AVS Forum banner
Status
Not open for further replies.
1 - 20 of 125 Posts

·
Premium Member
Joined
·
24,983 Posts
Discussion Starter · #1 ·
This is inspired by periodic "dissing" of on/off CR as a worthwhile performance measure that I encounter (most recently here http://www.avsforum.com/avs-vb/show...335#post2142335 )


Partly it seems to be the idea that because on/off CR tends to be much higher than ANSI (checkerboard) CR, it must be inflated or bogus, like peak W vs. rms W. But this isn't "cheating", it's because of the difference in how they're measured.


ANSI CR is currently good enough not to matter to perceivable subjective performance, and on/off CR is the "tall pole" in CR performance. There is a clear correlation between the steady improvements of on/off CR in digital pj's and perceived benefits in less foggy haze, lower black levels, and greater 3D effect.


I emphasize that I am saying this only in regard to black levels (which by definition should be zero), as apart from blacks, which include shadow detail, which depends on gamma performance.


CR is exactly analogous to S/N ratio in audio; any black level greater than zero is photon noise.


If an amp has a high S/N ratio, no one is going to claim that it means anything about the sound quality. Likewise if a pj has 5000:1 on/off CR, it shouldn't be assumed that it has good shadow detail, only that it won't be hindered by high black level.


It would be helpful to be clear on the difference between "black level" and "blacks".


Black level is purely objective and is simply the leakage light projected with no signal. It is calculated directly as Lumen output divided by on/off CR.


IMO "blacks" is purely subjective and is affected by gamma/shadow detail, brightness, and color reproduction, as well as CR. I think there's a lot of confusion when someone says a pj has good blacks and/or black levels based, for example, on scene where someone's wearing a really black looking shirt in an otherwise bright scene. Even LCD's can have inky-looking blacks in these circumstances.


It seems that the people who have a problem with on/off CR confuse it with "blacks" as I desribed it above, not to mention other unrelated performance characteristics.
 

·
Registered
Joined
·
1,481 Posts
I'm quoting myself from the other thread, located at http://www.avsforum.com/avs-vb/showt...58#post2143458

but I think Noah's word is worth of attention.

These are some ideas to discuss..


Gentlemen,

it's always a pleasure having an open discussion with people like Paul and Noah.


Even from different perspectives, I think sharing ideas is great!

Only $ my .02 contribution:


i) I wonder how we can see so high CR on magazines. It's just a question, not an answer! We are using the Gretag, that's one of most accurate, but still not perfect, instrument to collect data.


I've been using Sencore and CA-1, and I frankly don't think they outperform the CF-6000 bundle. But I still lack accuracy at low levels: I don't think that the PhotoResearch (that I personally don't own, I'm only quoting from Mark Hunter's experience and from what I saw in Las Vegas with him directly), besides being heavier and slower, can be so rock steadily pin focused to a perfect 0 IRE reading...


More, in my experience, a dark 0 IRE can vary a lot: it can be 0.044, 0.34 or 0.018 (I'm referring by heart to actual, measured values), or whatever you'd like in the low range.


A simple math shows us how much a CR can differ from these values: if we get a, let's say, 35 cd/mq 100 IRE value (as for a mean DLP, for example), this leads to 795, 102 or 1,944:1 CR.


Are these numbers reflecting a real situation? Judging from the experience I had with the InFocus X1 tested on HTProjectors, I might say yes, and that's why I acknowledged Noah because this was an actual and tested situation.


I promised to send you all the details, Noah, and I'll do: what I'd like to have is your personal contribution: what kind of instrument are you using? I'm sure that every number must be correlated to a real experience: that's why I always put actual graphs coming from the CF-6000 and I always describe the environment.

I think we should try to compare our visual experience to numbers, in order to seek a partial truth... and to see what those numbers do mean when referred to a projector!


ii) Has anyone considered that reducing the screen size can definitely improve CR? Once again, it's a simple math: we use a different approach: 200 cm. vs 244 cm. base screen, for projectors up to 7" CRT or 1024 x 768 rez. vs. 8 and 9" CRT and 1280 x 720 and higher digital projectors.


