I'd like to see a set of objective measures so that we can evaluate different designs and changes to the currrent one. A set of images and methods for evaluating the images would be a big step. I'm collecting images from various related web sites. See here. Brain could you set up a pinned thread for this topic?

Brain, thank you for making this a pinned topic.

It is my intention in starting this topic to provide a forum for the objective evaluation of the image quality of our DIY video projectors. This will help us to evaluate the step-wise refinement of Brain’s baseline design that is an ongoing part of participating in the Lumenlab Forum. As can be seen from the discussion started by Jones Rush here there are many factors to be considered. It is my hope that this topic will spur greater knowledge and understanding of the factors that contribute and detract from the image quality of our projectors.

In this early post I will try to compile with ongoing edits a definition of terms. One term that appears regularly when discussing light sources is color temperature. The following text about color temperature quotes from the work of Mort Abramowitz, a legend in the field of microscopy.

Color Temperature - “The concept of color temperature is based on the relationship between the temperature and radiation emitted by a theoretical standardized material, termed a black body radiator, cooled down to a state in which all molecular motion has ceased. Hypothetically, at cessation of all molecular motion, the temperature is described as being at absolute zero or 0 Kelvin, which is equal to -273 degrees Celsius.”

“Therefore, we can define the color temperature of a light source as the value of the absolute temperature of a black body radiator when the radiator color spectrum, or chromaticity, matches that of the light source. In the case of fluorescent lamps, which can only approximate the chromaticity of a black body, the corrected term correlated color temperature is applied through a calculated chromaticity value.”

“There are two important points to consider when examining color temperature phenomena. The color temperature value of a light source refers only to the visual appearance of the source, but does not necessarily describe the effect this source will have on photographs or digital images. Also, color temperature does not take into consideration the spectral distribution of a visible light source. In cases where a light source, such as a fluorescent lamp, arc-discharge burner , laser, or gas lamp, does not have a spectral distribution similar to that of a black body radiator, its color temperature alone is not a reliable means of selecting suitable filters or creating look-up tables for color balance corrections. Therefore, although two different light sources may be described as having the same color temperature, exposed photographic emulsions or digital images lacking proper white balance baseline adjustments may respond differently to the sources. When using fluorescent lamps or similar light sources, a per-wavelength comparison of sensitivity and spectral output is often necessary in order to determine the correct filters for color temperature balance.”

This description of color temperature captures many of the issues that challenge every designer of a video projector. In essence the light source is the life blood of the video projector.

Contrast Ratio - The ratio between the lowest and highest intensity that an LCD flat panel can display. A glaring (no pun intended) defect with this measure is that it can be increased by increasing the intensity of the back light to levels that exceed comfortable viewing brightness. Generally speaking contrast ratios exceeding 250:1 are perceived as acceptable. Better quality LCD panels can exceed 400:1.

ISO 13406-2 - Ergonomic requirements for work with visual displays based on flat panels - Part 2: Ergonomic requirements for flat panel displays. Includes standards for contrast ratio measurement, viewing angle, dead pixel count, and luminance among others.

Response Time - The amount of time it takes for a liquid crystal cell to go from active (black) to inactive (white) and back to active (black) again. This value is usually defined in milliseconds (ms). Values range from highs around 40ms to as low as 16ms.

Refresh Rate -

A general but fairly lightweight Dictionary of Multi-Media Projector Technical Terms

(To be continued)

Sites for test images:

http://www.lynxxx.nl/barco/patterns/

http://www.construnet.hu/nokia/Monitors/TE...nitor_test.html

http://members.aol.com/marcorsyscom/1920x1080.jpg

Google search on the terms "test image" (Thanks to h_ank for this intelligent suggestion.)

Sites describing test procedures:

http://brighamrad.harvard.edu/research/top...p/tutorial.html

Documenting test results:

(To be continued)

Technical references:

The following research paper investigates how Metal Halide lamps ignite and why a delay is required to re-ignite them.

Breakdown processes in metal halide lamps

Great thread.
Here is another one for ya.

Oh, and don't forget to buy one of these:

It can be bought at RadioShack, and goes by the name of "Solar Project Fun Kit" ($12.99). It is a solar cell. Take the lens thing off of the front, cut the wires off of the motor and hook a multimeter to those wires. By measuring the current, you'll have a lumen meter. This is very good to calculate the on/off Contrast Ratio.

Thanks Jones two great additions! I'll add the link to the test image group above.

Tom's Hardware has a couple of interesting articles about LCD Flat Panel Technology. The first article goes into some detail about the different types of LCD technology. The second article discusses LCD contrast ratios.

http://www.tomshardware.com/display/20020114/index.html

http://www.tomshardware.com/display/200402...t_ratio-01.html

here you can find many DD/Dts trailers, including HDTV WM9 clips for download.

Hi Jones,
Thanks for the trailer link. Should come in handy.

Oh, and if you're into those high definition 720p clips, don't forget checking here.

Stop it Jones. I can't get any work done.

Outstanding thoughts and contributions Jones and Voxelman. I would take it a step further-making everyones results' posts identical. For example;
1. list all materials used in projector
2. camera type and lighting conditions
3. 4 shots of box, 1 head on, 1 internal, 1 with lights on, 1 in the dark
4. a set number of specific images to compare and contrast.
I am sure I left things out that many consider to be important. Let us know what those things are. I really would like to see this happen. Thanks to Voxelman for bringing up the 'standards' issue. See you in the movies.

