PR-655 Low Light Readings Inconsistent - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 29 Old 02-29-2012, 01:16 AM - Thread Starter
Newbie
 
tech_df's Avatar
 
Join Date: Feb 2012
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Hello,

I'm getting very inconsistent data from a PR-655 reading a Cinemage B420 display at low light levels (3.5cd and higher). Occasionally, I get "low light level" errors (and no reading). Other times I get a deviation between readings as high as .0068. Even when averaging 20 or more readings, the deviation is too high to trust.

The PR-655 spec claims sensitivity down to 3.4cd. I've heard talk of "difficulty in the shadows" but I assumed that meant longer read times were required, not that a lower level of accuracy would result. Does anyone have direct experience that might suggest a realistic expectation for this device?

Mike Lary
Digital Factory
Seoul, Korea
http://digitalfactory.kr/
tech_df is offline  
Sponsored Links
Advertisement
 
post #2 of 29 Old 03-01-2012, 11:58 AM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 227
The low light capability of spectro devices is limited by the nature of the device. In spite of how expensive they are, they don't do low light levels with any reliability. Taking multiple readings and averaging them is a waste of time since none of the readings are accurate except by accident. In fact, taking multiple measurements and averaging them is like playing a slot machine and averaging 20 losing pulls trying to come up with a winning combination without ever actually pulling a winning combo. It's just not going to happen.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #3 of 29 Old 03-01-2012, 02:39 PM
Advanced Member
 
CalWldLif's Avatar
 
Join Date: Feb 2011
Location: South Bay'Los Angeles California.
Posts: 518
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 29
how low can you trust it?
if the panel has .010 mll
will the i1Pro be reliable across all levels?

Loving D65
CalWldLif is offline  
post #4 of 29 Old 03-01-2012, 03:21 PM
AVS Special Member
 
PlasmaPZ80U's Avatar
 
Join Date: Feb 2009
Posts: 7,252
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 106 Post(s)
Liked: 201
having a good colorimeter to profile with the spectro (like the new D3 or C6), can be very helpful for such measurements

Of course, if you're a pro calibrator and need something high end (given the spectro in question is a PR-655), the Klein K-10 represents the top end of all colorimeters as far as I know. There are others meters in between these two extremes, though I don't know too much about those.
PlasmaPZ80U is offline  
post #5 of 29 Old 03-01-2012, 03:49 PM
Member
 
ZandarKoad's Avatar
 
Join Date: Feb 2009
Posts: 136
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Doug Blackburn View Post

The low light capability of spectro devices is limited by the nature of the device. In spite of how expensive they are, they don't do low light levels with any reliability. Taking multiple readings and averaging them is a waste of time since none of the readings are accurate except by accident. In fact, taking multiple measurements and averaging them is like playing a slot machine and averaging 20 losing pulls trying to come up with a winning combination without ever actually pulling a winning combo. It's just not going to happen.

Not necessarily. An instrument can indeed have high accuracy and low precision. I'm not saying that spectros do, but I can certainly conceive that they may. This is a question I posed elsewhere and never received a satisfactory response (which is not to say anyone owes me such). Speaking of the i1Pro at low light levels:

Quote:


"Also, how would this "inaccuracy" manifest? If I let the spectro sample 5 to 10 times at low light, and it's readings were stable, could I rely on their accuracy? Or would it maintain high stability and precision but loose accuracy? I would think it would loose precision rather than accuracy..."

For example, if you fire a gun 500 times it could miss the bullseye every time, but have a perfectly balanced spread around the target. In this case, the gun would have low precision but high accuracy.

If the bullets hit the same exact spot every time, but the spot was not the bullseye, then the gun would have high precision but low accuracy.

If the bullet holes were spread out, and not centered around the target, then the gun would have low precision and low accuracy.

If the bullets struck the bullseye every time, and were not spread out, then it would have high precision and high accuracy.

Which of the four scenarios describes the spectro, I have no idea. I just don't know enough about how these doohickies work. I can tell you this: someday I will know.
ZandarKoad is offline  
post #6 of 29 Old 03-01-2012, 05:37 PM - Thread Starter
Newbie
 
tech_df's Avatar
 
Join Date: Feb 2012
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Thanks everyone for your responses.

Doug Blackburn: I agree that averaging multiple readings in this instance is of no use. I did several averages of 20 or more readings and each time had a drastically different result (enough to be in or outside the Rec. 709 spec in either direction).

