AVS Contrast Thread - Now with Dynamic Contrast Results! - Page 4 - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #91 of 200 Old 09-13-2008, 01:35 PM - Thread Starter
AVS Club Gold
 
Mark Petersen's Avatar
 
Join Date: Dec 2000
Posts: 4,671
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Cliff, many thanks for posting this. I think there is a typo on 100 IRE projector 1 or 2, blue tube though.

Is it possible to take one of the low APL test patterns (maybe the 0.5% APL pattern would be the best) from this thread and measure with that too? The reason I ask is that ignoring black out scenes what's really relevant as far as contrast is intraimage contrast. At very low APL's digitals can come close to approaching their native on/off limit and it would be very interesting to see the behavior of CRT's too. The full suite of test patterns aren't needed for this just the reference white and dark patterns for say 0.5%.
Mark Petersen is offline  
Sponsored Links
Advertisement
 
post #92 of 200 Old 09-13-2008, 01:50 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by darinp2 View Post

Interesting numbers Cliff. If you get a chance could you try a subjective test with Finding Nemo. Basically, measure approximately how long it takes before you can see the edge of the screen when the image blacks out in the Abyss scene.

Do you think your dealer friend will be bringing an RS20 over for comparison when they come out?

Approximately how much does a red retube for the G90 cost these days? Is there anything you could do in the meantime like lower your offset for the red tubes? Looks like you are way over a million to 1 just considering blue and green.

Also, if the Minolta LS100 that Ken uses is like the LS110 I have used then it may not be able to measure ANSI CRs over about 200:1 unless a dark light pipe is put on the end of it. I made a pipe out of some somewhat stiff paper type board that I covered with black velvet and then rolled up. I think I had to go to about 18" long for the pipe to get it to work well. If the room isn't taken into account then I'm guessing the Samsung could do 500:1 ANSI CR or more.

--Darin

Darin, do you still have that LS110? The reason I ask, is if I had one that I could keep here for a week or so, I could really dive in and take all of the measurements and scenarios for you guys!

It's hard for me as Ken is not around enough and when he is here we like to enjoy our time watching and not "work" in a sense.

As far as the retube, P19LUGS can be had right now for 1000 bucks each brand new and your right, if I did change the reds out, the numbers would definitely go way up. I'm damn near tempted to do it but I've got well over 3000 hours on these tubes now (the blues and greens) and am not sure if by next year there will be something out there on the digital side that will make me want to switch.

If not, I'll probably end up retubing again!

Cliff
overclkr is offline  
post #93 of 200 Old 09-13-2008, 01:51 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
The method of calculating CR is wrong. If you are trying to calculate the CR for white, the proportion of the three primaries is not equal, but determined by the formula for the standard. Green is close to 60% of the total. The second issue is the accuracy of the LS-100 for the black measurements. You need about .050 fL for black to be accurate. For example, that .002 fL reading is +/- .002 fL. The CR for white is probably very high, higher than stated, but the low accuracy makes an actual number meaningless.
odyssey is offline  
post #94 of 200 Old 09-13-2008, 01:53 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by Mark Petersen View Post

Cliff, many thanks for posting this. I think there is a typo on 100 IRE projector 1 or 2, blue tube though.

Is it possible to take one of the low APL test patterns (maybe the 0.5% APL pattern would be the best) from this thread and measure with that too? The reason I ask is that ignoring black out scenes what's really relevant as far as contrast is intraimage contrast. At very low APL's digitals can come close to approaching their native on/off limit and it would be very interesting to see the behavior of CRT's too. The full suite of test patterns aren't needed for this just the reference white and dark patterns for say 0.5%.

Mark hopefully I can source a minolta some how for a week or so and we'll really dive in deep.

As far as the typo, nope, that is my blue tube output. It just goes to show you how 2 of the same CRT projectors can vary. That 1209 FL number is correct and those measurements equate to exactly 5 foot lamberts per projector.

I hope you guys enjoyed me FINALLY getting around to getting this done. I'm actually like I said extremely happy and actually suprised at the result in a good way!

