Dimming curve question

Santacarl

Active member
Hey All...

After reading the thread in the Nutcracker section and as I watched John's videos....At one point he mentioned setting a low intensity level (~20% if memory serves) for videos on his matrix to prevent washing out of colors.

It got me to thinking.....and yeah I know Joe's probably thinging....rut roh there he goes again.....

I use dimming curves on my pixels.....I set each color level and incorporate all 3 color dimming curves into an overall RGB dimming curve. But I've noticed that any time there is a lot of white in one of my displayed graphics (think santa's beard) the white just overpowers any other color...especially those bordering the white...(think the rosy cheeks of Santa's face being washed out by the nearby sea of bright white in the beard.)

There is no place to adjust the dimming curve for all white...at least not that I can find....So, naturally, while the RGB dimming curve takes care of mixing the 3 colors.....there is no way to 'dim down' the pure white created when the 3 colors all come on at the same time to create white....

To my way of thinking.....the white just overpowers the individual colors that are competing for viewability at the same time...

Am I thinking of this correctly? Is there a cure?

Thanks

SC
 
Yeah Joe...I understand that....but pure white is sooooo bright that it seems to overpower the other colors....it's not the color I'm talking about...it's the intensity/brightness that I seem to be struggling with... Maybe when you combine all 3 @100% you just get the color white and that has noting to do with the intensity...I was just thinking that one of the main benefits to a dimming curve is that it allows you to adjust the brightness of a given color so that when you look at a green and blue they both appear as having the same brightness.... I know if I don't adjust mine and I turn on a blue and a green at same time the green APPEARS brighter......and I was trying to think up a way to adjust the whites so they too 'blend' and appear of the same intensity as the other colors to the eye....
 
I would take your graphic and adjust the whites to some level of grey. A little trial and error might give the results your looking for.

As for the appearance of green being brighter, it's not just appearance. Take a look at the specs I pulled off of aliexpress. They are pretty much the same for the ws2811 bullet nodes.

Red = 3.3 Lumens/300 mcd, Green = 10 Lumens/1000 mcd, Blue = 2.7 Lumens/300 mcd

In addition, the human eye is much more susceptible to green than red or blue. So we perceive an increase in green as an increase in brightness much more than red or blue.

I think Joe was saying if you want white to be dimmer, use 50% for RG&B to make a 50% lighter white, of course it doesn't quite work that way, but with some messing around you can come up with a better white (dimmer) to suit your needs. This won't apply to graphics as you can't isolate the white in a graphic to just change white, which is why I suggested turning white to gray in your graphic.
 
Dimming curves are there to match illumination levels for Red, Blue, Green and then modify the proportion to get secondary colors like yellow.

The problem you are having with White is not solved using a dimming curve. YOU solve it by picking the WHITE intensity.

White is so bright because you have all three leds on. As gadget stated -- you might want to try White at 50%, or some other illumination level, to get the intensity to match the pure Red, Green, and Blue lights nearby.

Joe
 
I would take your graphic and adjust the whites to some level of grey. A little trial and error might give the results your looking for.

Take a look at the specs I pulled off of aliexpress.

Red = 3.3 Lumens/300 mcd, Green = 10 Lumens/1000 mcd, Blue = 2.7 Lumens/300 mcd

In addition, the human eye is much more susceptible to green than red or blue. So we perceive an increase in green as an increase in brightness much more than red or blue.

Yeah....that's why I use dimming curves...I had to cut the green back considerably to get it's brightness close to the Red and Blue.... The way I did it was to take the weakest (less bright) color and see when it quit getting brighter.... That is my highest and baseline number for all 3. I took the other 2 colors and dialed them up until they matched the brightness of the less of the baseline color...as it turns out my green now stops increasing in intensity long before it would if left unchecked. When I turn on a R, G, and B at the same time they all appear the same in brightness.

But that does make me wonder.....going back to Joe's statement of 100% of all 3 colors=100% white.... In theory, if all 3 nodes are of the same relative brightness you would expect 'pure' white as a result.....but if green has more prominence and you bring in all 3 colors @100% wouldn't that, based on the specs you gave for pixels, make the resulting white have a slight greenish tint?
 