Since a smaller screen helps improving the light luminance value (i.e., the smaller the screen, the higher the luminance you'll read, no matter what instrument you're using...), I think it could be of some interest if the manufacturers will give us the actual screen size they use to raise up those astonishing values we read (2-3K:1 range)... I do believe they are true, but let's see how they are obtained!

Once again, that's why we describe the very actual measure environment...


iii) I don't think CR gives us a proper and ultimate description of 3D and details.


I think this is correlated to resolution AND the ability to reproduce appropriate levels of a gray scale.


This is, IMHO, why a 8 or 9" CRT is still superior: it acts as an analogue driving source that can render every single change on any surface.


Let me take as an example Tom Cruise and his partner's face in the famous tr. 4 of Mission Impossible 2.

While a CRT can resolve every single detail, both as a continuous color palette and a continuous image, thus describing what we perceive as a 3D image (located in space and depth), a digital projector is still limited by its resolution (matrix structure, even if a 1280 x 720 is rather close to human resolution and a 1920 x 1080 will closely fit the top, see Andrea Aghemo's article on HTProjectors on this topic) and ability to "write" an analogue surface like a human face. This has to do with CR, but certainly it's not the whole bunch: that's why I think CR is limiting our global approach to a classification of a projector, Noah.


But I think we have to discover so many other things!
 

·
Registered
Joined
·
3,534 Posts
I like Noah's thinking. I'd really like to see a "gamma tracking" score in projector reviews. The score would relate to how well the projector can reproduce the standard NTSC 2.2 (or 2.3?) gamma, from 0 IRE to 100 IRE. If 10 IRE measures the same as 15 IRE, for example, you can be fairly certain that shadow detail would suck on that projector given the current settings.


Another useful measure is the projector's color gamut...how well can it reproduce the primaries when compared to an NTSC monitor, for example.


A rating system similar to the CRI rating used by lightbulb manufacturers would be nice!


Of course, all of this would require instrumentation, probably something the average reviewer could not do.
 

·
Premium Member
Joined
·
24,983 Posts
Discussion Starter · #5 ·
"That link does not work."


Sorry; the one Andrea posted works.


Andrea,


"Has anyone considered that reducing the screen size can definitely improve CR?"


The only way I can see for this to happen is if light control is imperfect and the black level is dominated by reflected ambient light. Otherwise the smaller screen increases black level as much as white, with no change in CR.


"I don't think CR gives us a proper and ultimate description of 3D and details."


I agree, I just don't think that's reason to dismiss the usefulness of on/off CR numbers.


I confess I have never measured CR, but I believe I can still comment based on reasoning and what I have seen.


Regarding the difficulty of measuring very low light levels, one workaround is to measure close enough to the pj to get a good reading on black, and then use an ND filter with known transmission loss to measure white. For example, if you use a filter that transmits 10%, you've effectively increased the dynamic range of your instrument by 10X.
 

·
Premium Member
Joined
·
24,983 Posts
Discussion Starter · #6 ·
I changed this paragraph


"IMO "blacks" is purely subjective and is affected by gamma/shadow detail, brightness, and color reproduction. Even LCD's can have inky-looking blacks when they're part of a bright scene."


to this


"IMO "blacks" is purely subjective and is affected by gamma/shadow detail, brightness, and color reproduction, as well as CR. I think there's a lot of confusion when someone says a pj has good blacks and/or black levels based, for example, on scene where someone's wearing a really black looking shirt in an otherwise bright scene. Even LCD's can have inky-looking blacks in these circumstances."
 

·
Registered
Joined
·
1,481 Posts
Quote:
Originally posted by maxleung
I like Noah's thinking. I'd really like to see a "gamma tracking" score in projector reviews. The score would relate to how well the projector can reproduce the standard NTSC 2.2 (or 2.3?) gamma, from 0 IRE to 100 IRE. If 10 IRE measures the same as 15 IRE, for example, you can be fairly certain that shadow detail would suck on that projector given the current settings.