PlasmaPZ80U: I've considered the K-10 but I've been leaning toward a high end colorimeter because of the presumed higher level of accuracy. I've heard of some people using the K-10 and a PR-655 in tandem (the K-10 for fast patch reading and the PR-655 to verify the results), but this raises the question of how I can trust unverifiable readings in the shadows.

ZandarKoad: In this case, the device behavior was high precision and high accuracy until the bottom five steps of a 17 step grayscale target. Then, initially, I was able to get consistent averages from multiple batch readings. The darker the patches, the less accurate and less consistent the readings became.

Mike Lary
Digital Factory
Seoul, Korea
http://digitalfactory.kr/
tech_df is offline  
post #7 of 29 Old 03-01-2012, 05:48 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,636
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 40 Post(s)
Liked: 174
Typically what you do with a K-10 and a PR-655 would be to profile the K-10 from the PR-655, essentially teaching the K-10 about what the PR can see that the K-10 can't at high stimulus levels.

Then you use the K-10 through out and possibly double check with the PR-655 at the end.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #8 of 29 Old 03-01-2012, 06:02 PM - Thread Starter
Newbie
 
tech_df's Avatar
 
Join Date: Feb 2012
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by sotti View Post

Typically what you do with a K-10 and a PR-655 would be to profile the K-10 from the PR-655, essentially teaching the K-10 about what the PR can see that the K-10 can't at high stimulus levels.

Then you use the K-10 through out and possibly double check with the PR-655 at the end.

Can you please elaborate on what you mean by profiling the K-10? Is this a software-assisted process?

Also, do you have difficulty with the PR-655 at low light levels? If so, are there certain displays that are more problematic?

Mike Lary
Digital Factory
Seoul, Korea
http://digitalfactory.kr/
tech_df is offline  
post #9 of 29 Old 03-01-2012, 06:06 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,636
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 40 Post(s)
Liked: 174
Quote:
Originally Posted by tech_df View Post

Can you please elaborate on what you mean by profiling the K-10? Is this a software-assisted process?

Also, do you have difficulty with the PR-655 at low light levels? If so, are there certain displays that are more problematic?

I don't use a PR-655 very often, only a few times in testing.

It's just a light level issue, I don't think any specific displays would be better or worse, but darker ones like older plasmas or projectors would be troublesome.

Profiling is typically software assited. It takes a Red, Green, Blue and White reading and compares the results from the reference meter to the target meter and generates a correction matrix that adjusts the raw data coming out of the profiled colorimeter (a K-10 being a colorimeter v a PR-655 being a spectroradiometer).

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #10 of 29 Old 03-01-2012, 06:17 PM - Thread Starter
Newbie
 
tech_df's Avatar
 
Join Date: Feb 2012
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by sotti View Post

Profiling is typically software assited. It takes a Red, Green, Blue and White reading and compares the results from the reference meter to the target meter and generates a correction matrix that adjusts the raw data coming out of the profiled colorimeter (a K-10 being a colorimeter v a PR-655 being a spectroradiometer).

Thanks for that explanation.
What software do you use to profile the K-10 to the PR-655?

Mike Lary
Digital Factory
Seoul, Korea
http://digitalfactory.kr/
tech_df is offline  
post #11 of 29 Old 03-01-2012, 06:19 PM
AVS Special Member
 
turbe's Avatar
 
Join Date: Feb 2001
Location: Lake Tahoe, NV
Posts: 4,434
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 88
I personally recommend Klein's own software to profile the K-10 to a PR..

Need to find a Professional Calibrator? Click Here to PM me with your Display & City

Calibrator List - Pioneer ISFccc Interface

Calibration Reports - Pioneer

 

ControlCAL™
Designed by Calibrators for Calibrators™

No need to fumble through the Display's Menu with its Remote Control™

turbe is offline  
post #12 of 29 Old 03-01-2012, 06:44 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,636
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 40 Post(s)
Liked: 174
Quote:
Originally Posted by tech_df View Post

Thanks for that explanation.
What software do you use to profile the K-10 to the PR-655?

We'll I'd use CalMAN, because that's the calibration package I'm most familiar with

That probably has something to do with being the lead dev for the software. The K-10 software is just as capable for creating profiles, but isn't nearly as complete a package as CalMAN. It all really depends on what your needs are.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #13 of 29 Old 03-01-2012, 07:30 PM
AVS Special Member
 
turbe's Avatar
 
Join Date: Feb 2001
Location: Lake Tahoe, NV
Posts: 4,434
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 88
bit of confusion here..