Cliff
overclkr is offline  
post #95 of 200 Old 09-13-2008, 02:06 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by odyssey View Post

The method of calculating CR is wrong. If you are trying to calculate the CR for white, the proportion of the three primaries is not equal, but determined by the formula for the standard. Green is close to 60% of the total. The second issue is the accuracy of the LS-100 for the black measurements. You need about .050 fL for black to be accurate. For example, that .002 fL reading is +/- .002 fL. The CR for white is probably very high, higher than stated, but the low accuracy makes an actual number meaningless.


LS100 specs:

Accuracy 0.001 to 0.999cd/m2 (or fL): ±2% ±2 digits of displayed value

1.000cd/m2 (or fL) or greater: ±2% ±1 digit of displayed value

So basically, if the inacurracy is +2% or +2 digits of displayed value then my fluctuation, .001 FL measured times .02 gives me .002 FL. I think thats pretty accurate if correct and would not sway my numbers enough to drastically decrease the final result.

Correct?

Cliff
overclkr is offline  
post #96 of 200 Old 09-13-2008, 02:09 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by odyssey View Post

The method of calculating CR is wrong. If you are trying to calculate the CR for white, the proportion of the three primaries is not equal, but determined by the formula for the standard. Green is close to 60% of the total.

It looks like there is an error in the blue reading for the 1st projector at least. Probably a missing decimal point or something like that. For the 2nd projector his numbers have green at about 75% of the light total. It could be wrong, but my memory is that blue should be around 7% and his numbers have it at 6.5% for the 2nd projector. So, red looks weak there.
Quote:
Originally Posted by odyssey View Post

The second issue is the accuracy of the LS-100 for the black measurements. You need about .050 fL for black to be accurate. For example, that .002 fL reading is +/- .002 fL. The CR for white is probably very high, higher than stated, but the low accuracy makes an actual number meaningless.

I considered that when looking at the numbers and normally it does matter a lot, but here it doesn't. The reason is that the black level readings for blue and green have little to do with his final number because the red is dominating for video black. Here are his numbers:

Red tube - .188, .033

Green tube - .002, .004

Blue tube - .001, .001

Without considering the margin of error that is 0.229. Taking the error of +-2 in the last digit and +-2% of the reading, that only takes the high end to about 0.246. That is a difference of only around 7%. That isn't enough to give all those digits of accuracy, but enough for 200k:1 to be correct if the readings as posted are right (still wondering about the blue tube for projector 1). I didn't do the math for the error for the lower side, but it should also be within about 10%.

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #97 of 200 Old 09-13-2008, 02:11 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by overclkr View Post

Darin, do you still have that LS110? The reason I ask, is if I had one that I could keep here for a week or so, I could really dive in and take all of the measurements and scenarios for you guys!

Unfortunately, it is one that I borrow and which has to be available pretty much any time if it is called for.

Are you sure there isn't a mistake in the blue reading for projector 1? Looks like it should maybe be 997 FL. It shouldn't be way brighter than the red tube. It should be dimmer.

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #98 of 200 Old 09-13-2008, 02:12 PM - Thread Starter
AVS Club Gold
 
Mark Petersen's Avatar
 
Join Date: Dec 2000
Posts: 4,671
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
Quote:
Originally Posted by odyssey View Post

The method of calculating CR is wrong. If you are trying to calculate the CR for white, the proportion of the three primaries is not equal, but determined by the formula for the standard. Green is close to 60% of the total.

True but isn't that what we are seeing from his data? I presume that he set everything for white 100 IRE and then masked each gun and measured them separately and we can see that green is the dominant component. At the bottom (black) I don't think there will be any recognizable ratio.

Quote:


The second issue is the accuracy of the LS-100 for the black measurements. You need about .050 fL for black to be accurate. For example, that .002 fL reading is +/- .002 fL. The CR for white is probably very high, higher than stated, but the low accuracy makes an actual number meaningless.

Good point. .002 is very low and what makes this sort of a measurement a challenge. I don't know what the accuracy of the LS-100 is but that does seem awfully low. Edit: Sorry, I can see that Cliff and Darin already responded to this...
Mark Petersen is offline  
post #99 of 200 Old 09-13-2008, 02:31 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Something is wrong with the black readings for green and that's where the potential error is. Are the black readings adjusted for ambient? Also, is the green CRT biased off completely? It would be helpful to have readings going up in one IRE steps from black or better, one digital value steps from digital 16.