Another useful measure is the projector's color gamut...how well can it reproduce the primaries when compared to an NTSC monitor, for example.


A rating system similar to the CRI rating used by lightbulb manufacturers would be nice!


Of course, all of this would require instrumentation, probably something the average reviewer could not do.
maxleung, I'm sorry that you evidently never took a look at www.htprojectors.com :p !!!


I hope you'll like the site and encourage you to look at it criticizing in any manner!


We have online everything you're asking for...
 

·
Registered
Joined
·
1,481 Posts
Quote:
Originally posted by noah katz



The only way I can see for this to happen is if light control is imperfect and the black level is dominated by reflected ambient light. Otherwise the smaller screen increases black level as much as white, with no changer in CR.
Noah, unfortunately this is not true.


Please remember that a projector is NOT a linear device, apart from gamma's issues. Even with the same settings and same ambient light conditions, if you repeat the measures with a smaller screen, you'll get different values, that are inherently not linear.


You can have an immediate verification of this statement thinking why that are lot of people (let me talk about my magister Joel Silver who spends his time asking for smaller screen: does it makes sense using a 9" CRT with a 200 cm. screen? I'm not so hard positioned (I use a 244 base for 9" and 1280 x 720 DLPs), but I think this is correct. No doubt. Concentrating light helps achieving better performances.


That's why I've been insisting on a real hands-on experience: every single time I have to work with a measuring instrument, I always find new questions that I'm not able to answer with. And this means thinking over these open issues, and trying to find correct answers. Sometimes these might be silly, but you can even catch the heart of the problem...


Quote:
Originally posted by noah katz


Regarding the difficulty of measuring very low light levels, one workaround is to measure close enough to the pj to get a good reading on black, and then use an ND filter with known transmission loss to measure white. For example, if you use a filter that transmits 10%, you've effectively increased the dynamic range of your instrument by 10X.
Once again: I think you should buy a real instrument to spend a word about this.


Using a spectroradiometer you won't be affected by reading distance, but only by screen dimension. I've been discussing many times with Mark Hunter about this, and this can be a partial solution. Partial because, to get a good reading of black, you need to reduce screen size...

Not to mention that getting closer to the screen can give you some difficulties with the shadow the instrument itself will project on the screen.


I don't want to take into account using a colorimeter, that lacks accuracy at low level (this might have been a solution because in that case you're measuring incident light, i.e. illuminance, and you're aiming the sensor toward the projector). But you don' have accuracy, with such an instrument!


Sorry, but I must disagree on the ND filter usage too.


We are talking about measures, here, not calibration (that, in the end, is something related to art or personal tastes).

I've been playing for such a while with Lee filters, that are a cheap solution for filtering a CRT.

Have you ever seen what the transmission characteristics are? What you see as similar filters may differ A LOT in terms of transmitted spectrum, thus altering the "quality" of light that the projector is dealing with.


And this is EXACTLY what a measure should avoid, i.e. putting different "transfer functions" inside the measuring chain.


So I should use a "perfect" filter, with a flat response on the whole range: do you think that a photographic ND filter will actually meet those specs? ;)


I know lot of people follow this route, but I think this can be a good personal solution for a personal (again!) calibration, not an aid for measuring: it's only another, unpredictable variable...
 

·
Premium Member
Joined
·
24,983 Posts
Discussion Starter · #9 ·
Andrea,


"Please remember that a projector is NOT a linear device, apart from gamma's issues. Even with the same settings and same ambient light conditions, if you repeat the measures with a smaller screen, you'll get different values, that are inherently not linear.


You can have an immediate verification of this statement thinking why that are lot of people (let me talk about my magister Joel Silver who spends his time asking for smaller screen: does it makes sense using a 9" CRT with a 200 cm. screen?..."


I see the problem with a highly nonlinear CRT, but I was speaking of taking any pj, linear or not, and simply illuminating a bigger screen without any adjustments to it.


