I would like some guidance on how best to determine the whiteness, brightness, and uniformity of various surfaces for use as a projection screen. See Background at the end of this post for more details why I am doing this.
I will be placing 600 x 600 mm test samples on a large board, projecting light onto them from a digital projector, and photographing the samples. Then I'll open the images in Photoshop, and using histograms or other techniques I'll compare the whiteness, brightness, and uniformity of the samples. The results will end up in a PDF with side-by-side comparisons of the histograms and anything else from Photoshop that would show useful information about how the fabrics reflect light.
These technical objective tests would be followed by subjective viewing tests to see if the differences highlighted by the technical tests can actually be seen in practice.
What I would like to know from people more knowledgeable than I am, is how best to do the technical tests. I don't really need absolute figures for most of what I'm doing, only relative figures. So I doubt whether I need a spectrometer or calibration equipment. I'm hoping that I can use a digital camera to achieve what I want. But maybe I'm wrong about that, in which case I'm all ears.
Imagine a darkened room, a projector throwing light onto a 600 x 600 sample, and a digital camera on a tripod, probably behind the projector. After photos are taken, the images will be opened in Photoshop to determine certain characteristics. Here are my questions:
Q1: What should be the colour-balance setting of the camera?Just set it to daylight and stick with that setting for all photos, or automatically set the colour balance from one of the samples, and stick with that?
Q2: White purity. In Photoshop, the histogram for each image should indicate the colour purity. i.e. how closely the R, G and B histograms line up under each other. The histograms will be included in the PDF, but is there a way to distill one number from a histogram that represents how white is the light reflected from the sample?
Q3: I assume that "whiteness" is not too important a characteristic because the project-screen combination should be calibrated anyway, thereby removing any "non-whiteness" of the fabric. Is that a correct assumption, assuming the fabric is near-white and not coloured.
Q4: My initial testing indicates that some of the whitest fabrics are probably fluorescing (http://en.wikipedia.org/wiki/Optical_brightener) i.e. made to appear whiter than they otherwise would be. Anyone know if professional screen fabrics use optical brightners? Are they likely to cause image problems?
Q5: Uniformity. The Photoshop histogram should also reveal the uniformity of the light reflected from the sample. If I want a measure of how the brightness varies across each image (due to hotspots or defects in the material) what is the best way to portray that variation? The standard deviation?
Any suggestions as how best to evaluate different screen materials in terms of brightness, whiteness, purity, and uniformity, using a digital camera or other equipment, would be most appreciated.
Given that I'm wanting a projection screen, I suppose I could just buy a $150 budget model and endure all its problems…
… or pay 20 times as much for a Stewart Studiotek 130 screen and never have a problem…
… but I think there will a happy medium somewhere in between where I can obtain nearly the same image quality as the Studiotek 130, but at a tenth of the price. So I've obtained samples of screen materials and intend comparing them against white bed sheets, roller blinds fabrics, paper, and anything else that is white and flat, including painted surfaces. You might laugh at the latter, but surprisingly, a $100, well-painted, flat surface using the right paint, can give an image very close to that obtained from an upmarket screen material such as the Studiotek 130…
Basically, I want to find out if there really is something special about projector screen material that other materials cannot duplicate. If there is some special characteristic, I want to know what it is. But if another fabric has similar light-reflecting properties, I may decide to use it, combined with a good-sized roller mechanism from either Benthin (http://www.benthin.info/blinds/roller-1.aspx), or Rollease (http://www.rollease.com/skylinegalaxy.html).
This sounds great. I'm a complete screen noob, just getting my 1st projector and I'm totally lost. I really don't want to blow $3000 on a screen when something ~$500 will be almost as good.
Let us know your results!
