White CCTs

Has anyone calculated levels to closely match various CCTs with x7 based fixtures? The spec sheets list the CRI and Duv at various CCT so someone did some very detailed photometrics.

Maybe it's just my eyes. Somehow the HSI selections just don't seem quite right in white territory. I recently noted saturation 0 for a Desire 40 Lustr+ has Cyan, Blue and Indigo all a 0. That would mean the White at 85% is providing all the blue needed. That seems a bit extreme even for a 3200K halogen match.

I realize there are quite a few basic formats and even the individual arrays will vary some. I'm looking for a starting place to give to others. Even just a listing of the levels for the default presets would be a step forward.

It would also be nice if the consoles had various whites in their color pickers. I'm imagining something like the gel pickers where one "brand" could be "Whites" 3000K, 3200K, 3500K, 4500K etc.

Parents
  • I'll try to address your questions:

    1 - Yes, we did get some detailed settings made when we calculated CRI and Duv for various CCTs on our data sheets.  I'm sorry to say, however, that these settings were basically made by eye, on the fly, in order to maximize CRI and minimize Duv.  As far as I know, these settings were not recorded.  They are not the same as what is stored in the consoles for each fixture profile, since the console settings are automated and not tied to CRI.

    2 - All HSI settings and other automatic color mixes within the fixtures are set to produce the highest possible brightness for a given x-y point on the CIE color space in the fixture software.  Whatever the input protocol is for selecting color - HSI, RGB, Studio Mode, etc. - all selections are first effectively converted to an x-y point; then the fixture calculates the color mix from there.  This is not true in Direct Mode, of course, where there is no internal color calculation going on in the fixtures, just strict response to a DMX value for each color primary.

    I'm not exactly sure what you mean when you say that the HSI white points don't seem quite right, so I'll respond in a couple different ways.  

    If what seems odd is the fact that at many white settings, some colors are not used at all, I think we can simply assign responsibility to the imposed constraint of maximizing brightness.  We have toyed around with the option of maximizing other things like CRI or spectral match to a reference lamp type or gel color.  This adds significant complexity to the user, and it often can drop brightness by as much as 50%, so the usefulness comes into question.  We would certainly get a fuller spectrum and probably "whiter-looking" white by using all 7 colors at each CCT, but not necessarily at the highest brightness for that white point.

    It takes surprisingly little blue light, especially the shorter-wavelength indigo light that forms the basis for all white LEDs, to balance other colors to a low-CCT white.  The white emitters in Lustr+ fixtures have a relatively high CCT, so there is an abundance of indigo light coming through them, more than enough to balance the mix at 3200K.

    It could also be that the internal color calculations are simply not able to achieve the absolute precision that our eyes want to see, especially at white.  In order to keep electronics costs in check and to achieve sufficiently quick response times, the fixtures do the best they can to get to the selected color point within the time allocated, but with some variability in absolute accuracy.

    To get really theoretical about this, here is much debate over the usefulness of the CIE color space, because it only accounts for one, isolated part of the comprehensive human visual system, and even that component may not be representative of all "normal" human observers.  Right now it's the best tool we have, but we have seen over and over again that many observers can perceive differences between color mixes that CIE calculates as absolute metamers (color mixes with different spectral components that supposedly produce the same response on the human color receptors in the retina; in other words, that have the identical x-y coordinates on the CIE space).  Perhaps this could explain why the calculations are giving you something that seems off.  What calculates as a true white mix by the CIE math might actually produce a different response in your visual system (or video camera, or color meter, etc.), since no color-mixing LED fixture ever produces a truly continuous spectrum, as in natural daylight or tungsten sources.  However, I think this is perhaps the less likely explanation.

    3 - I don't have ready access to a chart of all fixtures across various CCTs of white.  This seems like something we should have handy, but I'm not certain that such a chart has ever been compiled.  However, the ETC Selador iPhone app can perhaps give you the rough starting points you're looking for.  (Honestly, I don't know whether that app is available yet in other formats besides iPhone.)

    Let me know whether this is at all helpful to you.  I will see what I can learn from our consoles team about getting a chart put together for this.