I don't see the relevance of your example to the validity of on/off CR.


Regarding use of ND filters, do you think their deviation from color neutrality is enough to make more than 5% differnce in the results?


Thanks
 

·
Registered
Joined
·
4,735 Posts
Quote:
Originally posted by Andrea Manuti
Please remember that a projector is NOT a linear device, apart from gamma's issues. Even with the same settings and same ambient light conditions, if you repeat the measures with a smaller screen, you'll get different values, that are inherently not linear.
I think it really depends on whether we are talking about measured CRs or real CRs. They are 2 different things. We strive to have the measured CR be as close to the real CR as possible but our test equipment and other things don't make them exact. The real CR is the intensity of the "white" divided by the intensity of the "black". Now can you explain how the real CR would change when you change the screen size for a DLP? How did the screen size affect the ratio of the intensity of the photons in the "white" divided by the intensity of the photons in the "black"? Was it reflections off the walls. Or are you claiming that the amount of total light that the projector puts out at "white" to the amount it puts out at "black" changes by screen size?
Quote:


So I should use a "perfect" filter, with a flat response on the whole range: do you think that a photographic ND filter will actually meet those specs? ;)
While it is true the filters aren't perfect, I think you can easily measure the amount that it filters. If it is 90%, then just about any of the sensors we have should be able to measure things that are 10:1 CR apart with great accuracy. And if filtering varies by color I'm guessing it doesn't make a huge difference. I think the key is making sure that we are well within the usable range of whatever sensor we use. This is one reason that I've suggested using a similar method to the filter one. Mine is:


1. Put the projector and sensor close to the screen to get a very bright image and measure 0 IRE and 20 IRE

2. Put the projector a ways from the screen to get a much dimmer image where 100 IRE won't overload the sensor. Now measure 20 IRE and 100 IRE.


Now you can determine the ratio of 100 IRE to 0 IRE by taking:


((20 IRE) / (0 IRE)) * ((100 IRE) / (20 IRE))


--Darin
 

·
Registered
Joined
·
3,534 Posts
Andrea, your website is very comprehensive...it'll take me a while to digest the information in it. Hmmm, look at all the flaws you found in the Infocus X1. :)


I really appreciate the gamma readings etc.
 

·
Registered
Joined
·
5,445 Posts
Andrea

I have checked on on htprojectors.com and with the measurements for x1 and Cine 7 there is not much differnce in full on/off. Bothe were from my memory between 100-200:1.


How come the x1 is so far from specs such as 1000:1 and Mr Phelps has measured his crt at 8000:1. Is the difference that you measure off screen and that the specs are measured in front of the lens? The difference is still very wide, eplainable or ?
 

·
Registered
Joined
·
5,445 Posts
I have checked now and the 0 ire values are close to the value for the Sony g90 crt that was 0.15 nits, the same as 0.15cd/msq. However the 0.15 could be lowered to get higher cr even though not optimal for pq. THe 100 ire values seem low for the Cine 7 compared to the g90. The g90 had more than 1000 nits and the Cine 7 is measured below 40 nits. What am I missing?
 

·
Registered
Joined
·
1,481 Posts
Some notes to everyone.


This is an example of actual measurements with an actual projector in an actual environment…

http://space.virgilio.it/andreamanut...nvironment.JPG


I like to make an accurate overview of what I do: that’s why we always publish instruments and conditions where the projectors were measured, including the surrounding hardware.


So, Olson, I really don’t know how William could have raised an 8000:1 full on / full off: all I can say is that I’ve changed the CA-1 even because of the lack of accuracy at low ends (at least, the version I had until September 2002), and I think that the Gretag is superior on this aspect.

I can only talk for the numbers I know and for what I’ve got, but I like to show what I have here…


The first answer that comes to my mind when comparing the X1 and Cine 7 full on / full off results, is: THAT’S WHY I WAS TELLING THAT CR IS NOT AN ISSUE FOR TAKING INTO ACCOUNT AN ACTUAL PJ PERFORMANCE!!!


