Ambient Light screen development thread.. - Page 5

Steve,

looks promising indeed. the only issue i see is something that you also eluded to in an earlier post. basically, we know that the RGB has different wavelengths. so you might want to figure out a dot pattern that would reflect that. in other words RGBW followed by RGBW may not be the right pattern needed. i could be RRWGGWBW or whatever other pattern produces the amount of dots needed based on the intensity of the wavelength.

i also agree with biglyle in that this goes way beyond the feasibility of most DIY'ers. if the application doesn't include a pole with a roller on the end of it

i have to admit i'm very intrigued by it all and i can't believe how much we've accomplished as a whole in the past couple months... we've taken this DIY forum to a whole new level of technology.

With my approach, each individual pixel on the screen is subdivided into 1/4 Red, 1/4 Green, 1/4 Blue and 1/4 white. (I added the white to boost the brightness of the image, because the first one I did was only RGB, and boy was it dark...) So, if you have a pure Red pixel, it is only reflecting from a quarter of the pixel (only the red dots are "turned on" while the other dots absorb, or are turned off). Different colors will give different degrees on "on" or "off" dots over the range of red, green and blue. I thought about boosting the brightness by adding even more white dots into the matrix, but the more white that is added, the more ambient light will be an issue.

It seems this is a bit of a drawback of the approach--and where the Chromavue or similar screens should blow this out of the water--apparently, theirs is a film combination, where each color, R, G or B, is completely reflected over the whole pixel, preserving the brightness of the pixel.

So my approach might be valid for a light cannon like my Benq 6100--but for lower lumen projectors, the image will definitely suffer.


So, without taking the optical film approach that Sony and others have done to get their black screens, is there a way to lay pure color reflectance of the particular wavelengths without absorbing any of these wavelengths?

Steve,

my thinking... is that you might consider using 2 separate pixel patterns and then alternating them.

edit. consider alternating a Y into your pixel in place of the GB. and then alternating that new pixel with your RGBW pixel.

PBMaxx--so the pixel pattern you are thinking would be:

RGBWRYW... ?

Any particular reason why?

absolutely. your original pattern should look like this.

RW
BG

green needs to opposite of red and blue needs to opposite of white for it to be an optimum pixel. (in my subjective opinion)

you stated that adding addition white between each color and the screen quickly begins to lose it's ambient qualities. this is because if you were to add a white in between each color you'd be increasing your white 'four-fold'.

in my subjective opinion the eye visually perceives the primary colors as being more 'visually' appealing then RGB and since big lyle has shown that yellow could or should play a role... my suggestion was to add Y to an alternating pixel.

so in other words, rather than increasing the white 'four-old' by adding a white in between each color... you instead increase the vibrancy about 25% since it is replacing one of the GB dots.

think of the GB in the same way as you think of lamp black in a bombadil grey mix... it's needed... but too much will only 'mute' visual quality of the entire screen.

Actually, one thing I should clarify is that each pattern is disposed multiple times throughout each pixel--in other words, each pixel doesn't just have one pattern of RGBW, but many, many repeated patterns, because the dots are so small.

I could try to estimate the size of a pixel on the screen, and get the square dots large enough so that each individual four dot pattern fits roughly in each pixel space... Hmmm....

Steve,

Just a thought.

I often use a small microscope from Radio Shack for different things I do. They have a small rectangular that is 60X to 100X that sells for around $10. It uses a small flashlight bulb to illuminate the area of interest and I assure you it will tell the tale about your printed pixels. You will see up close if there is any blending of colors etc..

I'm pretty sure that the latest sample (the one that I sent you) is doing pretty well with respect to blending of the image. I have pretty good eyes for extreme close-ups, and I can tell that the three colors are definitely there, without much blending.

Thanks for sending me the bit map so quickly. I downloaded the bmp and printed it. As soon as it drys I will put it up on my Dalite HCCV screen and shine my PB6200 on it.

Also I will put it under the microscope to see how my cheapo printer reproduced it. Since you have a photo quality printer I am sure your prints are much better.

Just out of curiosity I may print on a transparency and paint some SM on half and pearl on the other half of the back side. Who knows it might make a difference.

Steve,
This sounds interesting. Could you send me the file?

Ericglo

Steve,

I printed on both glossy and transparency. Placed both on my HCCV screen and the results were, as far as I can see identical to yours. It held up well under ambient light but was far too dark.