I like this idea, I will implement it on my website once I get it ... (hosting admins :angry:)

Anyway, the different pictures should allow us to collimate (pun?) some better ideas and suggestions to help.

Anywho good ideas all!

Also,


Yeah really jones! J/k those sites are awesome, even though I can't seem to download anything.

did you rt click and chose save as on file instead of clicking it?

yep

Btw, VERY quick response.

if you wanted to check the amount of light it puts out you simply could use a light meter that photographers use to measure light on an object, they arent that much money. This would easily allow you to see what adjustments make what amount of difference on projector brightness. It would be a standard we could measure light output with.

How much is not that much. Because my budget is getting a little low.

Here is a site that I found that helps determine your correct screen size/viewing distance.

Talking about "Standards of measurement", here is a nice pic someone at Avsforum did in order to show the difference in contrast ratio:

neat.

Wow, great post voxelman. I am going to enjoy following this thread.

Cheers,
Lee

Great site Voxelman... Great thread too!

Mark

Heh heh. I can't believe I avoided this topic for so long. I always assumed it had to do with converting standard measures to metric, etc.

I have made mention somewhere, recently, that it would be neat to have a .zip downloadable file, with a number of pics in it for people to project, when posting results. The result would be identical images being displayed. No other factors would come into play, such as the quality of the DVD player, component vs. SVHS inputs, etc.

Hey, Voxelman, this isn't you is it?



http://www.avsforum.com/avs-vb/showthread....light=rustoleum

No, I'm much more handsome.

OBTW, my projector is currently on the back burner as we just closed on the sale of our home and are moving this week. We will be living in temporary quarters without wood working tools etc. while our new home is under construction. I'm hoping to return to the project in the fall. The new house is going to have a purpose built home theater.

LOL, thats the Screen Doctor....

Well why wouldn't the screen doctor be intersted in DIY projectors?

I have a few shots I think would be useful to determine 1.) detail(sharpness, focus, etc.) 2.) contrast and 3.) wash out. Jr, please feel free to correct my terms, to something more accurate.

(Crap, 100k file limit. 300k was a bit much though.)

I will see what I can do about shrinking them a little.

Here is a colorful one, see if the white behind the characters is overly bright. Also shows some flesh (tones). Follow this link to download a full size pic.

This one shows great detail, in the books and the mantel, and the cage, and the other decorations. The contrast here is rather extreme, I think, so it is good to test that as well.

Okay, put your sunglasses on if you're using a reflector.

This one is bright. I imagine the 'sky' will be hard to look at on this one. Pretty poor lighting in my opinion, but you will encounter that in many movies. Don't judge the skin tones on this one, Harry looks a little green to me. Has a substantial amount of contrast though, from the sky to the shadows.

Other pics to download for standard comparison.

Marlin, Dorrey, and the Turtles

Very vidid shot of Dorrey and Marlin

This one is huge, I never noticed Bruce had so much detail to him.

Here's one... thought it was good for many different facets. Can't remember where I found it. It was on someone's home theater site.


cow1280x720.jpg

Mark

Great flesh tones on that one!

Yeah and the cow's eye-boogers really add a level of realism that you just don't see on pics from Finding Nemo...

She obviously didn't know she was going to have her picture taken that day, or she'd have worked a little harder on her curls.

I don't think it has been mentioned, we need a standard screen size, for us to make comparisons with. The difference in brightness is huge, from 120 down to 100 inches diagnally.

I suggest, when making posts of your results, to use a 100 inch 4.3 screen size, as a standard.

maybe we could have a section with fill in the blanks, for specs used in screen shots?

Has anyone taken a photographic light meter and measured the amount of light falling on their screen from a "standard" LL projector (projecting a white image) at a certain distance (like 120")?

I think it would be interesting to have a standard to compare projectors to... like an "atypical" projector with 400W bulb, Norpro, LL lens set, 120" from the screen, with the meter set to ISO 100, measures EV (x)...

This way, we can quantify which bulbs, reflectors, etc. work better than others... I personally haven't been able to get a Norpro, so for now I'm using a different chrome bowl... I'd like to know how it measures up.

Since we know that the digital pictures everyone posts don't tell us much... an image might look bright in a picture, but it's all relative... the shutter speed someone used for their picture could be 10 seconds for all we know...

The same criteria could be used to measure the amount of light reflected by a screen... to quantify which materials coatings, etc. are higher gain than others.

-N.

In order to measure the amount of light reflected by the screen, we have to be talking in foot lamberts. This measurement takes into account the gain of the screen and the size of the screen. The following formula is used to calculate Foot Lamberts:

Foot Lambert = Real Projector Lumens * Gain of Screen / Area of Screen in feet

At work here we use a Minolta CS-100 to measure color balance and FL to calibrate our projectors. (I believe you can rent them from camera shops) When we are calibrating normal scenes with no bright light points, we are only looking for a measurement of 5 FL. When we calibrate for the bright light points, we look for 20 FL. Keep in mind that these are light points, if the whole screen was filled with 20 FL, you couldn't look at the screen. These are big screens, so the initial lumens of the projector must be very high.

I think this would be a more accurate way of evaluating the output of these projectors because it is measuring what is coming back to your eyes. I plan to measure mine and set my white balance using this meter to obtain the best color I can get.

Has anybody seen published specs on the gain of different screen materials? (blackout cloth, painted wall, screen goo)