Results 1 to 10 of 10

Thread: Helmholtz-Kohlrausch effect and RGB power balance?

  1. #1
    Join Date
    Jul 2019
    Posts
    19

    Default Helmholtz-Kohlrausch effect and RGB power balance?

    There's perceived brightness and our eyes see green as brighter than red and blue but for lasers and LEDs there is also the Helmholtz-Kohlrausch effect which is dependent on wavelength and here we see a narrow band red light about 5x brighter than green and blue about 2.5x brighter than green. I've noticed displays and projectors have a comparably lower power blue channel, yet with laser show projectors I notice the opposite: our blues are higher powered than both red and green. If Helmholtz-Kohlrausch effect also applies to laser beams wouldn't we already see blue of same power significantly brighter than green?

    https://www.viewsonic.com/de/product...si_lumen-3.jpg

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

    Default

    That’s interesting. Is this band a set wavelength for red and blue or variable based some other effects.

    im thinking we sit in the dark so we are relying on rods. Take a look at this page
    https://en.m.wikipedia.org/wiki/Nigh...l_night_vision

    the graph shows where we see best at night. That may be different in daylight.

    pure guess

  3. #3
    Join Date
    Jun 2009
    Location
    St. Louis, MO
    Posts
    1,199

    Default

    This is talking about an apparent brightness with measurements in Lumens where the eye’s RGB sensitivity has already been factored in to the measurement. All I can contribute is that when I’ve adjusted RGB power in mw to correct for the eye’s sensitivity & matched the beam sizes my Lumia were nicely matched in term of intensity.

    Wetware at work...
    "There are painters who transform the sun into a yellow spot, but there are others who, with the help of their art and their intelligence, transform a yellow spot into the sun." Pablo Picasso

  4. #4
    Join Date
    Jul 2019
    Posts
    19

    Default

    Quote Originally Posted by kecked View Post
    That’s interesting. Is this band a set wavelength for red and blue or variable based some other effects.
    Do you mean the graph I linked to? It's Helmholtz-Kohlrausch effect intensity separately for red, green and blue wavelengths of LED illumination. The idea is a narrow band of red, green or blue wavelengths with one peak wavelength each will produce a much higher preceived brightness than a less narrow bands with less pronounced peak wavelength as with LCD monitors and lamp-based video projectors.
    Here's an example comparing OLED with quantum dots which illustrated what I mean: https://assets.pid.samsungdisplay.co...ral-output.jpg
    Laser show projectors on the other hand have pronounced peak wavelengths.

    im thinking we sit in the dark so we are relying on rods. Take a look at this page
    https://en.m.wikipedia.org/wiki/Nigh...l_night_vision

    the graph shows where we see best at night. That may be different in daylight.
    Right. The shift is not huge, green still appears the brightest compared to red and blue, although shifted more to cyan at night. The Helmholtz-Kohlrausch effect would still apply.




    Quote Originally Posted by laserist View Post
    This is talking about an apparent brightness with measurements in Lumens where the eye’s RGB sensitivity has already been factored in to the measurement
    Right both lumens and Nit take into account eye sensitivity to each wavelength, but they don't seem to take into account the Helmholtz-Kohlrausch effect. Video projector manufacturers now already list separate "LED lumens" and "ANSI lumens" for their LED-powered projectors for this reason. I am guessing the TV and monitor manufacturers don't yet as the Helmholtz-Kohlrausch effect doesn't affect both OLED and backlit LCD much due to their wider spectrum and no single peak wavelength for each color channel.


    All I can contribute is that when I’ve adjusted RGB power in mw to correct for the eye’s sensitivity & matched the beam sizes my Lumia were nicely matched in term of intensity.
    What ratio did you go with?

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

    Default

    I like 4:1:2. Rgb

  6. #6
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,489,446

    Default

    Color balance in general is well-worn territory here on the forum. (Google "Site:Photonlexicon.com Chroma" if you want to dive deep into the drama pool.) But at the end of the day it's important to remember that color perception is subjective. What *I* like might not be the same as what you like - regardless of what the literature says about the color response of the human eye.

    Laserists can argue for hours over what color a given beam is. (Seriously - we did this intentionally at FLEM a few years ago by putting a 575 nm "yellow" laser next to a 520 nm green, and then next to a 635 nm red. The color we "saw" would magically shift right before our eyes based on what we compared it to!) As an aside, if you really want to go down the rabbit hole regarding color perception, this TED talk is a cool place to start.