I don’t think, sincerely, that a Cine 7 can be compared to an X1, nor I think that even the dullest DLP fan would dare so much: so, the conclusion is up to you: is this number really meaningful?


And about G90 performance: it makes a little sense to me comparing a 7†with a 9†very bright pj as the G90 is. But, besides that, I never measured a G90 and don’t know how we could ever get 1000 cd/mq, Olson… The D-ILA G150HTE I’ve tested, had 74.296 cd/mq with a 244 based screen: and I don’t think that a G90 can be more than 10 times a DLA in brightness… But I can’t answer for someone’s else method!


Let’s go to what I’ve just took away from a projector I’m currently testing, a Yamaha LPX-500, an LCD 1024 x 768 unit.


I’m reporting these conditions, under which I made the test, that comes from a calibration I made some days ago:


Black level = -3

White level = -7

Picture mode = A

Color temp = 6500°K

Flesh tone = 4

Color balance

Offset R = -3

Offset G = 2

Offset B = -8

Gain R = -2

Gain G = 0

Gain B = 7

Gamma R = 1

Gamma G = 0

Gamma B = 3

Sharpness = 0


The connection was via a 10 meter DVI cable, and the output was from a Radeon 9700 card with Catalyst 3.0 drivers.


The actual location is my home (you’ll find some photos here and within the Special guest Forum located at http://www.avsforum.com/avs-vb/showt...hreadid=233170

I never claimed this to be a perfect HT theater, it’s only a real one, like many HT addicted have: we can see white walls and curtains, but, take it for sure, every single light is totally turned off when making measures!


I’ve made 2 sets of measures, the first with a 117 x 88 cm. image (hereafter, small screen, the second with a 158 x 199 set (large screen).

You might ask why these values: it’s only determined by the zoom factor of the Yamaha LPX-500: I decided to place it on a table and maintain the position fixed, to make the difference visible in the shots I made.


My purpose was to explain why CR can vary with the screen dimension.


If we have the well known relation (Jerry Whitaker, Video Display Engineering, Mc Graw Hill, page 307):


Luminous flux in Lumens = [ L (Luminance in cd/mq) * A (Area in square meters) ] / Screen gain


we can determine that a pj’s output in Lumens is the value we get from the CF-6000 (in my case, I’m making a luminance reading) * the actual area in square meters / screen gain.


So I assume that these values are 1.0234 mq for the first case and 1.8644 for the second (screen area) and 1.3 screen gain (Stewart StudioTek 1.3, 244 cm. base).


What we can see is that the values are quite similar, and this is another reason that shows why we at HTProjectors decided to go the Colorfacts way. I think what Mark is doing is absolutely remarkable!


I made two more measuring sessions, the first with the current Colorfacts version, and the second with a beta Mark Hunter kindly provided me with. I have to say I do prefer CF 3.0, but this is another story…


Here we are with numbers:


Colorfacts 2.25.0244


Yamaha LPX-500, 117 x 88 cm. base, calibrated device

100 IRE 96.43 cd/mq

0 IRE 0.64 cd/mq


Yamaha LPX-500, 158 x 119 cm. base, calibrated device

100 IRE 54.98 cd/mq

0 IRE 0.48 cd/mq



Colorfacts 3.0.0256 Color reference SMPTE 240M (HDTV)


Yamaha LPX-500, 117 x 88 cm. base, calibrated device

100 IRE 96.28 cd/mq

http://space.virgilio.it/andreamanut...0100%20IRE.JPG



0 IRE 0.58 cd/mq

http://space.virgilio.it/andreamanut...%200%20IRE.JPG



Yamaha LPX-500, 158 x 119 cm. base, calibrated device

100 IRE 55.41 cd/mq

http://space.virgilio.it/andreamanut...0100%20IRE.JPG



0 IRE 0.45 cd/mq


http://space.virgilio.it/andreamanut...%200%20IRE.JPG


So, what we are using for our considerations are the CF 3.0 results.