The luma equation for Rec 709 is:
Y = 0.2126 R + 0.7152 G + 0.0722 B

Green is slightly above 70%. For Rec 601, it's slightly less than 60%. Green dominates and an accurate reading for the green component of black is important. Blue hardly matters and red is only modestly important.
odyssey is offline  
post #100 of 200 Old 09-13-2008, 02:42 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by overclkr View Post

LS100 specs:

Accuracy 0.001 to 0.999cd/m2 (or fL): ±2% ±2 digits of displayed value

1.000cd/m2 (or fL) or greater: ±2% ±1 digit of displayed value

So basically, if the inacurracy is +2% or +2 digits of displayed value then my fluctuation, .001 FL measured times .02 gives me .002 FL. I think thats pretty accurate if correct and would not sway my numbers enough to drastically decrease the final result.

Correct?

Cliff

The +/- 2 digits means that a .002 fL reading can be .000 fL to .004 fL. Obviously, even .004 fL will result in a very high CR, but you don't know what it actually is. Also, ambient could be .001 fL or .002 fL and you have to measure it and subtract it. If you can't find a measurement method with the LS-100 that gives a black reading of at least about .050 fL, the CR will be too inaccurate to claim an actual number. You can say that it's very high, or even that it's at least X, but not an accurate number.

Do the red CRTs emit the measured black levels when the bias adjustment is set to minimum?
odyssey is offline  
post #101 of 200 Old 09-13-2008, 02:46 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by darinp2 View Post

Unfortunately, it is one that I borrow and which has to be available pretty much any time if it is called for.

Are you sure there isn't a mistake in the blue reading for projector 1? Looks like it should maybe be 997 FL. It shouldn't be way brighter than the red tube. It should be dimmer.

--Darin

Big dog, I measured it over three times as I was kind of put off from this as well. 1209FL is what projector B measured vs. 9969FL for projector A.

I wish I could get that meter here for an extended amount of time.

Maybe I could persuade Ken to part with it for a while.

Cliff
overclkr is offline  
post #102 of 200 Old 09-13-2008, 02:52 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by odyssey View Post

The +/- 2 digits means that a .002 fL reading can be .000 fL to .004 fL. Obviously, even .004 fL will result in a very high CR, but you don't know what it actually is. Also, ambient could be .001 fL or .002 fL and you have to measure it and subtract it. If you can't find a measurement method with the LS-100 that gives a black reading of at least about .050 fL, the CR will be too inaccurate to claim an actual number. You can say that it's very high, or even that it's at least X, but not an accurate number.

I think it's close enough when your talking about over 200K on/off CR.

Where it wouldn't be close enough for me is if I have measured on/off 30K to one.

As far as ambient light. There is ZERO ambient light at black in my room. It's completely light controlled.

I'm still confused as to why I would need to raise my black level to .050 FL in order to get an accurate number. If it's measurable within 2% or +2 then I would consider my numbers to be accurate.

Please feel free to correct if I'm wrong.

******on your edit, YES if I lower the bias then my black level measurements will go down but then my image will not be right on red bias. Ken has to turn up the bias to those numbers to get the correct greyscale.*******

Cliff
overclkr is offline  
post #103 of 200 Old 09-13-2008, 02:57 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by odyssey View Post

The +/- 2 digits means that a .002 fL reading can be .000 fL to .004 fL. Obviously, even .004 fL will result in a very high CR, but you don't know what it actually is. Also, ambient could be .001 fL or .002 fL and you have to measure it and subtract it. If you can't find a measurement method with the LS-100 that gives a black reading of at least about .050 fL, the CR will be too inaccurate to claim an actual number. You can say that it's very high, or even that it's at least X, but not an accurate number.

I already addressed that. When one reading out of six is .188 and one is .033 the .002 you are talking about makes little difference. If you disagree then please say why you disagree and how much inaccuracy you think the +-2 in the last digit makes to the final CR number.

I doubt the ambient light is 0.001 or 0.002 when shooting straight into the tubes in Cliff's room, but they could check that by turning the projector off.

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #104 of 200 Old 09-13-2008, 03:24 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by darinp2 View Post

I already addressed that. When one reading out of six is .188 and one is .033 the .002 you are talking about makes little difference. If you disagree then please say why you disagree and how much inaccuracy you think the +-2 in the last digit makes to the final CR number.

I doubt the ambient light is 0.001 or 0.002 when shooting straight into the tubes in Cliff's room, but they could check that by turning the projector off.