As I suspected my printer just isn't up to par.

Under 30X the Red looked the most uniform in coverage with very little blending into the adjoining squares. Since my printer uses CMY it is certain that the fidelity isn't there as it should be with a photo grade printer equipped with dedicated RGB cartridges.

The blue is the second best with the green being very speckled which reveals a high percentage of yellow underneath. It is like yellow squares with green dots on top.

I am now going to try painting the backside of the transparency with Pearl just for grins, we will see what happens.

Steve,
Try to make a grey as a percentage of black that is as similar to your RBG matrix under full spectrum light (or under a regular light bulb). This would be a perfect control to see if the RGB matrix is doing anything more than just grey. See if there is any difference between the matrix and a true grey. Also, if your projector allows it, project just through one filter (R, G, or B) at a time and see if the matrix appears lighter then the grey, as it should if really does have any ambient light benefits.

In an attempt to make something rollable that doesn't use UPW to achieve strong ambient light qualities, I rolled a panel with a slightly different approach that seems to get us close to that goal:

2oz Delta Gleams Silver Metallic
2oz Behr Deep Base
1oz Minwax Polycrylic Satin

I first rolled a panel with a coat of pure Silver Metallic, then rolled 3 coats of the above mixture on top -- and BAM -- up against a UPW matte screen it gives us great black levels, minimal white crush, and a strong contrasty image particularly with a ton of daylight shining on it. In controlled lighting, it looks similar to my previous mix that I posted, and better than UPW alone.

Its not perfect -- there is a slight bit of viewing 'cone', due to the large reliance on silver, but its minimal. In very bright white scenes, there are subtle roller marks visible -- but otherwise it works pretty well for a first try.

Perhaps biglyle and others have something that is improved over this mix-- I suspect they do -- but in the absense of any other 'posted' formulas I would recommend anyone looking or waiting for a starting point to dabble with to try this out.

Next steps? Its just a bit too silvery I think -- I want to lighten it up a touch - probably with a bit of pearl or the 'dreaded upw'. That may also make it more even for rolling a big screen. Perhaps a 'hybrid' of this and the last formula I posted. I'm running out of time before turkey day. Happy painting!

I did some experimenting last night, and, with my projector, went into the service menu, where they have "Blue curtain", "Red Curtain" and "Green Curtain" (which, as far as I understand, only shows the pure colors at once.

I then tried to match up the colors, in total darkness with some large printed dots. I iterated the colors until the color from the projector most closely matched the color on the page--which gave me a very orange red, a very yellow green, and kind of a cyan color. I then made a color matrix out of these three colors. While the image was slightly improved in ambient light, it cast a very green sheen over the entire image contained within the matrix (it must have been a pretty extreme green color, because I am red/green color blind and I could see it pretty well (perhaps I'm not the best one to be doing these experiments...))

So, back to the drawing board...

Steve Scherrer,

don't give up guys. looks like you got something good going on here.

-maxx

Biglyle do you have a release date set for your mix, I thought the pictures looked very good and I will soon be at the point where it is time to paint my screen. I do not want to rush you as I know you are trying to perfect it, just curious to when it will be released. If you do not recall I to have the H31 Optoma and that is why I am so intrested in your mix.
Thanks for all the hard work you guys are putting into all of this.

I tested a bunch of samples today, I significantly changed my color matrix, and added yellow to the mix as well. What I found is that my original RGB-white color matrix was, by far, the best of all--which surprised me a lot.

I made a bigger sample of the RGB-white and placed on my screen:

This one shows two samples of the RGB-white on different parts of the screen:


This one shows a closer-up version:



From HDNet, with my canned lighting about half to 3/4 brightness:



With canned lighting at full power:


Another shot:



Of course, this color matrix is merely printed on high resolution paper, so it is very crude to begin with. In the above pictures, the single page set off to the left is a gray scale print of the RGB-white color matrix that I put up for comparison's sake. It is not as apparent, but the color matrix does provide for a better image than the grayscale version.

I don't know if others can proceed with this, but it seems that I probably have gone as far as I can right now. It would be interesting if this concept could be incorporated into some other concepts floating around the boards right now. I have given the color matrix out to a few, so maybe others will come up with something...

Steve,
Thanks for the initiative. I have more questions and suggestions for you. You said the gray sample is a grayscale printout of the RGB matrix, but in "normal light" do they seem like similar grays from a distance (or by unfocusing your eyes when you look at them)? Does the matrix appear to have a color push?