We get:


Small screen


Base 117.50

Height 88.00

Area 1.03


100 IRE 96.28

0 IRE 0.58


Full on / full off 160.00


Output in Lumens 76.58




Large screen


Base 158.00

Height 118.00

Area 1.86


100 IRE 55.41

0 IRE 0.45


Full on / full off 123.13


Output in Lumens 79.47



Comments?


As we see, the CR value for the small screen is 160.00:1, while the large is 123.13:1.


I think this makes some kind of difference, even though I have to advice that your actual conditions may vary a lot: should I turn brightness and contrast controls all way up, we’ll get higher numbers.


Are these values too low? I don’t know about the others, I’m sure that the CF-6000 is affordable enough at high levels, and maybe could improve a little bit at low ones.


But I must underline that any other conclusions with other instruments that are not lab dedicated (and this is our choice at HTProjectors) should vary a little from these numbers.


That’s another reason why I’m so stubborn when talking about astronomical values for CR: I would like to determine if the measure was made with an actual calibrated projector, not with default conditions or with contrast at its maximum: this makes no sense in an HT environment, at least IMHO and for my experience.


If there are mistakes in these formulas, I’m ready to correct them: I think we all can make mistakes and I’m surely the first.


Should you have different values, that can prove a 2000:1 CR, show them: but please indicate exact conditions as I did.


I would only like to find numbers that are shared among enthusiasts: but I want to talk with facts, not with suppositions, ideas or someone’s else experience. So let’s talk about OUR values, not what we heard from another person: the only contribution I can give is show you what I did get, showing you the environment and photos.


Then you’ll draw your conclusions!


In the end, just an example about one of the improvements we have on CF 3.0: this is the color gamut for the Yamaha LPX-500 when compared to high def coordinates: it shows some poor green, but great red and blue: not bad for an LCD unit!

http://space.virgilio.it/andreamanut...art%20HDTV.jpg
 

·
Registered
Joined
·
5,445 Posts
Thanks Andrea for the elaborate reply.

Where do photons go and where do they die?


The amount of energy in the shape of visbile photons is what constitutes a light output in lumens.


The photonic energy that reaches the screen either is reflected or is absorbed and dies. The photons that are reflected take a direction that is determined by the incident angle towards the screen and the gain chartacteristic of the screen. A perfect lambertian, 1.0 gain. should distribute the photonic energy over 2 pi steradiand of solid angle, that is in all directions but through the screen.


We must ask ourselves.

1 How much of the photonic energy is absorbed by the screen.

2.How much effect does the incident angle centre compared to the screen edge have on light reflection characteristics.

3 How much light is actually reflected right back ot reach our eyes.


Andrea

Perhaps you could pm William Phelps to compare methods. Your readings seem consistent on the 0 ire end but not at the high end. It would be as interesting to know how he reached his numbers as to read about how you reached your numbers. I guess that Mr Phelps reading are with the light meter facing the projector.


Also how much of the screen in percent is used for the full on / off in your system. The larger that area the lower the lux would be from the crt but not from the lcd or dlp right?
 

·
Registered
Joined
·
4,735 Posts
Andrea,


I haven't read your whole response, but if you are measuring 0 IRE with anything else on the screen then you are doing it wrong and getting much lower on/off numbers than are realistic. Those other bright things will throw your numbers off, even though the Gretag device is pretty directional. Especially with white walls where the bright items on the screen are bouncing back into the sensor. What you described above was a whole lot closer to an ANSI contrast test than an on/off contrast test.


Could you try the CR measurement tool builtin to ColorFacts for on/off and see what you get? Your numbers should end up being much more realistic.


If your numbers vary that much between your screen sizes it is your tools and setup that are at fault, IMO.


Also, I didn't think ColorFacts could be used to get lumens reliably. Maybe I'm wrong there. I've only used their numbers for ratios not for absolute brightness/area.


