results are definitely worth it. Not only do the colors on the projectors actually match what is on the screen at all brightness levels, but I can also produce a ton of colors that were completely inaccessible before - things like oranges, cyans, and purples that only used a tiny amount of one component color relative to the other are not only usable, but usable at any brightness. I didn't recognize half the cues in my workspace anymore because they were suddenly being displayed accurately!

    The only slight drawback is that, compared to running them on default settings, some power will be lost to gain the extra dynamic range, but if I'm not competing with any other significant light sources and am running all color corrected projectors, I'll take the extra definition over a bit of extra brightness any day. Beyond also has the ability to store multiple color correction profiles, which can be picked either by effect or in the properties of each cue. So in my case, proper looking yellow is a balance of ~100% red + 80% green, which I have set in the (default) "fully balanced" profile, but setting the "high power" profile to display yellow at 100% R/100% G allows me to have certain cues run balls out for a bit of extra brightness if desired.

  4. #4
    Join Date
    Feb 2005
    Location
    Switzerland
    Posts
    1,410

    Default

    I have this since years in two of my projectors, I developped with a friend a special electronic board for this purpose, which digitizes the analog signals of the ILDA color signals, apply a correction on them, and convert it back to analog.

    Kvant even had a prototype in their hands as they were interested.

    It is working very well, but it's very tricky to do quick and clean analog data aquisition. There is some noise at low level and acquisition speed is at 5 us now, so room to improvement but unfortunately I had to stop development...

    Click image for larger version. 

Name:	PopulatedBoard.jpg 
Views:	21 
Size:	447.0 KB 
ID:	52701
    Click image for larger version. 

Name:	Software.png 
Views:	21 
Size:	119.5 KB 
ID:	52702

  5. #5
    Join Date
    Nov 2008
    Location
    Cleveland Ohio
    Posts
    2,599

    Default

    Cool! Thanks for sharing. I think doing it in software makes the most sense since you mKe your own profile for each device but taking something like eye magic unit and having it monitor the output while generating the desired curve would be much faster. Then the device just needs to translate the input voltage on the color lines to give the right curve. Seems much less intensive that way. You can take all day to train the device. Then it's just a curve fit for the voltages. That seems quick enough. Can have different profiles for beams and graphics.

  6. #6
    Join Date
    Jun 2010
    Location
    Blacksburg VA
    Posts
    137

    Default

    Published January 1994...
    One quad op amp, 4 diodes, 8 fixed resistors, 2 variable resistors.
    Repeat for as many color channels as you have.
    Linear Technology Application Note 57

  7. #7
    Join Date
    Feb 2005
    Location
    Switzerland
    Posts
    1,410

    Default

    Quote Originally Posted by kecked View Post
    Cool! Thanks for sharing. I think doing it in software makes the most sense since you mKe your own profile for each device but taking something like eye magic unit and having it monitor the output while generating the desired curve would be much faster. Then the device just needs to translate the input voltage on the color lines to give the right curve. Seems much less intensive that way. You can take all day to train the device. Then it's just a curve fit for the voltages. That seems quick enough. Can have different profiles for beams and graphics.
    Especially if you have several different lasers, I'm pretty confident that the best thing is to have a laser which has a natively perfect linear color balance. If you apply 2.5V on its input, it should output half of its total power, and same for the whole 0..5V range.

    Then it's easy to apply the needed color balance from the software (linear curve, log, high power, balanced etc), without the need to spend time to create a set of curves for each situation for each laser.

    Also very important, you know exactly what power you output, and you can use the software BAM with confidence!

  8. #8
    Join Date
    Dec 2008
    Location
    veenwouden
    Posts
    2,580

    Default

    So bassically it comes down to having a good driver that actually is totally linear. Sweet problem solved then..... Thanks poul!!!!


    Interested in 6-12W RGB projectors with low divergence? Contact me by PM!