The reason I ask is because your approach (so valid in theory that I tried it myself about 2 weeks ago) needs a fundamental question answered: Is the improvement coming from the grayness of the matrix or from the colors within the matrix?

Any scientist or statistician knows that to best see the effect of one variable, you should make every other variable consistent. So in this case-- gray vs. matrix-- if you can make everything else the same in the printouts-- gray "value", hue (color push), printout size-- and equalize the projector factors-- image on screen, location of printout from center, etc.-- then you can better answer the question: Is the matrix doing anything that the gray is not?

In your pictures, only the last one has the grayscale and the matrix under similar images from the pj, so is the only one that can really be used for comparison. Even then, on the background screen, the face seems "warmer" on the right than the left, just as it seems "warmer" on the matrix than on the grayscale.

Also how is gray vs. matrix under the single color "curtains" of your projector?

You think you've gone as far as you can? HA! We all know that when it comes to DIY there's ALWAYS more that can be done!

You're absolutely right--I'm sure I've got a long way to go. I felt a little like the project was starting to consume me a little bit, in that I spent all of last night working on it, and felt a little burned out. But I slept on it, and I already have some new ideas for this... (I try to get out, and it keeps pulling me right back in!)

But I suppose I did answer a few questions:

The grayscale matrix on the left in those images was about the same tone (is that the right word?) as the color matrix from a distance (The grayscale did look to be about the same as the color--perhaps just a wee bit darker than the color matrix).I believe the color matrix is providing some benefit as opposed to just the gray--colors on the screen were a bit richer. I also tested it using the white curtain, red curtain, blue curtain, and green curtain. Each one seemed to show an effect using the color as opposed to the grayscale. It wasn't huge, by any stretch of the imagination, but it did seem present. I would like to try some things to enhance the color contribution from the matrix, and have some ideas on that. (Would optical brighteners help, or even fluorescence?) Anyway, I will try to put up some images of the comparison using a large sheet of gray and a large sheet of the color matrix in the same place of a projected image, but I'm afraid that the distinction or effect of the color matrix may be too subtle, and not adequately captured by my camera. I also have to rely on the color expertise of my wife, a bit, since I have a red/green color deficiency in my own eyes. (I can still see red and green, certain shades are tough to make out, that's all...)

The other color samples I tried--I varied the amount of white, red, green, blue, and added some yellow to some samples. I did not get a very good image with any of these samples--as I stated in my previous post. The ones with the yellow gave me a very distinct yellow/green push on the image--it was not good at all. I also tried a sample that was very red and, as you might expect, gave a very distinct red push from in the projected image.

Steve,
I tried this matrix you sent me and it is way to dark. I have no way of measuring gain, but this is real low probably .2 or .3. I tried to figure out a way to keep each individual color, but lighten them considerably. I had no success with this approach though. I am sure there is a way to do it (maybe another program), but I couldn't figure it out. I think if there is a way to lighten the color that this might be successful, if you can keep the gain around .8 to 1.0.

Ericglo

Yeah, I agree--it is way too dark. The first sample that I tried had only RGB dots, so I lightened up the image by adding a white dot after each RGB pattern. This definitely lightened things up a bit.

That's why I posted in the previous thread that it would be nice to brighten up the image by using, possibly, optical brighteners, or possibly fluorescence?--something that won't suck up the light as much.

A lot probably has to do with the printer used to print the inks as well--I only make the matrix--I would be surprised if anyone who had the matrix were getting the same colors as anyone else when printed.

Steve,
I appreciate your efforts, and I'm sure other do too. This is great stuff.
Anyone know of a printer that can print in RGB? CMYK is designed to absorb light. (Cyan absorbs red light, etc...) A printer with RGB ink would, in theory, be twice as bright.

Really? I am certainly no expert on this stuff (and I'm not sure that my approach is even working...), but does it matter what the base ink combinations are? Even if CMYK is used, shouldn't the ink work in the same way, as long as the CMYK is combined together to give RGB dots?

Steve,

My understanding of RGB is that each of the colour values changes based on the color that is to be reproduced.

The range being 0 to 255 for each color.

So for example a dark green would be R 0 G 100 R 0
while a bright green could be R 0 G 255 R 0.