Also, since William Phelps is "the guy" for certain stuff, his numbers are rarely disputed. I'm not sure exactly how he did his measurements although he gave a reasonable writeup, but if anybody wanted to know how to do them correctly, he would be one of the people to ask.


Thanks,

Darin
 

·
Registered
Joined
·
1,481 Posts
Quote:
Originally posted by Ohlson



Andrea

Perhaps you could pm William Phelps to compare methods. Your readings seem consistent on the 0 ire end but not at the high end. It would be as interesting to know how he reached his numbers as to read about how you reached your numbers. I guess that Mr Phelps reading are with the light meter facing the projector.


Also how much of the screen in percent is used for the full on / off in your system. The larger that area the lower the lux would be from the crt but not from the lcd or dlp right?
Mathias,

the Gretag has a reading angle that is something about 8°, if I remember correctly.


Before using it, you have to aim the meter, i.e. moving it find the "right spot" where the sensor is directed: the software sports a little window, right in the middle of the screen, that is something close to a 5% window. So we're sure that the sensor points directly to the maximum.


When I make my measurements, I always open a wider window, as you can see from the shots, so I'm "rather confident" ;) that I'm doing an accurate reading.


I forgot mentioning that I have a photoradiometer from DeltaOhm too, the HD9221 with a luminance sensor, that has a 6° reading: I've been performing a double check, and the Gretag and the DeltaOhm sported EXACTLY same values.


This is something uncommon, since I've been used to a certain degree of approximation when dealing with these instruments.


Another thing I forgot mentioning is that I prefer using the windowed patterns that Colorfacts presents and not the Contrast ratio wizard: I've found this one less accurate, even if gives you higher CR values: but they're somehow "uncertain", so I prefer the other way that appears more consistent to me.


This is the reason, Darin, why I'm not using the CR measurement tool: I've seen values ranging from 400: to 1,200:1 at same time!


In any case, I'll do further investigations to understand if what I got is definitely acceptable.


But let me dissent from the fact that the surrounding walls can have such a great importance: in my case, they're covered enough with paintings and other stuff, and the real white area is not so important.


Anyhow, if you're familiar with the Gretag you'll be confirming (I think you are from your words, " I only use their numbers for ratios not for absolute brightness/area") you know very well that a little deviation from the optimal positioning of the sensor alters the "Aim sensor" procedure: you see values ranging from 8 to 20 cd/mq with some 5-10° vertical movements.


The CF-6000 allows me to do a reading that INCLUDES the screen, while the Sencore and the CA-1, along with the CF-100 (that is Milori's colorimeter) work on an illuminance base.


So the difference is that I point the sensor toward the SCREEN, they work toward the PROJECTOR.


I make a direct cd/mq (or nits, if you prefer) reading; that is native, i.e. the scale of the software works accordingly to the physical unit we are using.


Every colorimeter's software, as far as I know (Sencore, CA-1, CF-100, the ones I've been using), that was originally used for direct TV readings, expresses its output in cd/mq.


But this is true only if you're working with a direct set TV, where the reading is a luminance one.


If you use the colorimeter to analyze a front projector, you have a reading in cd/mq: this is NOT CORRECT, since the physical unit you're dealing with is LUX (=> sensor facing the pj, illuminance!). So you have to do some math, multiplaying by 3.14 and approximating a spherical surface with a plane. Later I found a precise mathematical explanation of this (see VESA Flat Panel Display Measurement Standards, Version 2.0, June 1, 2001, Appendix A215, p. 267, "Illuminance inside an integrating sphere", where you read that the relation between Illuminance, E, and Luminance, L, is exactly E = 3.14 * L). But you have to submit to same degree of approximation...


This is something I never get acquainted with when I used to writing for Digital Video magazine, some time ago: we had only the Sencore and struggled a lot to find an acceptable math to do this. A transformation that is, BTW, not correct dimensionally: lux and cd/mq are not the same!