    --Rob

Reply
  • I'll try to address your questions:

    1 - Yes, we did get some detailed settings made when we calculated CRI and Duv for various CCTs on our data sheets.  I'm sorry to say, however, that these settings were basically made by eye, on the fly, in order to maximize CRI and minimize Duv.  As far as I know, these settings were not recorded.  They are not the same as what is stored in the consoles for each fixture profile, since the console settings are automated and not tied to CRI.

    2 - All HSI settings and other automatic color mixes within the fixtures are set to produce the highest possible brightness for a given x-y point on the CIE color space in the fixture software.  Whatever the input protocol is for selecting color - HSI, RGB, Studio Mode, etc. - all selections are first effectively converted to an x-y point; then the fixture calculates the color mix from there.  This is not true in Direct Mode, of course, where there is no internal color calculation going on in the fixtures, just strict response to a DMX value for each color primary.

    I'm not exactly sure what you mean when you say that the HSI white points don't seem quite right, so I'll respond in a couple different ways.  

    If what seems odd is the fact that at many white settings, some colors are not used at all, I think we can simply assign responsibility to the imposed constraint of maximizing brightness.  We have toyed around with the option of maximizing other things like CRI or spectral match to a reference lamp type or gel color.  This adds significant complexity to the user, and it often can drop brightness by as much as 50%, so the usefulness comes into question.  We would certainly get a fuller spectrum and probably "whiter-looking" white by using all 7 colors at each CCT, but not necessarily at the highest brightness for that white point.

    It takes surprisingly little blue light, especially the shorter-wavelength indigo light that forms the basis for all white LEDs, to balance other colors to a low-CCT white.  The white emitters in Lustr+ fixtures have a relatively high CCT, so there is an abundance of indigo light coming through them, more than enough to balance the mix at 3200K.

    It could also be that the internal color calculations are simply not able to achieve the absolute precision that our eyes want to see, especially at white.  In order to keep electronics costs in check and to achieve sufficiently quick response times, the fixtures do the best they can to get to the selected color point within the time allocated, but with some variability in absolute accuracy.

    To get really theoretical about this, here is much debate over the usefulness of the CIE color space, because it only accounts for one, isolated part of the comprehensive human visual system, and even that component may not be representative of all "normal" human observers.  Right now it's the best tool we have, but we have seen over and over again that many observers can perceive differences between color mixes that CIE calculates as absolute metamers (color mixes with different spectral components that supposedly produce the same response on the human color receptors in the retina; in other words, that have the identical x-y coordinates on the CIE space).  Perhaps this could explain why the calculations are giving you something that seems off.  What calculates as a true white mix by the CIE math might actually produce a different response in your visual system (or video camera, or color meter, etc.), since no color-mixing LED fixture ever produces a truly continuous spectrum, as in natural daylight or tungsten sources.  However, I think this is perhaps the less likely explanation.

    3 - I don't have ready access to a chart of all fixtures across various CCTs of white.  This seems like something we should have handy, but I'm not certain that such a chart has ever been compiled.  However, the ETC Selador iPhone app can perhaps give you the rough starting points you're looking for.  (Honestly, I don't know whether that app is available yet in other formats besides iPhone.)

    Let me know whether this is at all helpful to you.  I will see what I can learn from our consoles team about getting a chart put together for this.

    --Rob

Children
  • You've given me quite a bit of information to chew over. Thank you!

    I'm familiar with the limits of CRI & CIE color spaces and not really concerned with hitting any detailed points about them. What I am looking for is for non-technical users. When a teacher asks for something "not so purple," what is the best answer? Even better would be to have some authority beyond my opinion.

    Since my current focus is on "normal" theatrical use, direct mode but with a console is where I noticed something off. Since there isn't anything in color calculations going on in the fixture that might be where the oddness starts to creep in. It may also be the balance for brightness and I am seeing the spectral discrepancies due to the large bleed through of blue inherent in white LEDs. Regardless I'm looking at what can be done with an Ion/Paradigm/Smarfade and the fixtures. Would whites be better/brighter in another mode? I have to admit I find it hard to believe HSI can accurately describe something as complex as the color of light.

    My last attempt to do a detailed white match was incandescent S4s with Palettas. Yes a significant drop in brightness to get very warm.

    As you say, a chart seems like something you would have handy. If you can find or make one that would be great. I'll look into the Selador App (not an iGuy,) but was told it has the same colors as a console, no CCTs. (But there seem to be colors for frosts ie. R3000 but that's something else entirely!)

Related