--Darin

You are technically correct, but it means nothing. I can misadjust my G90 and have the CR at infinite and it would mean nothing. If one red CRTs is emitting 33x too much light at black and the other 188x, the CR measurement means nothing because the projectors are broken. If red CR is determining almost the entire white CR, something is obviously wrong. When you can ignore the green CR and it's supposed to be 70% of the total, something is obviously wrong.

The ambient measurement is important. It takes very little reflecting from the lens or CRT screen to have a large effect at these low levels. Were there absolutely no sources of light in the room except for the CRTs?
odyssey is offline  
post #105 of 200 Old 09-13-2008, 03:37 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by odyssey View Post

You are technically correct, but it means nothing. I can misadjust my G90 and have the CR at infinite and it would mean nothing. If one red CRTs is emitting 33x too much light at black and the other 188x, the CR measurement means nothing because the projectors are broken.

Nobody said nothing was wrong. Cliff already reported that something was wrong and his CR with things that way. Your argument about something like 100% error with the blue and green readings doesn't change that his margin of error in the total number is likely way less than that (probably) 10% error or less. And the numbers he got do mean something. They tell us within a reasonable error what kind of CR he is getting and they mean that even if you take the highest margin of error of the meter, if he can get close to the same CR with new red tubes as he is getting with the others now, he would be over 1 million to 1, just like I originally said. He wouldn't have a number that goes to very much accuracy, but would indicate things like that.
Quote:
Originally Posted by odyssey View Post

The ambient measurement is important. It takes very little reflecting from the lens or CRT screen to have a large effect at these low levels. Were there absolutely no sources of light in the room except for the CRTs?

The LS100 reads mostly from a very narrow beam. He could have some LEDs on and not be even close to 0.002 FL when shooting into the tubes. Cliff has valued his black levels and known about minimizing other lighting for a long time. I would have been very surprised if he had light in his room that would be in the same range as even the black levels from his green tubes to his screen, let alone affect the readings when shooting into the tubes close to that much.

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #106 of 200 Old 09-13-2008, 03:46 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by overclkr View Post

I think it's close enough when your talking about over 200K on/off CR.

Where it wouldn't be close enough for me is if I have measured on/off 30K to one.

As far as ambient light. There is ZERO ambient light at black in my room. It's completely light controlled.

I'm still confused as to why I would need to raise my black level to .050 FL in order to get an accurate number. If it's measurable within 2% or +2 then I would consider my numbers to be accurate.

Please feel free to correct if I'm wrong.

******on your edit, YES if I lower the bias then my black level measurements will go down but then my image will not be right on red bias. Ken has to turn up the bias to those numbers to get the correct greyscale.*******

Cliff

Lets assume that the white reading for green is 10,000 fL. If the black reading is .002 fL that again means that it's .000-.004 fL with the +/- 2 digit potential error. The CR is between 2.5M:1 and infinity. The .002 fL reading gives you 5M:1, + infinity and -50%. If the black reading is .050 fL that means that it's .048-.052 fL. The CR is between 192,308:1 and 208,333:1. It would be 200,000:1 +/- 4.2%. I was aiming at 5% accuracy and that's why I picked .050 fL.

Are you able to turn down the green bias and have the LS-100 read 0?
odyssey is offline  
post #107 of 200 Old 09-13-2008, 03:48 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by overclkr View Post

I'm still confused as to why I would need to raise my black level to .050 FL in order to get an accurate number.
If it's measurable within 2% or +2 then I would consider my numbers to be accurate.

The point is basically that with .050 you have +-3 in the last digital (one from the percentage and 2 more) and so that is what you need to be within about 6% margin of error. But you don't need to be within 6% margin of error. You just need to realize that if you get something like a million to 1 it might be plus or minus 20%, or even higher depending on the readings. Like if you got readings around 0.012 for black you would probably be around +-20% in your final figure. Your average was actually higher than that and so I would say you got to within a reasonable range for what the CR is that you are getting. 200k:1 plus or minus around 10%. Of course with the red problem like you pointed out originally.

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #108 of 200 Old 09-13-2008, 04:01 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
The main problem is that the measurement is not the CR of white but almost entirely that of red only. As I already said, you can claim that the white CR is very high or that it's at least X, but you can't claim what it actually is.