In any case, I hope that the 1,000 cd/mq might be an error: as a physical sense to what I'm saying, let me quote once again from a book I have (Jesus, I've been spending a lot of money with this stuff, now it's time to show it off!), Wyszecki & Stiles, Color science, Wiley Classic Library, 2000: we can read at p. 25 that luminance of a green ceramic electromagnetic source operating at 400 Hz is something that barely reaches 70 cd/mq, while at p. 28 we see that a gas-filled lamp has 1,200 cd/mq, while a 750W projection lamp is rated for 2,400!

More, I can add that all the way through the VESA book I mentioned, while talking about actual luminance measures, the average value reported as example never exceeds 100 cd/mq for flat panel display: I think this should be a fair comparison to a front projector!


So, I hope that the 1,000 cd/mq for the G-90 is a misunderstanding, otherwise I guess William should calibrate his sensor a little bit... ;)


In any case, I do not want to share the truth: once again, I'm here and I can repeat all the work I've done and fix my errors: should anyone had different experience, I'm glad to confront them.

But, please! Only real and actual numbers, and exact indication of instruments and conditions!


Darin,

I'm not talking about lumens, I'm using the CF-6000 to take a 100 IRE value expressed in cd/mq that is compared with the DeltaOhm and is perfectly matching and consistent with the physical definition.

I sincerely don't catch your point:


" If your numbers vary that much between your screen sizes it is your tools and setup that are at fault, IMO. "


What I've got is exactly what I've expected from the definition itself of luminance. If you go through what I wrote, you'll find that there's a match (obviously is not exact, but it's close) between the derived lumens value (76.58 vs. 79.47, that is pretty the same) when we focus our attention to the PROJECTOR capability to output light energy (lumens) that should not vary with the screen (it's a characteristic of the DUT itself), while we get different values REFLECTED from the screen, that is what I get from the CF-6000.


Do you really think that putting the same "numbers" of photons (i.e. coming from the same projector under the same conditions), as Mathias says, on a 1.03 sq. meter can be THE SAME that on a 1.86 sq. meter area???


C'mon, Darin, I think you should reconsider the physical phenomenon and reason a little bit about what you're saying: you are practically stating that what I get from the same projector (light output, energy, call it as you want) illuminating 2 different areas that have a 1.85:1 ratio, will produce the SAME effect?


So I'm using 78 lumens to obtain some level of light on a 1.03 sq. meter screen; after that, the same 78 lumens fall on a 1.86 sq. meter screen, and I have to obtain the same result?


With the help of the Holy Spirit, Darin?


And the energy that is used for illuminating the remaining surface is for free?


I think we shouldn't confuse lumens with cd/mq... but it's only MY point!!!

:p
 

·
Registered
Joined
·
1,481 Posts
Quote:
Originally posted by darinp



Also, since William Phelps is "the guy" for certain stuff, his numbers are rarely disputed. I'm not sure exactly how he did his measurements although he gave a reasonable writeup, but if anybody wanted to know how to do them correctly, he would be one of the people to ask.


Thanks,

Darin
Darin,

I do agree, and have just emailed William to make him aware about this thread. I think it's the minimum I can do, since I was one the guys who mentioned his measures.

I hope he'll be participating, since I find this thread really fascinating, and think his contribution will be of great value!
 

·
Registered
Joined
·
1,481 Posts
Quote:
Originally posted by darinp
Andrea,


... What you described above was a whole lot closer to an ANSI contrast test than an on/off contrast test.


Could you try the CR measurement tool builtin to ColorFacts for on/off and see what you get? Your numbers should end up being much more realistic.


Thanks,

Darin


Darin,

I have a terrible ringing bell in my brain (just got up from bed while trying to fall asleep... :p), and I think I should reconsider what you told.


I'll do some further investigations with CF 3.0 to see if the value I'll get from the CR wizard will lead to other conclusions: even if I definitely do not believe in CR as #1 performance index, I think accuracy is a primary goal!


So, just give me some more time and I'll post something else...


Thanks, Darin!
 
1 - 20 of 125 Posts
Status
Not open for further replies.
Top