You keep talking as if the world is linear -- it is NOT.

Your eyes do not respond linearly to different colors -- they are most sensitive to RED. That is why when I flew jets for the Marine Corp back in the VietNam days -- lights in the cockpit were RED. That is also why street lights are RED for STOP -- you see RED better than every other color -- hence RED to STOP!!!.

Do research on what studies were done to produce good looking color TV and Monitors --- there is a ton of level adjustments per RGB color depending on intensity to make you, the User, have a nice picture on the screen.

With RGB pixels -- and you building matrices and MegaTrees and looking to have secondary colors and matching intensities -- you are back in the 1950's all over again trying to figure out how to manipulate led intensities for good looking displays.

RGB pixels are to the point were you ask for a color -- and it might give you that color (think secondary colors, e.g yellow, purple, orange). There is a lot of work left to do to get pixels next to each other to have different colors AND similar "PERCEIVED" illumination levels (WHITE being the worst). The eyes do not respond in a linear fashion to color as stated above.

The world is one big non-linear mess and little old you is trying his best to make it fit into a linear one ... don't try too hard -- you will get a head ache:)
 
There is no such thing as pure white either. Warm white contains a lot of red. Daylight contains a lot of blue. Really it just comes down to finding something you like and sticking with it.

I can see what your doing with the r g b dimming curves, but is that messing you up when you try and make something white? The r g b might be a different mix than the blend you get from your dimming curves?
 
Thanks Joe....that's helpful. I am starting to get my arms around the non linear piece of this puzzle. I think functionally and I suppose that nudges me in the direction of linear thought. Too much time spent in the military following those accursed checklists maybe! :wink2:

I know you jet jocks had checklists too but you guys didn't always have time to reference them which is a good 'thang' when you're traveling mach + :eek:mg:

Anyhow I had them force fed to me and I learned to rely on a step by step approach and was 'trained' to be dumb as a bag of rocks and to not get inquisitive and get out of the sandbox. :biggrin2:

That haunts me in this hobby, let me tell ya.....

I really liked your analogy on old TV technology....in a former life I used to repair TVs and in those days we had to degausse and set screen intensity levels to get the RGB guns in the CRT balanced.... I can't tell you the number of sets I readjusted after the owner had set his own screen levels! Maybe that's where I got the obsession to balance colors from! :hmm:

I'll take your advice and not hurt my head over this issue....It would appear it is beyond our current capability.

You keep talking as if the world is linear -- it is NOT.


The world is one big non-linear mess and little old you is trying his best to make it fit into a linear one ... don't try too hard -- you will get a head ache:)
 
I can see what your doing with the r g b dimming curves, but is that messing you up when you try and make something white? The r g b might be a different mix than the blend you get from your dimming curves?

It doesn't seem to.....I get a bright (pure) white and not some off white like warm white.....I don't know if the dimming curves are influencing the color of my white but I do wonder if the dimming curve still works when white is involved...and if it does wouldn't it make sense that an equal blend (brightness) of all 3 colors would by definition make pure white? Otherwise the green which is apparently so strong based on the post above would be giving me a greenish 'flavor' tint wouldn't it?
 
If you do photography, I found for vidwo on pixels are best worked backwards. Are there any incandescent fixtures in the display? Mixing color temperrature will make most LED displays appear cold and blue completely lacking in red. The Pure wite is a cold cool white fluroscent color.

How this applies is to use a digital camera and set it for a color temperature of your reference white. In one of the LED strings I use, white to match white is 256 red,50 green and 20 blue. Cutting way back on green and blue stopped the washout and made the string match traditional mini lights even in red. Hope this helps.
 
Last edited:
It doesn't seem to.....I get a bright (pure) white and not some off white like warm white.....I don't know if the dimming curves are influencing the color of my white but I do wonder if the dimming curve still works when white is involved...
Not sure there. Joe can probably tell you if curves are applied when white is selected as a color. Even then the curves you've made to make R G and B more similar in brightness may make no difference at all when making white.

and if it does wouldn't it make sense that an equal blend (brightness) of all 3 colors would by definition make pure white? Otherwise the green which is apparently so strong based on the post above would be giving me a greenish 'flavor' tint wouldn't it?
Remember, white is not a color, it's an interpretation from our brain from the three cones in the back of our eyes. These three types of cones respond to wavelengths of light at (you guessed it) the red, green and blue spectrum, but they do so at different levels. For arguments sake, we can call R,G,B of 255, 255, 255 "pure white". Of course this doesn't really mean anything until you define the wavelengths of R G and B that you are producing, and in what intensities they are emitting at. Then they need to be compared to the levels of how the eye reacts to each of those wavelengths and intensities.