I am still not convinced that there is a problem with the red CRTs. They are probably shot, but I would still check the bias of the green CRTs. Using this method, the reading of black for the green CRTs should be higher than .002 fL.

Are the x and y readings for 100 IRE white from the screen available?
odyssey is offline  
post #109 of 200 Old 09-13-2008, 06:30 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by odyssey View Post

Are you able to turn down the green bias and have the LS-100 read 0?

I will try it when I get the meter back. I have a list of things to do in here now.
overclkr is offline  
post #110 of 200 Old 09-13-2008, 06:33 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by darinp2 View Post

The point is basically that with .050 you have +-3 in the last digital (one from the percentage and 2 more) and so that is what you need to be within about 6% margin of error. But you don't need to be within 6% margin of error. You just need to realize that if you get something like a million to 1 it might be plus or minus 20%, or even higher depending on the readings. Like if you got readings around 0.012 for black you would probably be around +-20% in your final figure. Your average was actually higher than that and so I would say you got to within a reasonable range for what the CR is that you are getting. 200k:1 plus or minus around 10%. Of course with the red problem like you pointed out originally.

--Darin

I agree 100% big dog fo sho. I gotta tame those rasters.

Even with this red issue, I gotta tell ya big dog, and you know, I'm lovin' this setup.

200K to one? THAT IS SICK!!!!!!

Cliffy
overclkr is offline  
post #111 of 200 Old 09-13-2008, 07:12 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by overclkr View Post

I will try it when I get the meter back. I have a list of things to do in here now.

Before you replace the red CRTs, make sure that the green CRTs are not biased too low.

I want to clarify a few things. The method of adding up the six "white" readings and dividing by the six "black" readings is correct. In a perfect system, the CR of the three colors will be the same, but they never are with CRTs. The more the three CRs vary, the larger the error in grey scale tracking. One calibration method to improve grey scale tracking is to adjust bias so that the three CRs are the same. You leave the lowest CR color alone and raise the bias of the other two to match its CR.

The luma equation does not precisely define the proportion of the primaries at D65 for the G90 and you need to measure that. You can set something like 70 IRE white to D65 from the screen and then measure the proportions from the CRT faces. If the measured proportions at other luminance levels vary from this, it's an indication of the tracking error. While the luminance proportion of the primaries at black may not matter that much, it can be an indication of what's hapenning above black at luminance levels that do matter.

If you don't have enough complications already, I should mention that the LS-100 accuracy spec is for Illuminant A. The errors for measurement of the G90 primaries can be much larger. I think that you should be satisfied that your CR is very high and give up on trying to measure it accurately.
odyssey is offline  
post #112 of 200 Old 09-13-2008, 08:49 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by odyssey View Post

Before you replace the red CRTs, make sure that the green CRTs are not biased too low.

I want to clarify a few things. The method of adding up the six "white" readings and dividing by the six "black" readings is correct. In a perfect system, the CR of the three colors will be the same, but they never are with CRTs. The more the three CRs vary, the larger the error in grey scale tracking. One calibration method to improve grey scale tracking is to adjust bias so that the three CRs are the same. You leave the lowest CR color alone and raise the bias of the other two to match its CR.

The luma equation does not precisely define the proportion of the primaries at D65 for the G90 and you need to measure that. You can set something like 70 IRE white to D65 from the screen and then measure the proportions from the CRT faces. If the measured proportions at other luminance levels vary from this, it's an indication of the tracking error. While the luminance proportion of the primaries at black may not matter that much, it can be an indication of what's hapenning above black at luminance levels that do matter.

If you don't have enough complications already, I should mention that the LS-100 accuracy spec is for Illuminant A. The errors for measurement of the G90 primaries can be much larger. I think that you should be satisfied that your CR is very high and give up on trying to measure it accurately.

Ok, let me make a long story short here. Are you telling me that my primary colors cannot be accurate? Are you saying that my G90 stack cannot match the colors of the latest DCI specs? Not sure if I'm reading you right here hence the question?

Cliff
overclkr is offline  
post #113 of 200 Old 09-13-2008, 08:54 PM
AVS Club Gold
 
Art Sonneborn's Avatar
 
Join Date: Jun 2000
Location: Battle Creek,MI USA
Posts: 22,307
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 13 Post(s)
Liked: 49
A thing of beauty !!!!!!!! We all know what sequential contrast adds to image depth. Just look at ANSI CR and on /off with those G90s. This gives something for the fixed pixel device manufacturers to aspire to. Some of us have seen it and enjoyed it. Thanks for the effort.

