AVS Forum banner

Status
Not open for further replies.
1 - 8 of 8 Posts

·
Registered
Joined
·
398 Posts
Discussion Starter · #1 ·
I've been trying the Guy Kuo solar cell method for grayscale calibration of my projector. While the solar cells are actually quite precise, my main problem was getting an accurate 6500k white from the PJ to start with. Visually (subjectively) comparing the projected white to a gray card, or 6500k fluorescent tube (Tom Morrow's method) was difficult for me and seemed to have too much potential for error. Starting the calibration with a more accurate - yet still inexpensive:) - objective white measurement would be better.


This is my idea.


Materials -

A. Primary Green #139, Primary Red #106, and Dark Blue #119 Gel Sheet color filters from filmtools.com $15.75

B. Solar cell from Radio Shack as per Guy Kuo's thread. - actually I used 4 of them wired in parallel to get a little better accuracy in low light. ~$20.00

C. 6500k 18" reference tube and fixture as per Tom Morrow's thread. ~$25.00

D. ~10" diameter cardboard tube painted black - I used a left over piece of a concrete post hole form from Home Depot. A smaller housing or spare lens cap would work just as well.




1. Attach wires and mount all 4 solar cells inside the black cardboard tube or lens cap.


2. Place a colored filter over the open tube and shine the 6500k light through the filter and onto the solar cells. With a digital multimeter, record the solar cell current for the Red, Blue, and Green colors using this 6500k source. Obviously, make sure the tube, filters, and light source are in the same position for each measurement. Using these particular filter colors, this yields the correct ratio of red, green, and blue current to match 6500k.


Interestingly, the Green/Blue ratio was an even .67 (2/3) and the Green/Red ratio was .51 (1/2)


3. Now place the cardboard tube in front of the center of the screen facing the projector. Project a 100 IRE pattern, then repeat the current measurements using the same Red, Blue, and Green Gel filters. Adjust the Red and Blue gain on the PJ so that the RATIO of Green/Blue and Green/Red currents is equal to the corresponding ratios from the 6500k tube measurements in step 2. When those ratios exactly match, the PJ should also be projecting 6500k. Keep in mind the color filters do not exactly match the PJ red, blue, and green CRT colors, so adjusting the gain of one gun may slightly affect the current measurement of the other 2 filter colors.


4. After 6500k white has been set, perform the grayscale calibration per Guy Kuo's thread


Mike
 

·
Premium Member
Joined
·
4,892 Posts
Well Mike, this sounds like the "way to go". I was concerned about trusting my eyes to the 6500K comparison (I'm color blind!) Any comments?


About the only thing I can imagine as a source for error is the fact that the light from the 6500K source contains all color components and the filter may not block the other colors well, whereas the individual PJ lenses should be pretty pure. Since you're just dealing with ratios maybe it's ok.


Cary
 

·
Registered
Joined
·
398 Posts
Discussion Starter · #3 ·
Hi Cary,


I find the solar cells to be more precise than "eyeballing it". Repeat measurements with Red and Green were within 1%. Blue was within 2%. Part of that error may be in the 6500k light itself.


Yes, the exact colors of the CRTs are irrelevant for this purpose as you are only looking at the combined "white" projector output.


Even the gel filters do not necessarily need to be R,G, and B - just so long as all colors are represented. Using these primary colors probably makes it a little more accurate. And using colors that are close to the CRT colors makes the CRT gain adjustments a little easier.


Mike
 

·
Registered
Joined
·
4,916 Posts
Mike,


Been meaning to call you for a week. How did everything go. I called you wife when you were in the hospital and she said it went well.

Now here are my comments on your method, and I'm just thinking out loud.

In measuring CRT's I have found that the primary colors of the three tubes will differ from manufacturer to manufacturer as far a where they are on the Gamut.

So my question to you has to do with your choice of color filters and why you chose them? Are they representative of the poits on the gamut that the CIE set up for CRT's and if so, how do you know that the CRT's you have in your PJ have those primary colors? Like I said, I have measured significant differences between MEC, Sony and NEC CRT primary colors.

Throw into that color filtered lenses that some manufacturers put on to try and get the primany colors where they should be and the mix gets even more confusing.

think you concept has merit and should be investigated more. I am going to be getting an updated sensor in the next few weeks for Colorfacts and will have to travel up there to see and compare what you are doing so we can see if there is any accuracy. Or you need to come down here and we can go through a bunch of different PJ's and compare results.

Give me a call.


Terry
 

·
Registered
Joined
·
398 Posts
Discussion Starter · #5 ·
Hi Terry,


I've been meaning to call you too. I'm fine now, although 20 lbs lighter -don't ask. I can't decide which was worse, major surgery or spending a week with basic cable and an old magnavox 19" TV. (Tried and tried to calibrate that thing but just couldn't get anywhere):)


The CRT colors are not relevant for this measurement. Only the combined white is "seen" by the sensor/filter device. The originating tube colors could be yellow, cyan, magenta, or whatever as far as it knows. So long as the light eventually hitting the sensor is identical to the 6500k reference, it should be accurate. I believe some colorimeters work in exactly this way - only using a prism to separate the colors instead of filters.


I chose primary red, blue, and green just to keep things as simple as possibe.


I'd like to come down when you get your new sensor. First, it would be interesting to see how accurate the 6500k 18" tube really is. Also, maybe we could do a grayscale start to finish with each and compare the end results.


Mike
 

·
Registered
Joined
·
87 Posts
Are you placing the solar cells on the screen? interesting, I've been using the solar cell on the tube, for greyscale tracking.


But for 6500k this looks like the way to go. The filters need to exactly complimenting each other, though.


BTW do you use white screen or white window, as a CRT is more accurate (brighter) with a white window, at least that is what they say.


So my next purchage will be a 6500k >90 CRI tube and a set of color filters :)


Soon we will have a color meter kit.
 

·
Registered
Joined
·
398 Posts
Discussion Starter · #8 ·
The solar cells are near the screen but still facing toward the PJ. It would be possible to do the whole calibration reading off the screen itself, but the solar cells are not sensitive enough when measuring the low IREs.


Yes, I'm using small windows on black background.


I found this site where someone is already selling a similar kit for certain digital PJs.

http://www.enhancedht.com/Merchant2/...ategory_Code=S


I've ordered the photocell and slide filters to compare with the solar cell method and will report later.


Mike
 
1 - 8 of 8 Posts
Status
Not open for further replies.
Top