In our hobby, dimming and changing RGB outputs using some form of PWM (WS2811 chip for example), creates changes in voltage, current, and junction temperature, all of which effect the spectrum of light an LED will produce. We know we can't just put in 128,128,128 and expect to get the same white at 1/2 the intensity of 255,255,255.

As a very practical example, look at the values of R,G,B that Technician has given for his LED white to imitate a white incandescent light. 256, 50, and 20.


Here is some literature you might find interesting. I'm by no means an expert, but understand some of the relationships that exist with light, led's and how they all intermingle regarding our devices.

http://www.physicsclassroom.com/class/light/Lesson-2/Visible-Light-and-the-Eye-s-Response

http://light-measurement.com/spectral-sensitivity-of-eye/

http://www.beercolor.com/color_basics1.htm (yup. color of beer. you simply can't argue the information from this website! :) lol.)
 
Washed out color got me to thinking from my old tv repair days as video systems are not linear. Pixel dimming is linear as they are PWM directly to their DMX value. For more info on the curves for correcting for the non linear video to properly display it, a google search on Gamma Correction may be helpful. I'm including a link to a Wikipedia article on the subject.
http://en.wikipedia.org/wiki/Gamma_correction

The diffeerence between brightness between pixels and a CRT disply is shown in the graph below.
600px-GammaFunctionGraph.svg.png

Using a Gamma Correction for your video will drastically darken the majority of the videw resulting in less washout and vivid color. while preserviing peak white.
 
Last edited:
Washed out color got me to thinking from my old tv repair days as video systems are not linear. Pixel dimming is linear as they are PWM directly to their DMX value. For more info on the curves for correcting for the non linear video to properly display it, a google search on Gamma Correction may be helpful. I'm including a link to a Wikipedia article on the subject.
http://en.wikipedia.org/wiki/Gamma_correction

The diffeerence between brightness between pixels and a CRT disply is shown in the graph below.
View attachment 28422

Using a Gamma Correction for your video will drastically darken the majority of the videw resulting in less washout and vivid color. while preserviing peak white.

Gamma correction....yikes....that sounds pretty overwhelming.... How does one go about adjusting the gamma and is more or less a trial and error process?
 
If you do photography, I found for vidwo on pixels are best worked backwards. Are there any incandescent fixtures in the display? Mixing color temperrature will make most LED displays appear cold and blue completely lacking in red. The Pure wite is a cold cool white fluroscent color.

How this applies is to use a digital camera and set it for a color temperature of your reference white. In one of the LED strings I use, white to match white is 256 red,50 green and 20 blue. Cutting way back on green and blue stopped the washout and made the string match traditional mini lights even in red. Hope this helps.

No incadescents....everything is LED and pixel..... So how do you "set a digital camera for a color temperature"? I don't take a lot of pictures of the display because the never come out looking good.....poor color quality usually...
 
To take good pictures of lights requires all manual camera settings. Auto will see most of the scene as dark and try to adjust for it blowing out the lights. As stated most LED lights are not color balanced for incandescent replacement. Under very low light most auto settings on cameras assume fireworks, candles, incandescent light sources so photos tend to severly lack red. Start with a manual light source setting of fluroscent cool white or daylight.

If your camera has a manual white balance setting, turn on a box or pile of pixels all white and use that as the white reference to set color balance.

If your camera has a histogram view, learn how to use it for setting exposure so lights are not overexposed.

Adjusting Gamma for video on a matrix involves either setting dimmer curves, or using a video editor to change the gamma of the video to play on a Linear display. Either will work. This will return the video to looking more normal instead of bright advertising sign video.
 
Last edited:
Back
Top