    All that being said, if you start with the CIE 1931 color space chart and plug in 3 common wavelengths for red, green, and blue (635 nm, 520 nm, 445 nm), you'll find that the center of the white space of the color chart corresponds to a power ratio of about 3 Red : 1 Green : 4 Blue. (These numbers are estimated based on my memory, as I just discovered that I can't run Chroma at work to check these numbers because I can't install the Matlab library on my new desktop!)

    Of course, each time you change one of the wavelengths you have to run the calculation again, because the ratio will be different. So for the old school laserists who remember when your solid state choices were usually 660 nm red, 532 nm green, and 473 nm blue, you might have a calculated ideal power ration of around 5 Red : 1 Green : 3 Blue.

    The problem is that Red is expensive. True, Blue used to be expensive too, before the 445 nm diodes were a thing. But Red has always been expensive. So it was common to see commercial projectors that didn't have as much red as the calculations would suggest would be needed. Many of those early projectors were green-heavy. Later, when blue got cheap, you started to see a shift towards blue-heavy color balance in projectors. And today you still see that quite often. One nice thing is that a blue-heavy projector will often be capable of producing a piercing, "cool" white color, which is sometimes desirable.

    Over time, most people just got used to seeing projectors with less than the recommended level of red. "Good enough" really was good enough! To the point that when I first saw a projector that had the "correct" ratio, I thought it had too much red...

    The take away is this: at the end of the day, color is both a set of numbers on a label *AND* the subjective perception of the viewer. And when the two don't agree, the viewer's perception is the one that wins out. So if the crowd likes it, go with it and to hell with what the math says.

    Adam

  7. #7
    Join Date
    May 2009
    Location
    UCSB
    Posts
    715

    Default

    I think there is some confusion about what everyone is talking about. On the one hand there is the luminance of different wavelength sources (the ratio of how bright a specific color appears for its power, generally specified for an 'average non-dark-adjusted eyeball) and there is an independent perceptual effect 'Helmholtz–Kohlrausch effect' where saturated colors (pure red, green, or blue) appear brighter than nonsaturated colors (ie, grey/white) when they are placed next to each other.

    It is similar to asking: What ratio of R/G/B to I use to get brown? Brown is not a 'real' color--it is really a shade of orange (same way that grey is a shade of white), and the only way to project a brown laser beam is to project an orange laser beam at half brightness next to a white laser beam.

    The luminance of colors plays into the color balance that is used if you wanted the pure red/green/blue beams to appear the same brightness, which is related to the lumens/watt for a given color (below)

    If you are using sources at 640/532/455nm the lm/w values are 120/603/33 so after normalizing to green you expect a 4:1:18 R/G/B ratio to get equal brightness of the 3 lasers.

    That isn't usually what we aim for though, instead we aim to have a white beam when all 3 lasers are on at 100%, which works out to much less blue light. Side note: the discrepancy is a lot less significant at less extreme wavelengths, at 473nm the expected blue/green ratio drops to 6:1 for equal brightness vs 3:1 for neutral white--one of the reasons why 473nm laser shows usually 'look different than 455nm based shows--the pure blue is not as washed out compared to the red/green

    The color balance that is used in laser projectors to achieve a 'white' beam (ie, the usual 3:1:4 or similar) is instead calculated using a color space chart, which quantifies the relationship between wavelength, relative intensity ratio, and color.

    The the CIE 1931 RGB color space achieving white requires the usual ratio of about 3:1:4 depending on the white you are targeting. These calculations can't easily can be done by hand, but as mentioned earlier there is a utility called 'chroma' floating around on the forum that can help with those calculations.

    As best as I can tell, this discrepancy is related to the logarithmic response of our perception to brightness combined with the huge overlap between our green/red cones as shown in this plot of sensitivity vs wavelength for red/green/blue cones in the average eyeball



    This brings us to the Helmholtz–Kohlrausch effect which is also related to this discrepancy. The chart below shows pure red/green/blue (plus pink/yellow) of the same luminescence as the grey background. The red in particular stands out as being much brighter than the grey backgrond




    What this means is that when projecting a show with a mix of different colors (notably, white and red or blue) one would expect that a pure red beam should appear brighter than a white beam of same luminance (not of same power, that is a whole different story), depending on the surrounding colors.