Art

My HT


iRule rules my theater
 

"If she's amazing she won't be easy,if she's easy she won't be amazing"

 

Bob Marley

Art Sonneborn is offline  
post #114 of 200 Old 09-13-2008, 09:22 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by overclkr View Post

Ok, let me make a long story short here. Are you telling me that my primary colors cannot be accurate? Are you saying that my G90 stack cannot match the colors of the latest DCI specs? Not sure if I'm reading you right here hence the question?

Cliff

No, that's not what I said. The G90 primaries are not accurate for either consumer SD or HD and are far off from DCI primaries, but those are different topics.

You really should not be concerned about sequential CR. If you want to be concerned about something, it should be ANSI CR and color accuracy. The measurements show problems with grey scale tracking...the 100 IRE blue is much too high and the low end much too red. How much of a problem this is depends on how far up the luminance scale this goes.
odyssey is offline  
post #115 of 200 Old 09-13-2008, 09:23 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by Art Sonneborn View Post

A thing of beauty !!!!!!!! We all know what sequential contrast adds to image depth. Just look at ANSI CR and on /off with those G90s.

You may have seen where I posted that I asked the main teacher in the THX booth how he thought a G90 would do for black performance and he told me it would do very poorly (or words to that effect). This was just after telling me that the 30k:1 for the JVC RS20 was just marketing and didn't mean anything. Makes me wonder about this industry. At least the manufacturers have been giving us higher and higher on/off CRs even as some people in the industry have been downplaying on/off CR.

Cliff,

There is one thing I'm really curious about. There has been some discussion about how easy or hard it is to see colors down low. My experience is that a skew toward blue seems relatively easy to see at the black levels of some of these digitals (I'm thinking LCDs of a year or two ago mostly), but red is pretty difficult to see down low. If you put up a blackout (like watch the one in Finding Nemo) is the skew toward red extremely obvious with your setup when looking at the screen?

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #116 of 200 Old 09-13-2008, 09:26 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by odyssey View Post

You really should not be concerned about sequential CR. If you want to be concerned about something, it should be ANSI CR and color accuracy.

Are you referring to just the G90, or images in general when you say that? I doubt Cliff can do much about the ANSI CR of his G90s, so it would seem like you aren't referring to those, but I'm not sure. If images in general, why do you say that?

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #117 of 200 Old 09-13-2008, 09:42 PM
AVS Special Member
 
odyssey's Avatar
 
Join Date: Jan 2001
Posts: 1,088
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by darinp2 View Post

Are you referring to just the G90, or images in general when you say that? I doubt Cliff can do much about the ANSI CR of his G90s, so it would seem like you aren't referring to those, but I'm not sure. If images in general, why do you say that?

--Darin

Do you think that an 80 ANSI CR, especially in a good room, is not something to be concerned about and does not have a significant impact on overall image quality? In the case of the G90, it's also a very good proxy for MTF, which is very low at the higher spatial frequencies. In any case, the ANSI CR of the G90 without room effects, should be about 150-175:1.
odyssey is offline  
post #118 of 200 Old 09-13-2008, 09:47 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,231
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 57 Post(s)
Liked: 79
Quote:
Originally Posted by odyssey View Post

Do you think that an 80 ANSI CR, especially in a good room, is not something to be concerned about and does not have a significant impact on overall image quality? In the case of the G90, it's also a very good proxy for MTF, which is very low at the higher spatial frequencies.

So, what do you want him to do? The G90 has maybe 125:1 at best from what I've gathered and I'm pretty sure Cliff has darkened his room pretty well. I also don't know how it was measured as the LS110 I've used can get some light from the brighter rectangles and read lower than what is actually coming off the screen. With the LS110 I found that it couldn't measure more than about 200:1 ANSI CR without a light pipe on the front of the lightmeter to make sure light from brighter rectangles is minimized. I don't know about the LS100 model in regards to this, but it measures from an even bigger angle than the LS110 (1 degree instead of just 1/3rd of a degree) and so might be even worse for measuring ANSI CR.