Would trying to print a fixed color value for each of RGB only optomize a vary narrow range of each of the RGB colors?
i.e. R 200 G 200 B 200

There is a RGB value color table with the numbers for various grays which, when each color value is equal, produces a range of neutral greys.

Even if was to find the optimum color values for each of RGB, wouldn't the dot size have to match exactly the projected dot?

Peter

Others can correct me if I am wrong, but as I understand it, you want each dot to match as precisely as possible the particular wavelength of solid green from the projector, solid red and solid blue--so that you get the maximum reflection of that color from the projector. Since any one pixel is comprised of a combination of only the three main colors, the screen itself has a number of dots of the three colors contained in the equivalent pixel space (I believe I am using samples that are about 40-50 dots per inch or so)--and the individual pixel, whatever color it is, reflects the equivalant proportions of colors off of the individual dots.

That's the theory, anyway.

I've been reading these high-ambient-light DIY screen threads with quite a bit of interest. Our new house is still several months in the future, and I'm looking to project a pretty huge image -- in the range of 166" diagonal (4:3). Clearly it would be great if I didn't have to buy a screen.

I have one comment to throw into the mix which may or may not be helpful pursuant to your commentary regarding RGB mixtures. While equal RGB levels on a screen do produce a continuous grayscale, standard procedure for converting an IMAGE to grayscale uses the following formula:

red * .29
green * .60
blue *.11

Those are the percentage contributions of each wavelength to the overall luminosity of the image. (Many years ago, when graphics software was rare, I wrote software that performed image transforms that are simple by today's standards, but were pretty interesting at the time.) Of course, not all images are equal, and the lightsource used to produce the image has a tremendous effect on what actual percentages apply to produce a grayscale which has the correct luminosity and tone, but the percentages listed above are the generic rule of thumb I have always encountered. I believe these are the same percentages Photo Shop uses for it's built-in grayscale conversion, actually.

I'm not entirely sure how this might help, but I figured I'd throw it out there and see if you guys might find a use for it.

Interesting--and it has me thinking...

For anyone still interested, here is my finished mix in 2 version, one for light controlled rooms, and one for ambient light. Let me know if they work for you.

ambient mix

1/2 cup minwax varnish (satin)
1/2 cup deep base
1/2 cup UPW
1/2 cup behr WOP or delta Pearl
1/2 cup Delta Silver Metallic
1/4 cup Folk Art Sterling silver Metallic
1/4 cup Folk Art metallic aluminum
15 drops Folk art bright red metallic
15 drops delta bright red - transparent
6 drops delta phalo green - transparent
3 drops delta phalo blue - transparent
3 drops flok art metallic amethyst
15 drops folk art metallic inca gold
15 drops delta yellow - transparent


Light controlled room mix

1/2 cup minwax varnish (satin)
1/2 cup deep base
1/2 cup UPW
1/2 cup behr WOP or delta Pearl
1/2 cup Delta Silver Metallic
1/8 cup Folk Art Sterling silver Metallic
10 drops Folk art bright red metallic
10 drops delta bright red - transparent
4 drops delta phalo green - transparent
2 drops delta phalo blue - transparent
2 drops flok art metallic amethyst
10 drops folk art metallic inca gold
10 drops delta yellow - transparent

Be very careful when adding the drops of the colors so as to not get big blobs, the metallic colors are quite thick, so slowly let the drops fall and keep count. This will give a very neutral color grey. The darkness of the screen can be adjusted by adding more or less of the darker "silvers" (aluminum and the sterling)
I have done all my testing using an Optoma H31 with "rolled on" screens. I have no doubt that this would show even better results when sprayed, especially if sprayed on a very shiney surface such as a mirror (props to MMAN).

Good luck

biglyle, can you try something out for me? Mix up 20 parts of the Folk Art Metallic Aluminum, 10 parts of WOP, 2.5 parts of UPW and a few drops of the Metallic Inca Gold. Just curious to see how it looks up against your mix.

Biglyle -- Thanks for all the hard work! From reading your list I have no doubt it is a great mix! Everyone should rush out and give it a try.

I just finished painting my latest test mix and it too includes some of the folk-art metallics which I've found work great--they are just a little less translucent than delta. I have not gotten into using the pigments to correct the colors just yet.

My only issue is still getting a nice smooth roll on a large screen --- I'm still having trouble eliminating roller marks. Can you share how you are rolling? Are you using the 1/4" nap or ? Horizontal or Vertical? What's the basecoat?

Best