CalMAN can use the actual calibration tables/memories in the K-10, CalMAN can also do the Profiling but I recommend using Klein's own software to do the actual profiling of the K-10 to the PR (not CalMAN to do the Profiling).. After profiling is done using Klein's software, use CalMAN and select the table/memory # in the Klein you profiled with.

Don't use the Klein software for the display calibration..

Need to find a Professional Calibrator? Click Here to PM me with your Display & City

Calibrator List - Pioneer ISFccc Interface

Calibration Reports - Pioneer

 

ControlCAL™
Designed by Calibrators for Calibrators™

No need to fumble through the Display's Menu with its Remote Control™

turbe is offline  
post #14 of 29 Old 03-02-2012, 01:08 AM - Thread Starter
Newbie
 
tech_df's Avatar
 
Join Date: Feb 2012
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by turbe View Post

bit of confusion here..

CalMAN can use the actual calibration tables/memories in the K-10, CalMAN can also do the Profiling but I recommend using Klein's own software to do the actual profiling of the K-10 to the PR (not CalMAN to do the Profiling).. After profiling is done using Klein's software, use CalMAN and select the table/memory # in the Klein you profiled with.

Don't use the Klein software for the display calibration..

Thanks for that clarification. Display calibration/profiling software isn't my focus right now. I'm strictly looking to determine how to get valid data from the shadows. If creating a matrix to force a colorimeter to mimic the behavior of a spectro would get me there, I'll look into that. However, since the verification of low level readings can't be verified I don't see that being an optimal solution.

This has been an ongoing discussion between me and our local PR authorized reseller. After testing two PR-655s side by side and seeing the problem, they suggested that the readings in question were below the sensitivity threshold of the device and that the luminance level reading the device gave me (which is within the device spec) was grossly off the mark. This makes sense, but I don't understand how this could be the reality given the popularity of the display in question and of the 655 in color critical environments. Their suggestion was that we use a higher end model (670 or 730).

I calibrated some TVLogic displays without any problems (aside from the primaries being off), so I'm wondering if the B420 problem has to do with the black level requirement. Our viewing environment is very subdued, so it doesn't seem there's a conflict with how we set the black level. I'd like to hear anyone's thoughts on that.

Mike Lary
Digital Factory
Seoul, Korea
http://digitalfactory.kr/
tech_df is offline  
post #15 of 29 Old 03-02-2012, 04:23 AM
AVS Special Member
 
buzzard767's Avatar
 
Join Date: Dec 2005
Location: Naples, FL & Wausau, WI
Posts: 4,002
Mentioned: 5 Post(s)
Tagged: 1 Thread(s)
Quoted: 385 Post(s)
Liked: 903
You might want to look into the Jeti Specbos 1211

Buzz
THX Certified Video Calibrator

 

buzzard767 is online now  
post #16 of 29 Old 03-02-2012, 05:59 AM - Thread Starter
Newbie
 
tech_df's Avatar
 
Join Date: Feb 2012
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by buzzard767 View Post

You might want to look into the Jeti Specbos 1211

What have you found to be the Specbos' capability, range-wise, in practical use?

Mike Lary
Digital Factory
Seoul, Korea
http://digitalfactory.kr/
tech_df is offline  
post #17 of 29 Old 03-02-2012, 06:11 AM
AVS Special Member
 
buzzard767's Avatar
 
Join Date: Dec 2005
Location: Naples, FL & Wausau, WI
Posts: 4,002
Mentioned: 5 Post(s)
Tagged: 1 Thread(s)
Quoted: 385 Post(s)
Liked: 903
Quote:
Originally Posted by tech_df View Post

What have you found to be the Specbos' capability, range-wise, in practical use?

My only use was to compare my i1Pro to one of them. Forum member Chad B has one so you might want to PM him.

Buzz
THX Certified Video Calibrator

 

buzzard767 is online now  
post #18 of 29 Old 03-02-2012, 10:03 AM
Member
 
subraman's Avatar
 
Join Date: Jul 2002
Posts: 116
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Doug Blackburn View Post

The low light capability of spectro devices is limited by the nature of the device. In spite of how expensive they are, they don't do low light levels with any reliability. Taking multiple readings and averaging them is a waste of time since none of the readings are accurate except by accident. In fact, taking multiple measurements and averaging them is like playing a slot machine and averaging 20 losing pulls trying to come up with a winning combination without ever actually pulling a winning combo. It's just not going to happen.