I'm trying to figure out what you are recommending Cliff do. Do you think his black performance overall would be better with a 3k:1 on/off CR and 300:1 ANSI CR LCD, since the LCD would kill his G90s for ANSI CR? Basically, I'm trying to figure out if you are saying that ANSI CR is more important than sequential CR along with what you think Cliff should do. He can control his sequential CR to a certain degree with the way he has things setup, but unless his room is light enough I don't think there is a whole lot he can do about the ANSI CR other than switch technologies.

EDIT: I see that you added that the G90 should be about 150:1-175:1 without room effects. That is higher than numbers I've seen before and much higher than I remember measuring for a G70 trying to take room effects out. I realize the G70 is a different model though. Have you measured that much yourself with your G90 or another G90?

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is online now  
post #119 of 200 Old 09-13-2008, 10:19 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by odyssey View Post

No, that's not what I said. The G90 primaries are not accurate for either consumer SD or HD and are far off from DCI primaries, but those are different topics.

You really should not be concerned about sequential CR. If you want to be concerned about something, it should be ANSI CR and color accuracy. The measurements show problems with grey scale tracking...the 100 IRE blue is much too high and the low end much too red. How much of a problem this is depends on how far up the luminance scale this goes.

BITE!!!! SWEET! You bit!

Don't talk smack you cant back up dude. You will be so badly misproven that you may as well go ahead and delete your account when you decide to back up those comments and I disprove them.

You are FLAT OUT WRONG when it comes to ANSI CR. I have seen the BEST so far in my eyes and there is no way it is going to match what my stack is capable of. While there are things about HIGH ANSI projectors like the 3 chippers that I have proclaimed many of times smoke the stack in certain material, it's the OVERALL performance of when I watch film that I prefer with the stack.

It's much more fluid, responsive, and just flat out more pleasing to my eyes and I run at 60hz! When my eyes have to adjust much less to material and yet retain that HOLY CRAP factor especially with on/off performance like I am getting now, there is just nothing out there that can match it and I will take no less.

Projector manufacturers need to take heed and shoot for this kind of performance even if it's a big ass 3 chipper that can crank over 20 foot lamberts on a big ass screen. I want to turn off ALL THE LIGHTS and when I go black on the screen, it's as good as the source can provide whether it be good or bad, cause when it's good, oh my GOD is it good.

Not that I am discounting ANSI performance as yes, it does have a HUGE advantage especially if your running BIG ASS screens. That is where ANSI really shines big time.

If you however reduce that "big ass" screen size to a more manageable level like mine, in a smaller 100% light controlled room, then it becomes a lot less important because in a room like mine, I actually will not except ANY less than the on/off performance that I get now regardless of the ANSI number.

Also keep this in mind, MOST home theaters in AVS do not use screens even as large as mine.

I just cannot imagine even a smaller screen than mine and looking at grey vs. black in an optimum viewing environment and I mean especially without using "TRICKS".

I'm talking unity gain here with the projector doing all of the work instead of pairing it with the screen. I am getting this with the G90's and I expect nothing less.

Cliff
overclkr is offline  
post #120 of 200 Old 09-13-2008, 10:31 PM
AVS Special Member
 
overclkr's Avatar
 
Join Date: Jul 2003
Location: Chicago
Posts: 7,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Quote:
Originally Posted by Art Sonneborn View Post

A thing of beauty !!!!!!!! We all know what sequential contrast adds to image depth. Just look at ANSI CR and on /off with those G90s. This gives something for the fixed pixel device manufacturers to aspire to. Some of us have seen it and enjoyed it. Thanks for the effort.

Art

It's thanks to you and Ken big dog. It is a thing of beauty and you know as well as I how our tastes differ these days and am so glad that we can both appreciate eachother as not only good friends, but compadres.

You have been an integral part of my learning experience here and I really have you to thank for so much.

The best part is you are in 100% agreement with me when it comes to on/off and low APL performance.

Make it accurate, make the projector do it, and shoot for the whole spectrum. We are all happy with bright to an extent. But we want black as well.

I would like to ask you a question if you dont mind and am curious if you have measured the ANSI performance of your HT5000 with Ken?

If not, can we and post it here for inquiring minds?

Cliffy
overclkr is offline  
Reply Digital Hi-End Projectors - $3,000+ USD MSRP

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off