Indeed let us know
It has taken a few months, but I've almost finished gathering screen samples (about 25 so far), my Lumix GH2 has arrived with which I'll be taking photos, and my new BenQ 7000 projector (DOA) has been replaced with a W7000+. I am about to start serious testing. Initial results on five samples, taken at 0º, constant exposure, white balance set to the Center Stage XD fabric, then using Photoshop to determine the Lab values, are:
The little I know about Lab values says that the "L" number is the luminance, and each step is the minimum brightness difference detectable by the human visual system. Thus, for example, the StudioTek 130 is five steps brighter than the Center Stage XD, but only 1 step brighter than the paint. The "a" and "b" coordinates contain the colour value and should be 0 for neutral gray. Which brings me to a point where I need some technical assistance.
Note: I'm not looking for absolute colours and brightnesses in my testing, just relative values. I'll be happy if I can say that this fabric is brighter than that one, but (relative to the Centre Stage XD, my colour reference) both have a certain amount of colour shift.
So, I want a single number that indicates how close to being neutral each image is. What I do is select a small representative portion of the image, average that selection, and use Photoshop to tell me the colour numbers. Let's just say the Lab numbers (L, a, b) for three fabrics are:
A: 80, -2, -6
B: 85, 7, -1
C: 83, -4, -4
What is the best way to indicate how far each colour is from gray?
Any suggestions most appreciated.
Wow. Too bad I missed this thread to begin with. Your time and work you put into this is to be commended but unfortunately it is a difficult way to test screens. I wouldn't worry about gray compared to a reference so much as this can be corrected withcalibration. If you are looking for a reference material to compare materials I'd probably get a sample of the Studiotech 100. If you want an actual reference then use one of these.
Some screen manufacturers don't even use these and use their basic white screen as a reference. That's why gain figures between different manufacturers don't always correspond with gain against a reference. Stewart uses a reference material.
If you are going to test this you need a light source that will be relatively the same distance your projector will be from the screen and measure at similar angles. 100 in wide screeen with 120 inch throw will equal 600mm wide with 720mm throw distance. This will help with screen uniformity testing if you test the corners vs the center. The reasons many screens hotspot are:
1. Poor setup of projector throw and angle for the screen. Generally shorter throws are more prone to hot spotting as the angle of incident = angle of reflection causes the outer part of the screen to not reflect as much light to the viewer as the center. Curved screens can help with this especially wider 2.35:1 screens.
2. Quality of lens in the projector, if the lens has poor uniformity the screen will too.
3. Coating on the screen not uniform. Stewart is excellent at this in my opinion.
There are a number of factors in picking a screen for your environment. It is also difficult to judge screens with samples as it is difficult to get a good idea of uniformity and artifacts on small samples. Also, it can be tough tell the differences even if you have a 50/50 screen. We are typically drawn to the brighter sample and with the exception of judging brightness it can make it difficult to judge other screen qualities unless you equalize the brightnessbetween samples. TV manufacturers know this very well and that is why you see the TVs in 'Torch' mode on many showroom floors.
Using a brand new projector and putting an image on the wall and saying I can go with a screen this size is not a good measure for determining screen size but you see it recommended frequently. Make sure you have a screen that will be bright enough for your application. You need to account for bulb aging when figuring out your screen size and brightness. Bulbs typically lose more than 50% of their light output over their lifespan.
By the screens you are testing it looks like your are deciding between an acoustically transparent screen or not and one for darker room conditions and one for lighter room conditions. I think you need to narrow the field first. I'm not sure if you plan on getting either the Studiotech 130 or Firehawk microperfed? If so there are different requirements for microperfed screens and placement of the speakers behind it. They typically work better if the speakers are about a foot behind the screen and some EQ will be needed. It is actually beneficial for the placement of speakers to be a foot back from the woven screns as well but if you move the speakers closer it doesn't affect the sound as much as a microperfed screen.
Hope this helps.
Ps. When you test with a camera for comparison, you need to test with all the same camera settings (i.e. ISO, aperture and shutter speed). Make sure you choose settings that will allwo the difference in the screens brightness. Auto mode on the camera could skew the results.
In case someone visits this thread and wants to know the results of my testing, check out: http://www.avsforum.com/t/1483279/comparison-of-fabrics-for-use-as-a-projector-screen
|All times are GMT -7. The time now is 11:10 AM.|
Powered by vBulletin® Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
User Alert System provided by Advanced User Tagging (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.