    For the most part it seems like most shows that I have seen do not take much effort in trying to maintain constant luminance in their projection (that would require adjusting the laser brightness to compensate for varying scan speed within a frame, plus doing the color space calculations, correcting for persistent of vision effects, etc...). In my opinion with vector scanned type laser shows, the scene has so much contrast to begin with that I only perceive the color variations across the scene, and the brightness variations are largely lost on me. I think this is largely due to the fact that we perceive brightness logarithmically, but perceive color ratiometrically, so we are much more sensitive to small changes in color/hue than we are to brightness/luminance.
    Attached Thumbnails Attached Thumbnails Cones_response.png  


  8. #8
    Join Date
    Nov 2008
    Location
    Cleveland Ohio
    Posts
    2,599

    Default

    Nice summary. For me the ratio 4:1:2 is all I remember. 642:520:465 or 642:532:445 or 638:520:445 or any other similar combination BUT 660/642:520:462 is my favorite then it's more 6:1:3. I like a cooler white.

    532 gives nicer yellows. 462 nicer blues 660/642 nicer reds than either alone

    It's really subjective. And then there is BROWN. I have really tried to work with more subtle shading and frankly no one notices but me.

    The argon palette is different than the diode palette is different than the Ar/Kr palette etc... I try to make color combinations that lead to interesting sets like pastel, earth, neon, metallic....

  9. #9
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,489,446

    Default

    Quote Originally Posted by krazer View Post
    For the most part it seems like most shows that I have seen do not take much effort in trying to maintain constant luminance in their projection (that would require adjusting the laser brightness to compensate for varying scan speed within a frame, plus doing the color space calculations, correcting for persistent of vision effects, etc...)
    Some laser show software incorporates logic to try to reduce the luminance variation between simple and complex frames. The end result is subtle, but you can see the difference when you view it side-by-side with software that does not do this (or if someone with root access disables the feature on the product that has it).

    Bill Benner demonstrated this on the ILDA cruise in 2008 with Pangolin's LD2000 software. I remember being quite impressed, particularly because it was an impromptu demonstration that was completely unplanned. A bunch of senior laserists were just sitting around chatting about a wide range of topics and this issue came up when discussing using a diffraction grating to make a fan of beams vs using scanners. Bill drew the analogy to hidden auto-brightness code in LD2000. When people pressed him on this subject he decided to fire up LD2000 to show us. Believe me, the difference was readily apparent between "enabled" and "disabled", and there were more than a few people in the room who were surprised that this feature existed. (Alas, it's not something a standard user can change, at least as far as I'm aware, although with Beyond this may no longer be the case.)

    To the larger question of luminance, color balance, and the Helmholtz-Kohlrauch effect in particular, my opinion is that because of the large number of variables at play, even if you could understand each and every one of them you still won't arrive at a precise answer because of the subjective nature of our vision. As mentioned in the Ted talk I linked to above, visual information by itself is meaningless: It's only when you add context and experience that it actually takes on meaning and has value. I do agree, however, that general predictions can be made.

    Vision is really tricky. The example in the video of two identical shapes with the same spectral information representing two completely different objects is a powerful reminder of the massive part our brain plays in making sense of the visual world. The quote, "How do we see anything?" is poignant. But it's also the reason why people will always be able to argue about things like color balance or apparent brightness. And honestly, I don't mind that - provided everyone at the outset understands that it's OK to disagree about something that is subjective.

    I'm thankful that this topic came up, however, as I had never before heard of the Helmholtz-Kohlrauch effect, and it's been a pleasant trip down the wikipedia rabbit hole.

    Adam

  10. #10
    Join Date
    Jul 2019
    Posts
    19

    Default

    Great information. Maybe a bit complicated for me. I'm not sure we addressed the Helmholtz-Kohlrauch effect or maybe I didn't understand the part of the text which did.
    To reference the wikipedia page, "LEDs in the dashboard and instrument lighting are designed for use in mesopic luminance. In studies, it has been found that red LEDs appear brighter than green LEDs, which means that a driver would be able to see red light more intense thus more alerting the green lights when driving at night.". This info corresponds with the data from Viewsonic. Red LED appears brighter than green, blue also. https://www.viewsonic.com/de/product...si_lumen-3.jpg

Posting Permissions

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