The data averaging Photo Research performs is not quite this simple minded. They average the raw uncorrected measured data for the cycles specified and only then are the photometric and colorimetric results calculated (from the averaged data). The uncertainty (in the measurement) improves by the square root of the number of measurements averaged. (eg. averaging 4 cycles has 2 times the signal-to-noise ratio as a single measurement.)
subraman is online now  
post #19 of 29 Old 03-02-2012, 10:06 AM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 227
Quote:
Originally Posted by ZandarKoad View Post

Not necessarily. An instrument can indeed have high accuracy and low precision. I'm not saying that spectros do, but I can certainly conceive that they may. This is a question I posed elsewhere and never received a satisfactory response (which is not to say anyone owes me such). Speaking of the i1Pro at low light levels:



For example, if you fire a gun 500 times it could miss the bullseye every time, but have a perfectly balanced spread around the target. In this case, the gun would have low precision but high accuracy.

If the bullets hit the same exact spot every time, but the spot was not the bullseye, then the gun would have high precision but low accuracy.

If the bullet holes were spread out, and not centered around the target, then the gun would have low precision and low accuracy.

If the bullets struck the bullseye every time, and were not spread out, then it would have high precision and high accuracy.

Which of the four scenarios describes the spectro, I have no idea. I just don't know enough about how these doohickies work. I can tell you this: someday I will know.

None of this has any application to this particular question. We're not talking about "any meter" we are talking about a spectroradiometer making readings at lower light levels than it is capable of measuring. My analogy of averaging 20 losing pulls on a slot machine trying to come up with a winning combo is appropriate because that would be the exact scenario you would encounter in this specific set of circumstances. It's not a matter of precision OR accuracy... it's a matter of using the meter outside the range in which it is capable of operating... outside the meter's range of operation.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #20 of 29 Old 03-03-2012, 08:24 AM
AVS Special Member
 
Chad B's Avatar
 
Join Date: Oct 2002
Location: Piqua, OH
Posts: 2,087
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 80 Post(s)
Liked: 416
Quote:
Originally Posted by buzzard767 View Post

My only use was to compare my i1Pro to one of them. Forum member Chad B has one so you might want to PM him.

It's very software and meter FW dependent at the low end.
I just updated the Jeti to the latest FW 1.52 and did some testing with the latest CalMAN 4.6 beta. With the right settings, I got very good repeatability and believably accurate results at 10% on my front projector at .044 fL. Reading at that level took about 45 sec or so. That's pretty incredible performance.
I used to use an i1Pro, and I can say without a doubt the Jeti 1211 gives consistent and believable results at levels far, far below what the i1Pro does.

ISF/THX calibrator, HDTVbyChadB
Jeti 1211

Latest reviews:
Sony 65X950BLG 65UB9800Samsung UN65HU8550UN65HU9000, UN85HU8550, LG 55EC9300
Chad B is online now  
post #21 of 29 Old 03-03-2012, 10:51 AM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 227
Quote:
Originally Posted by subraman View Post

The data averaging Photo Research performs is not quite this simple minded. They average the raw uncorrected measured data for the cycles specified and only then are the photometric and colorimetric results calculated (from the averaged data). The uncertainty (in the measurement) improves by the square root of the number of measurements averaged. (eg. averaging 4 cycles has 2 times the signal-to-noise ratio as a single measurement.)

It doesn't matter WHAT P.R. does with internal averaging of readings. Once the meter reaches its low light measurement limit, measurements are just random crap. And averaging 3 or 10 or 100 crap measurements still yields a crap measurement.

Averaging measurements only has value if you are still within the operating range of the instrument. Period.

And we weren't even talking about how measurements are averaged internally by the meter. We were talking about the operator averaging readings provided the meter in conditions where the light level is lower than the capability of the meter.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #22 of 29 Old 03-03-2012, 01:10 PM
Member
 
subraman's Avatar
 
Join Date: Jul 2002
Posts: 116
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Doug Blackburn View Post

It doesn't matter WHAT P.R. does with internal averaging of readings. Once the meter reaches its low light measurement limit, measurements are just random crap. And averaging 3 or 10 or 100 crap measurements still yields a crap measurement.

Averaging measurements only has value if you are still within the operating range of the instrument. Period.

And we weren't even talking about how measurements are averaged internally by the meter. We were talking about the operator averaging readings provided the meter in conditions where the light level is lower than the capability of the meter.

Only an idiot or someone who watches too many CSI type shows (and believes what he sees) would think that you can magically pull a signal out of a totally noisy measurement. But when you are just at the borderline situation where there is still a signal that is not totally buried under random noise, averaging is a viable solution, when used by someone who understands the signal and the test conditions. This is not a PR original, but scientists (who know a lot more about it than you do) have been using this technique for a long time.
I see no explicit mention anywhere of how the averaging was performed. You may have jumped to one conclusion, but I am pretty sure the OP was talking about letting the meter do the averaging. I personally wouldn't think anyone would spend their time manually averaging 20 or more readings!
subraman is online now  
post #23 of 29 Old 03-03-2012, 05:14 PM
Advanced Member
 
fastl's Avatar
 
Join Date: Sep 2007
Location: Boston
Posts: 588
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 13 Post(s)
Liked: 11
I'm also curious as to how you can ACCURATELY measure really low light levels. Lets say that you want to ACCURATELY measure display gamma down to 2 percent of peak white (120 cd/m^2) and that your display gamma is nominally 2.4. Plug the numbers into the following equation, and it indicates that the light output level at 2 percent is 0.01 cd/m^2. What (affordable) system can ACCURATELY measure 0.01 cd/m^2? Also begs the question as to how gamma measurements are being made. Is everyone measuring down to say 20 percent, and then ignoring the remainder of the dynamic range?

120*(0.02)^2.4 = 0.01 cd/m^2
fastl is offline  
post #24 of 29 Old 03-04-2012, 09:51 AM
Member
 
subraman's Avatar
 
Join Date: Jul 2002
Posts: 116
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Finally, I have always wondered if signal averaging (not result averaging) is part of Calman's secret low-light handler for the I1-pro. Of course we'll never know....
subraman is online now  
post #25 of 29 Old 03-04-2012, 10:33 AM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 227
Quote:
Originally Posted by subraman View Post

Only an idiot or someone who watches too many CSI type shows (and believes what he sees) would think that you can magically pull a signal out of a totally noisy measurement. But when you are just at the borderline situation where there is still a signal that is not totally buried under random noise, averaging is a viable solution, when used by someone who understands the signal and the test conditions. This is not a PR original, but scientists (who know a lot more about it than you do) have been using this technique for a long time.
I see no explicit mention anywhere of how the averaging was performed. You may have jumped to one conclusion, but I am pretty sure the OP was talking about letting the meter do the averaging. I personally wouldn't think anyone would spend their time manually averaging 20 or more readings!

You are making an argument that was not part of the original inquiry and not part of any of my responses. So you're arguing with yourself. I never - EVER - indicated in any post that I was addressing the "gray area" between enough light for accurate measurements and insufficient light for anything but a meaningless measurement. Once the readings of a spectro meter start jumping around radically... the light level is too dark and the readings are meaningless. Period. It doesn't matter if the meter is averaging these bad readings or not... the readings are still bad. If the spectro device is in the "gray area" where it's borderline failing to take accurate measurements, the measurements will vary, but they will continue to cluster towards something... for example, the readings may jump around, but would always be over towards "too red" (or some other color). But even then "averaging" several readings that are all "too red" might or might not be helpful/useful there are too many "it depends" variables. Once the readings begin jumping around... too red for one reading, too blue for another, too yellow for a 3rd... you're below the limit of the meter and no amount of averaging readings will make them any more accurate than the next bad reading. But because the meter is likely to be averaging several readings, you may not realize the first one was too red, the second too blue, the third too yellow, etc. So the averaging may provide a false sense that the meter is doing something useful when it is not.

Sometime averaging readings is useful... especially in materials testing environments when you are trying to determine the strength or lifespan or some other physical property of something. In meters... averages can be very bad/misleading. It depends on what is being measured and on what precision is acceptable.

There's also the issue of the light being measured... blue light is only about 7% of what we see and measure when we are measuring white or a shade of gray. So blue errors will, more than likely, become serious before errors involving red (approx 21% of the light we see for neutral gray or white) or green (about 72% of the light making up gray or white).

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #26 of 29 Old 03-04-2012, 10:43 AM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 227
Quote:
Originally Posted by fastl View Post

I'm also curious as to how you can ACCURATELY measure really low light levels. Lets say that you want to ACCURATELY measure display gamma down to 2 percent of peak white (120 cd/m^2) and that your display gamma is nominally 2.4. Plug the numbers into the following equation, and it indicates that the light output level at 2 percent is 0.01 cd/m^2. What (affordable) system can ACCURATELY measure 0.01 cd/m^2? Also begs the question as to how gamma measurements are being made. Is everyone measuring down to say 20 percent, and then ignoring the remainder of the dynamic range?

120*(0.02)^2.4 = 0.01 cd/m^2

It all depends on how dark you want to measure and whether you are interested only in luminance or if you also need to know if red, green, and blue are present in the proper proportions. You can EASILY spend $3000+ on a luminance meter sensitive enough to measure 0% to 5% white on panel displays and projectors. For ~$200 you can get a luminance meter that might get you from 1.5% white and higher. If you want accurate information about color balance at those levels... you'll probably need a one of the newer colorimeters selling for circa $600-$800 (like the Chroma 6 Spectracal sells). One of those should be good down to 5% white no problem... might not measure 0% on a dark panel or projector, but may be OK at 2% or even a little lower depending on how dark the display is. BUT... the colorimeter might not be accurate at those levels unless it is fairly new, recently calibrated, or characterized against an accurate specro meter.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #27 of 29 Old 03-04-2012, 11:14 AM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,636
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 40 Post(s)
Liked: 174
Quote:
Originally Posted by Doug Blackburn View Post

There's also the issue of the light being measured... blue light is only about 7% of what we see and measure when we are measuring white or a shade of gray. So blue errors will, more than likely, become serious before errors involving red (approx 21% of the light we see for neutral gray or white) or green (about 72% of the light making up gray or white).

That's not entirely accurate.

What a spectroradiometer is measuring is watts per steradian per nm. The watts of energy of 420nm are just as easy to measure as energy at 500nm or 720nm. The color matching function we lay over the spectrum is what we then interpret as luminance or the lack of. The intesity of the light (in watts) is only thing the spectrometer sees.



you can see that the green curve which is the Y curve runs the entire length of the visible spectrum, but is very small at the extremes. This doesn't mean that the X or Z curves can't be measuring something that has enough intensity to be valid.

It is very much a signal to noise issue, the amount of processing you can do in software is something that is an ongoing area of research and I would bet money that with better algorithms we could drag more range out of the sensors, than thought possible. But it would have to be done down on the sensor level, not with the XYZ results.

Than again a profiled colorimeter would probably still blow it out of the water.
LL

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #28 of 29 Old 03-04-2012, 11:37 AM
AVS Special Member
 
Chad B's Avatar
 
Join Date: Oct 2002
Location: Piqua, OH
Posts: 2,087
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 80 Post(s)
Liked: 416
I have owned/used a few different i1Pro meters, and without a doubt the color that is most inconsistent with it is red. With bright measurements there's no problem, but on the lower quarter or third of a grayscale run the problems start to manifest themselves. Red changed significantly with how long ago the dark reading was taken and the quality of the dark reading.
Thankfully the Jeti does not have that problem.

ISF/THX calibrator, HDTVbyChadB
Jeti 1211

Latest reviews:
Sony 65X950BLG 65UB9800Samsung UN65HU8550UN65HU9000, UN85HU8550, LG 55EC9300
Chad B is online now  
post #29 of 29 Old 03-06-2012, 10:27 PM - Thread Starter
Newbie
 
tech_df's Avatar
 
Join Date: Feb 2012
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
I took measurements from the B420 with a PR-670 to get a better idea of the black level. What I learned is that the 655 was giving wild variance because the readings were actually below the specs of the device (not slightly above the base spec as the device was reporting). The actual luminance level was substantially lower. The 670 gave its share of failed readings as well, but the greater sensitivity did prove useful. However, neither device can provide accurate data in the shadows, so I'll be looking for another solution.

For those of you using a colorimeter or spectro-calibrated colorimeter to pull color information from the shadows, how many readings do you have to average to feel confident in the data?

Mike Lary
Digital Factory
Seoul, Korea
http://digitalfactory.kr/
tech_df is offline  
Reply Display Calibration

User Tag List

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