  9. #9
    Join Date
    Nov 2008
    Location
    Cleveland Ohio
    Posts
    2,599

    Default

    That's why I mentioned the eye magic gizmo. I have one but never ever powered it up as green diodes showed up at the same time. With that it does make the output linear and may well do the logarithmic curve too not sure. Point is with such a feedback system you could tune a laser by determining its curve and then applying a correction curve. Sort of exactly like an audio system eq sweep with auto eq compensation for the room. Hey there you go use the same device to measure a venue for beam safety.

    If all you want to do is measure the curve of each diode and save the result, you don't need much other than a detector and a way to store the data and a way to make an accurate voltage. I guess you could do it by hand with a volt meter, a solar cell, and a beam splitter. That cuz you saturate the solar cell pretty quick. Think I'll try this out. What to do with the data I'm not sure.

    What I'd like is to be able to use colors just above loading threshold for really suble shadow effects to try and make 3d like perspectives like done in pencil drawings. Having this accuracy would be important to this. Having more than 256 step from 8 bit would help too. Can't be that hard these days to do even 24 bit. Think of the subtleties you could do in abstracts. I think with a capital stupid you might be even be able to fix the ugly yellow from 515nm green but maybe that's inherent in the color curve of those nm mixes.

    If you full time monitored the diodes you could add lots more color variety with say 515, 532. 660, 638. 488,473, 465, 445 and even 405 for fluorescents on say a painted object in a show. There goes the ilda standard....

    Kvant got something here and could have much more.

    One last point. No auto color will ever satisfy everyone with "white". The pang pallet has never made the "white" I prefer and never will. So hand tweaking will never go away. The green you like is based on the green you are use to. When I went from ion to dpss. I hated it. When I went back to diode 515. I then hated that. There is a lot of personal taste in this.
    Last edited by kecked; 08-20-2017 at 04:38.

  10. #10
    Join Date
    Feb 2005
    Location
    Switzerland
    Posts
    1,410

    Default

    Quote Originally Posted by kecked View Post
    That's why I mentioned the eye magic gizmo. I have one but never ever powered it up as green diodes showed up at the same time. With that it does make the output linear and may well do the logarithmic curve too not sure. Point is with such a feedback system you could tune a laser by determining its curve and then applying a correction curve. Sort of exactly like an audio system eq sweep with auto eq compensation for the room. Hey there you go use the same device to measure a venue for beam safety.

    If all you want to do is measure the curve of each diode and save the result, you don't need much other than a detector and a way to store the data and a way to make an accurate voltage. I guess you could do it by hand with a volt meter, a solar cell, and a beam splitter. That cuz you saturate the solar cell pretty quick. Think I'll try this out. What to do with the data I'm not sure.

    What I'd like is to be able to use colors just above loading threshold for really suble shadow effects to try and make 3d like perspectives like done in pencil drawings. Having this accuracy would be important to this. Having more than 256 step from 8 bit would help too. Can't be that hard these days to do even 24 bit. Think of the subtleties you could do in abstracts. I think with a capital stupid you might be even be able to fix the ugly yellow from 515nm green but maybe that's inherent in the color curve of those nm mixes.
    Agree, you want the bigger possible dynamic to take profit of the (relatively) low resolution.

    Quote Originally Posted by kecked View Post
    One last point. No auto color will ever satisfy everyone with "white". The pang pallet has never made the "white" I prefer and never will. So hand tweaking will never go away. The green you like is based on the green you are use to. When I went from ion to dpss. I hated it. When I went back to diode 515. I then hated that. There is a lot of personal taste in this.
    Of course manual work is still mandatory, even if you have perfectly linear colors in each system, but with different RGB power, or different wavelength as you said, you'll have to harmonize the white level between all to have the best rendering when you use all the systems together.
    Btw it's not really the green shade of diodes that is bad imho (still less punchy than 532, ok), but the mix with other colors.
    Same as 462nm, lovely blue but the magentas you get from it I find weak compared to with 445. That's why I use both 445 and 462 in my best projectors And in a perfect world, I would use both 532nm + 520nm as well !

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •