Official JVC RS20 / HD750 Calibration and CMS thread - Page 28 - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #811 of 1782 Old 01-08-2009, 02:03 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Quote:
Originally Posted by KTTV Images View Post

Googer,

Great work.
I think you have indeed found the Non-Linearity in the system that has been, more or less, escaping all of us. It is not the Gamma but the idea that our operating points for the RGB signals themselves are being mis-set and are not operating in a linear range of input vs output response.

The RGB operating points (separately or in combination) are being pushed and shoved to some point too close to saturation (-or off) by the various experimental adjustments being made in the CMS and with the normal Color Brightness and Contrast controls.

If this is the case it should be relatively easy to test for saturation effects on the RGB channels visually if you have access to a HD-DVD player by running the following test. What I am proposing here is just perhaps a new test to assist in judging what is happening when the CMS and basic settings are manipulated in attempt to tame the RS20 CMS Lion.

Note that this test does not involve (or replace ) the color measuring equipment necessary for CMS alignment. It involves only a HD test disk and HD player to provide test signals for convieniently judging the response of the projector to a full range of RGB input signals looking for signs of saturation in one or more of the these channels.

1. Use the HD-DVD DVE test disk.
Joe Kane has apparently not released a full Blue Ray version of his Digital Video Essentials disk -- only a DVE Basics disk aimed more at the consumer. Mine is in the mail but I doubt it has the needed test signals on it for this test. The HD DVD version has on it full R, B, G, Cy, Ma and Y Ramps and Steps .

2. Set up the projector for the basic Contrast level you have decided on using and then set the brightness for the correct black level using the pluge , and lastly set the Color and Hue controls using the color filters (or if it makes more sense -to the values you believe are the ones you want to use with any new setup for the CMS).

3. Now play the test signals (found under 1920x1080 - Tests and Calibration - Video Calibration) and look at the "Ramps and Steps" for each of the colors R, B, G, Ma,Cy, and Y to find out if you are running into saturation at the high end or if you are bottoming out at the low end. The normal operating point for each of these colors should be such that none of them reach saturation or bottom out and that all the steps are fully reproduced. the Steps and Ramps cover the full input range for RGB.

By using this test for each setup of the CMS settings you experiment with I think you should be able to quickly find out what colors are going into saturation or bottoming out. If this happens it will be because the operating point is not set correctly for each of the colors and they are responding in a non-linear way to the full range of color inputs to the projector, and you should be able to quickly identify the color and the degree of the problem for each color.

This should be done before looking at real images since it will give you an idea what problem you are looking for in the images (vis.. the red colors may be saturating as they go above 70%..etc.).

Kt

Thanks Googer and KT, this is really useful.
Manni01 is online now  
Sponsored Links
Advertisement
 
post #812 of 1782 Old 01-08-2009, 02:25 AM
AVS Special Member
 
Highjinx's Avatar
 
Join Date: Dec 2002
Location: Australia
Posts: 1,769
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 23
Quote:
Originally Posted by KTTV Images View Post

Googer,



1. Use the HD-DVD DVE test disk.
Joe Kane has apparently not released a full Blue Ray version of his Digital Video Essentials disk -- only a DVE Basics disk aimed more at the consumer. Mine is in the mail but I doubt it has the needed test signals on it for this test. The HD DVD version has on it full R, B, G, Cy, Ma and Y Ramps and Steps .


Kt

I guess one could use Spectracals Calman HTPC Pattern Generator for this step?

May the success of a Nation be judged not by its collective wealth nor by its power, but by the contentment of its people.
Hiran J Wijeyesekera - 1985.
Highjinx is offline  
post #813 of 1782 Old 01-08-2009, 05:06 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Could someone please confirm if the 75% color patterns on the AVS rec709 dvd are 75% saturation or 75% stimulus? I'm not talking about the primary/secondary sat, but the fifth option in the menu (RC1).
Manni01 is online now  
post #814 of 1782 Old 01-08-2009, 07:22 AM
AVS Special Member
 
lovingdvd's Avatar
 
Join Date: Jul 2002
Posts: 8,818
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Quote:
Originally Posted by Manni01 View Post

Could someone please confirm if the 75% color patterns on the AVS rec709 dvd are 75% saturation or 75% stimulus? I'm not talking about the primary/secondary sat, but the fifth option in the menu (RC1).

Probably your best bet is to ask the author in the AVS HD 709 thread if you haven't already.
lovingdvd is offline  
post #815 of 1782 Old 01-08-2009, 07:34 AM
AVS Special Member
 
lovingdvd's Avatar
 
Join Date: Jul 2002
Posts: 8,818
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Quote:
Originally Posted by KTTV Images View Post

If this is the case it should be relatively easy to test for saturation effects on the RGB channels visually if you have access to a HD-DVD player by running the following test. What I am proposing here is just perhaps a new test to assist in judging what is happening when the CMS and basic settings are manipulated in attempt to tame the RS20 CMS Lion.
Kt

Thanks KTTV for the great post. I was thinking that it would be nice to have the 5% scales patterns for the secondaries since I already had them for the primaries (and the pj will output these as well in its test patterns for the primaries).

Thanks to a large concerted effort of everyone here I think at this point we have a very good understanding of the issue. The challenge is that so far there appears to be no way to get around it, short of a CMS fix by JVC.

There are several things that can and have been tried, and all of these eventually lead to a significant "gotcha" that really hampers the image.

So far for those that find THX mode undesirable (for whatever reason and there are a good number of them, despite it still being a decent overall option) it seems the best it seems the best we can do is find a good balance. I'll post shortly on my results from last night.
lovingdvd is offline  
post #816 of 1782 Old 01-08-2009, 09:35 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
As a way to get back into shape before going for a 5th attempt, I have done a first series of contrast measurements (with a Tecpel 531 lightmeter) which confirm Mark's findings regarding the way RGB offsets affect contrast...

I've also measured on/off in "untweaked" THX preset (to make sure measurements where not affected by my settings) in normal/high mode with different iris positions (best on/off seems to come predictably with normal lamp, iris closed).

Full data here: http://www.avforums.com/forums/dlp-l...ml#post8540051

Comments and analysis welcome
Manni01 is online now  
post #817 of 1782 Old 01-08-2009, 10:06 AM
Advanced Member
 
Googer's Avatar
 
Join Date: Jan 2002
Location: Pittsburgh, PA, USA
Posts: 674
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 15
Quote:
Originally Posted by Highjinx View Post

I guess one could use Spectracals Calman HTPC Pattern Generator for this step?

Yep, Grayscale Calibration tab, select the color you want to check, # of segments (21 is good since it's enough to easily see any differences at the high levels without being so many that differences get smoothed-out in visual judgment), and don't forget to select Video Levels.
Googer is offline  
post #818 of 1782 Old 01-08-2009, 10:15 AM - 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 Manni01 View Post

As a way to get back into shape before going for a 5th attempt, I have done a first series of contrast measurements (with a Tecpel 531 lightmeter) which confirm Mark's findings regarding the way RGB offsets affect contrast...

I've also measured on/off in "untweaked" THX preset (to make sure measurements where not affected by my settings) in normal/high mode with different iris positions (best on/off seems to come predictably with normal lamp, iris closed).

Full data here: http://www.avforums.com/forums/dlp-l...ml#post8540051

Comments and analysis welcome

Nice work Manni! Interesting that the normal bulb setting provides the best contrast. Did the C4Home data also show this? For most folks though, the smallest iris aperture and normal bulb combination won't provide sufficient lumens. So what they'll want to do is use the high bulb setting (assuming they can tolerate the increase in noise) and then use the smallest iris aperture setting to achieve the desired ftL. In other words, the high bulb with smaller iris aperture combination should always yield better contrast than normal bulb with a larger iris aperture assuming that both are achieving the same ftL.

Oh yeah and what a difference it made in backing off the offset: 12k:1 vs almost 30k:1
Mark Petersen is offline  
post #819 of 1782 Old 01-08-2009, 10:20 AM
AVS Special Member
 
kutlow's Avatar
 
Join Date: Sep 2006
Location: Wetumpka, Al
Posts: 2,356
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Quote:
Originally Posted by Mark Petersen View Post

Nice work Manni! Interesting that the normal bulb setting provides the best contrast. Did the C4Home data also show this? For most folks though, the smallest iris aperture and normal bulb combination won't provide sufficient lumens. So what they'll want to do is use the high bulb setting (assuming they can tolerate the increase in noise) and then use the smallest iris aperture setting to achieve the desired ftL. In other words, the high bulb with smaller iris aperture combination should always yield better contrast than normal bulb with a larger iris aperture assuming that both are achieving the same ftL.

Oh yeah and what a difference it made in backing off the offset: 12k:1 vs almost 30k:1

I am not understanding this. I thought the cinehome measurements said high lamp mode yields more contrast. On a 55 x 132 screen shouldnt I use high lamp?
kutlow is offline  
post #820 of 1782 Old 01-08-2009, 10:22 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Quote:
Originally Posted by Mark Petersen View Post

Oh yeah and what a difference it made in backing off the offset: 12k:1 vs almost 30k:1

Well, thank you for allowing me to double my contrast.

I haven't had time to watch a film with the new settings (I'd like to run a quick greyscale optimisation before) but I expect the difference to be quite dramatic. I hope it will also have a positve effect on the "ansi wash-out" I felt in mixed/high apl scenes.

I'll report later!
Manni01 is online now  
post #821 of 1782 Old 01-08-2009, 10:26 AM
Advanced Member
 
Googer's Avatar
 
Join Date: Jan 2002
Location: Pittsburgh, PA, USA
Posts: 674
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 15
BTW, I believe I've at this point deduced exactly how JVC's CMS works. Rather than working on a perceptual level, I'm pretty sure it's working strictly on a technical level of HSL. The sliders behave 100% exactly like they do in MS Paint for example if you go to define a custom color. When high levels of a color are washing out because we turn the brightness control for it too high, that corresponds exactly with using increasing luminance above 120 in the Windows custom color creation dialog. This is also why decreasing saturation is decreasing measured light output. From a technical point-of-view of HSL, it actually isn't but perceptual point-of-view it is due to the differences in correct light output for different colors. We of course don't care about what's going on technically and are trying to control what our eyes see and this difference is what's causing the CMS to behave differently than how you intuitively expect it to.
Googer is offline  
post #822 of 1782 Old 01-08-2009, 10:26 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Quote:
Originally Posted by kutlow View Post

I am not understanding this. I thought the cinehome measurements said high lamp mode yields more contrast. On a 55 x 132 screen shouldnt I use high lamp?

There may be some difference due to my environment (throw ratio, zoom, etc).

Plus these are my first contrast readings, so I may be wrong in the procedure and/or the calculations (which is why I provided the raw data).

As Mark said, you don't have much of a choice, on a 132" screen you probably need high lamp anyway to get enough lumens, unless you use a very high gain screen. Also the difference is only 10% (at the same iris setting), so you're not losing that much (provided my readings/calculations are correct).

EDIT: I had a quick look at Cine4Home measurements, the readings between high/norml mode are almost identical at the same zoom setting, it's the zoom (min or max) that makes most of the difference.
Manni01 is online now  
post #823 of 1782 Old 01-08-2009, 10:53 AM
AVS Special Member
 
lovingdvd's Avatar
 
Join Date: Jul 2002
Posts: 8,818
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Quote:
Originally Posted by Manni01 View Post

There may be some difference due to my environment (throw ratio, zoom, etc).

Plus these are my first contrast readings, so I may be wrong in the procedure and/or the calculations (which is why I provided the raw data).

As Mark said, you don't have much of a choice, on a 132" screen you probably need high lamp anyway to get enough lumens, unless you use a very high gain screen. Also the difference is only 10% (at the same iris setting), so you're not losing that much (provided my readings/calculations are correct).

EDIT: I had a quick look at Cine4Home measurements, the readings between high/norml mode are almost identical at the same zoom setting, it's the zoom (min or max) that makes most of the difference.

Manni you may want to remeasure to make sure and also makes sure your meter is working properly. From everything that's been posted and my own experience with multiple RS20's yours would be the first case where CR was not higher in high lamp mode with same iris setting at same throw.
lovingdvd is offline  
post #824 of 1782 Old 01-08-2009, 10:53 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Quote:
Originally Posted by lovingdvd View Post

Probably your best bet is to ask the author in the AVS HD 709 thread if you haven't already.

Thanks LovingDVD, I've just followed your advice.

Thanks also for yor great contrast tutorial, it was really helpful to make my first measurements.
Manni01 is online now  
post #825 of 1782 Old 01-08-2009, 10:57 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Quote:
Originally Posted by lovingdvd View Post

Manni you may want to remeasure to make sure and also makes sure your meter is working properly. From everything that's been posted and my own experience with multiple RS20's yours would be the first case where CR was not higher in high lamp mode with same iris setting at same throw.

I've done a few runs (with my own settings and in THX), and they were all consistent. I how no idea how to test the accuracy of the lightmeter... There may be something wrong with my PJ?
Do the values look allright (when not comparing high/normal lamps)? Are my calculations right (not that it's very complicated...). Could you please check the raw data in the spreadsheet?

Also how do you explain the fact that Cine4home measures are identical in normal and high lamp at the same zoom?
Manni01 is online now  
post #826 of 1782 Old 01-08-2009, 11:02 AM
AVS Special Member
 
lovingdvd's Avatar
 
Join Date: Jul 2002
Posts: 8,818
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Quote:
Originally Posted by Manni01 View Post

I've done a few runs (with my own settings and in THX), and they were all consistent. I how no idea how to test the accuracy of the lightmeter... There may be something wrong with my PJ?
Do the values look allright (when not comparing high/normal lamps)? Are my calculations right (not that it's very complicated...). Could you please check the raw data in the spreadsheet?

Also how do you explain the fact that Cine4home measures are identical in normal and high lamp at the same zoom?

Look at Cine4home's CR numbers. High lamp beats normal lamp every time in CR given the same throw and iris position. Something is not adding up.

What lumens do you calculate when measuring at the screen in both scenarios?
lovingdvd is offline  
post #827 of 1782 Old 01-08-2009, 11:07 AM
AVS Special Member
 
Frank Derks's Avatar
 
Join Date: Jul 2002
Location: Region A,B,C
Posts: 1,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 87 Post(s)
Liked: 84
Quote:
Originally Posted by Mark Petersen View Post

Boosting the RGB offsets above 0 is what kills on/off because it raises the black level significantly. Reducing lumen output a little will affect contrast much less so. How much less has to be determined.



I think they are independent from each other but lowering contrast does end up effectively doing the same thing as lowering lumens. This is what I found:

Setting the black level floor (via the RGB offset adjustments) and the peak white (via the RGB gains) determines the full dynamic range (and therefore contrast). The brightness and contrast adjustments just set where video level 16 and 235 fall within that range. So lets say the black level is elevated (via RGB offsets), lowering the brightness adjustment will have no affect on the black level and lowering it will simply crush blacks. Lowering the contrast adjustment will lower where 235 hits as far as lumen dynamic range so it affectively does the same thing to full whites as reducing the RGB gains, but not by the same mechanism.

If reduced lumen output does work out to be the way to go to get the CMS to track, it's worth playing around with both the contrast and RGB gains to see which may yield the best results.

It's a given that at the end of the chain there are three panels that have to do with 8 bit dynamic range (at best) to modulate the light output.
Within this available range the grascale and cms adjustments must operate without clipping the maximum inputted video level.

The white balance adjustment (grayscale) to get D6500 limits the dynamic range for blue and green due to weaker red from UHP lamp.

The $6000 question is are the CMS matrix calculations applied before or after the gray scale calculation in the projectors hardware/software implementation.


If CMS is applied after the white balance adjustments then the adjustment ranges in the cms for green and blue get reduced significantly.
(I guess roughly buy the same ratio of the lamps red to green light output ratio! 20..30%)

Example: grayscale may require green gain to be reduced by 30%
Gray scale is a relative adjustment while to move the green around in the gamut requires an absolute adjustment.
The cms adjustment range for green suffers because the green stimilus it has to work with is reduced by 30%
For blue it is the same but since blue is closer to the desired location in the gamut it is less of a problem.

This would be a major design flaw.

I think if this is the case the best way to proceed is to white balance the image with a red color filter to get the most dynamic range for green or blue (depending on which color is strongest). I suspect blue is a bit weaker than green so the trick is to balance blue and red with the color filter and get green in line with the gain control.

This should minimize the impact on lumens output and improve the black level.
(Minimizing the hit on the contrast ratio)

With the overal whiteness setting (contrast) it must be possible to adjust for the maximum light output before the panels max out.


Its widely accepted that accurate gray scale is are the expense of max lumens. (For UHP lamps with weaker reds that is.)

It seems that in a similar fashion accurate colors will be at the expense of the contrast ratio.
Frank Derks is offline  
post #828 of 1782 Old 01-08-2009, 11:11 AM
AVS Special Member
 
lovingdvd's Avatar
 
Join Date: Jul 2002
Posts: 8,818
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Anyone measure ANSI with high lamp and iris -15? I'm only getting 202:1 using two different meters. I'm using quite a bit of vertical shift (the amount needed when the pj is mounted just a bit higher than the top of the screen) and a relatively small amount of horizontal shift (the amount needed to compensate for the off-center lens over a 15' throw).

ANSI goes up to about 230:1 if measured at iris 0. Likewise I didn't have a chance to measure it but I'm pretty sure it will go up about another 10% if I switch to normal lamp.

Also this is not Greg's modified ANSI CR where a sample is taken from multiple squares and averaged. This is just from a single reading from the center. I may try the modified version as I bet it will measure a bit better off center.

On/off is measuring nicely at 45,000:1. This is at mid throw.

I haven't played with it yet, but I may experiment a bit with using normal lamp and iris at x that gives me the same lumens as I have now. On/off and black level will take a hit, but I'll be curious to see if the ANSI somehow makes any noticeable difference.
lovingdvd is offline  
post #829 of 1782 Old 01-08-2009, 11:17 AM
umr
AVS Club Gold
 
umr's Avatar
 
Join Date: Oct 2002
Location: USA
Posts: 10,134
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 29
Quote:
Originally Posted by Frank Derks View Post

It's a given that at the end of the chain there are three panels that have to do with 8 bit dynamic range (at best) to modulate the light output.....

Many panel driver circuits have more than 8 bit depth. I believe JVC uses a 10 bit design. Sony uses 12 bit in their SXRD designs.
umr is offline  
post #830 of 1782 Old 01-08-2009, 11:21 AM
AVS Special Member
 
lovingdvd's Avatar
 
Join Date: Jul 2002
Posts: 8,818
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Quote:
Originally Posted by Frank Derks View Post

It's a given that at the end of the chain there are three panels that have to do with 8 bit dynamic range (at best) to modulate the light output.
Within this available range the grascale and cms adjustments must operate without clipping the maximum inputted video level.

The white balance adjustment (grayscale) to get D6500 limits the dynamic range for blue and green due to weaker red from UHP lamp.

The $6000 question is are the CMS matrix calculations applied before or after the gray scale calculation in the projectors hardware/software implementation.


If CMS is applied after the white balance adjustments then the adjustment ranges in the cms for green and blue get reduced significantly.
(I guess roughly buy the same ratio of the lamps red to green light output ratio! 20..30%)

Example: grayscale may require green gain to be reduced by 30%
Gray scale is a relative adjustment while to move the green around in the gamut requires an absolute adjustment.
The cms adjustment range for green suffers because the green stimilus it has to work with is reduced by 30%
For blue it is the same but since blue is closer to the desired location in the gamut it is less of a problem.

This would be a major design flaw.

I think if this is the case the best way to proceed is to white balance the image with a red color filter to get the most dynamic range for green or blue (depending on which color is strongest). I suspect blue is a bit weaker than green so the trick is to balance blue and red with the color filter and get green in line with the gain control.

This should minimize the impact on lumens output and improve the black level.
(Minimizing the hit on the contrast ratio)

With the overal whiteness setting (contrast) it must be possible to adjust for the maximum light output before the panels max out.


Its widely accepted that accurate gray scale is are the expense of max lumens. (For UHP lamps with weaker reds that is.)

It seems that in a similar fashion accurate colors will be at the expense of the contrast ratio.

Thanks for the excellent write up Frank! I never did quite grasp the whole concept behind using the filters, so hopefully folks familiar with this will experiment and report back.

Your theory makes total sense. With this in mind do you think it is somehow possible to adjust RGB gain and offset at different bright level using a custom gamma that can somehow help us work around this issue? Mark seemed to be thinking down this path earlier.

It sounds like there is some potential here to at least improve things using the gamma curves. But again, hopefully not at the expense of light output and reduced CR.
lovingdvd is offline  
post #831 of 1782 Old 01-08-2009, 11:22 AM - 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 lovingdvd View Post

Also this is not Greg's modified ANSI CR where a sample is taken from multiple squares and averaged. This is just from a single reading from the center. I may try the modified version as I bet it will measure a bit better off center.

IIRC, I think multiple readings from different squares and then averaged is supposed to be the way ANSI contrast is to be performed. I don't know how gregr measures it but the single reading in the center using two reversed patterns (one with a white center rectangle and another with black center retangle) is how most people seem to be doing it nowadays. This later approach is what I would call a modified measurement. Fwiw, my modified ANSI readings were relatively low too but throw plays a role and I haven't yet measured it in the short throw configuration (which probably gives the best ANSI).
Mark Petersen is offline  
post #832 of 1782 Old 01-08-2009, 11:30 AM
AVS Special Member
 
jigesh's Avatar
 
Join Date: Apr 2002
Location: Hopewell Junction, NY
Posts: 1,080
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by umr View Post

..I believe JVC uses a 10 bit design....

Right. As per the 20th slide of this presentation, their D-ILA device can allow up to 24-bit (= 2^24) gray scale steps, but they use 10 bits (1024 states).
jigesh is offline  
post #833 of 1782 Old 01-08-2009, 11:34 AM
AVS Special Member
 
lovingdvd's Avatar
 
Join Date: Jul 2002
Posts: 8,818
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Quote:
Originally Posted by Mark Petersen View Post

IIRC, I think multiple readings from different squares and then averaged is supposed to be the way ANSI contrast is to be performed. I don't know how gregr measures it but the single reading in the center using two reversed patterns (one with a white center rectangle and another with black center retangle) is how most people seem to be doing it nowadays. This later approach is what I would call a modified measurement. Fwiw, my modified ANSI readings were relatively low too but throw plays a role and I haven't yet measured it in the short throw configuration (which probably gives the best ANSI).

Yes now that you mention it I do recall something about ANSI CR being formed by averaging squares. And I can't recall what Greg's tweak is in the process that he uses and calls his Modified ANSI CR. Like you said I think a lot of folks these days use the term ANSI CR to refer to a measure off the checkboard and to differentiate from on/off, without necessarily following any true ANSI measuring standard. Which is generally OK I think we just need to know what folks are talking about when they post a number.

Do you recall what / if you measured for your ANSI CR with high lamp iris -15? Could it have been down around the 200:1 level?
lovingdvd is offline  
post #834 of 1782 Old 01-08-2009, 11:35 AM
AVS Special Member
 
Frank Derks's Avatar
 
Join Date: Jul 2002
Location: Region A,B,C
Posts: 1,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 87 Post(s)
Liked: 84
Quote:
Originally Posted by lovingdvd View Post

Thanks for the excellent write up Frank! I never did quite grasp the whole concept behind using the filters, so hopefully folks familiar with this will experiment and report back.

Your theory makes total sense. With this in mind do you think it is somehow possible to adjust RGB gain and offset at different bright level using a custom gamma that can somehow help us work around this issue? Mark seemed to be thinking down this path earlier.

It sounds like there is some potential here to at least improve things using the gamma curves. But again, hopefully not at the expense of light output and reduced CR.

Gamma adjustments have to operate in the same dynamic range available for driving the panels. It can't make up for the lost blue and green dynamics range due to D6500 white adjustment.
Frank Derks is offline  
post #835 of 1782 Old 01-08-2009, 11:40 AM - 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 Frank Derks View Post

I think if this is the case the best way to proceed is to white balance the image with a red color filter to get the most dynamic range for green or blue (depending on which color is strongest). I suspect blue is a bit weaker than green so the trick is to balance blue and red with the color filter and get green in line with the gain control.

Actually after red, green is the next weakest and then blue has quite a bit of excess, so that blue has to be cut the most. So a person would want to color balance the CCF with a mix of green and red I would think.

Quote:


This should minimize the impact on lumens output and improve the black level.
(Minimizing the hit on the contrast ratio)

With the overal whiteness setting (contrast) it must be possible to adjust for the maximum light output before the panels max out.

Yup exactly.

Quote:


Its widely accepted that accurate gray scale is are the expense of max lumens. (For UHP lamps with weaker reds that is.)

It seems that in a similar fashion accurate colors will be at the expense of the contrast ratio.

Yes true too.
Mark Petersen is offline  
post #836 of 1782 Old 01-08-2009, 11:43 AM
AVS Special Member
 
Frank Derks's Avatar
 
Join Date: Jul 2002
Location: Region A,B,C
Posts: 1,884
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 87 Post(s)
Liked: 84
Quote:
Originally Posted by umr View Post

Many panel driver circuits have more than 8 bit depth. I believe JVC uses a 10 bit design. Sony uses 12 bit in their SXRD designs.

Good call.
However if controls max out the number of bits are not important.
Frank Derks is offline  
post #837 of 1782 Old 01-08-2009, 11:43 AM - 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 lovingdvd View Post

Yes now that you mention it I do recall something about ANSI CR being formed by averaging squares. And I can't recall what Greg's tweak is in the process that he uses and calls his Modified ANSI CR. Like you said I think a lot of folks these days use the term ANSI CR to refer to a measure off the checkboard and to differentiate from on/off, without necessarily following any true ANSI measuring standard. Which is generally OK I think we just need to know what folks are talking about when they post a number.

Do you recall what / if you measured for your ANSI CR with high lamp iris -15? Could it have been down around the 200:1 level?

I measured under 200:1 (190ish) at iris -15 and long throw and something like 230ish:1 with iris open. Don't quote me on the exact numbers but it was in that ballpark.
Mark Petersen is offline  
post #838 of 1782 Old 01-08-2009, 11:45 AM
umr
AVS Club Gold
 
umr's Avatar
 
Join Date: Oct 2002
Location: USA
Posts: 10,134
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 29
Quote:
Originally Posted by Frank Derks View Post

Good call.
However if controls max out the number of bits are not important.

Of course, but panel limits and the calibration goals tend to bound the solution more than the controls.
umr is offline  
post #839 of 1782 Old 01-08-2009, 11:47 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Quote:
Originally Posted by lovingdvd View Post

Yes now that you mention it I do recall something about ANSI CR being formed by averaging squares. And I can't recall what Greg's tweak is in the process that he uses and calls his Modified ANSI CR. Like you said I think a lot of folks these days use the term ANSI CR to refer to a measure off the checkboard and to differentiate from on/off, without necessarily following any true ANSI measuring standard. Which is generally OK I think we just need to know what folks are talking about when they post a number.

Do you recall what / if you measured for your ANSI CR with high lamp iris -15? Could it have been down around the 200:1 level?

AFAIK Gregr uses an average of the four squares in the center of the standard ansi checkboard pattern (4 black, four white).
Manni01 is online now  
post #840 of 1782 Old 01-08-2009, 11:55 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,220
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 48 Post(s)
Liked: 93
Quote:
Originally Posted by lovingdvd View Post

Look at Cine4home's CR numbers. High lamp beats normal lamp every time in CR given the same throw and iris position. Something is not adding up.

What lumens do you calculate when measuring at the screen in both scenarios?


I didn't have time to measure anything at the screen, as I was doing a quick test with On/off.

Re Cine4home, I wasn't looking at the preprod2 (where the on/off is very slightly higher in high lamp mode, like less than 2%) but at the preprod1 (when they are exactly identical).

I'll run new measurements to double check, but I am very at this stage.
Manni01 is online now  
Reply Digital Hi-End Projectors - $3,000+ USD MSRP

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