CurtPalme.com Home Theater Forum

[kal]

[kal]

[kal]

Here's another question: When you take xyY measurements for greyscale you typically use a small window in the middle of the screen of varying shades of grey. When people refer to their peak like output at 100IRE, are the referring to the small window ftL reading or the whole screen displaying a 100IRE white?

Mine vary widely. I get 14+ ftL on the small window at 100IRE white with contrast set to 65 while a full screen 100IRE white only gets me 6 ftL!

I know this has to do how CRT's work. I'm just surprised at the difference when I actually measure it. I can easily push the small 100IRE window into blooming when I crank the contrast above 90, but can't even get close to blooming if I have a whole white 100IRE screen.
Kal

[kal]

Here's another question: When you take xyY measurements for greyscale you typically use a small window in the middle of the screen of varying shades of grey. When people refer to their peak like output at 100IRE, are the referring to the small window ftL reading or the whole screen displaying a 100IRE white?

Mine vary widely. I get 14+ ftL on the small window at 100IRE white with contrast set to 65 while a full screen 100IRE white only gets me 6 ftL!

I know this has to do how CRT's work. I'm just surprised at the difference when I actually measure it. I can easily push the small 100IRE window into blooming when I crank the contrast above 90, but can't even get close to blooming if I have a whole white 100IRE screen.
Kal

[zGman]

Hi Kal,
Thanks for the link to the new test patterns - I will try that soon.
Yes there's a big difference between the windowed pattern and full
screen, and in some of my testing, I found a different greyscale
profile and luminance curve between the two methods. Which is
the right way...? I am not so sure. Its pretty rare any content is full
grey screens anyhow. And one of the major strongpoints of CRT
projection (in my opinion) is the ability to really punch out a bright
object (car headlights, sunny reflections, etc) in an image.
Cheers,
Galen

[r.bauer]

Maximum lightoutput is related to projected surface area. Peak lumens are always measured at 10% windowed pattern.
Use this pattern when calibrating your projector. This type of Average Picture Level (APL) is very similar to the APL of video or film material.

It all comes down to beam current. There is just a maximum that can be delivered during a certain period of time. This originates in design of the electron gun (all these electrons emiting just from the tip of the Cathode), the RGB end-stage (That 100V peak-peak video signal has to be delivered with a certain amount of miliAmps at 120MHz or more. This is very difficult, many people forget what kind of a very special circuitry is needed to achieve these specs!) and the High voltage power supply which has to deliver enough miliAmps at 34,7kVolt.

One last thing: Rember that when calibrating a CRT projector you are just compensating for CRT-aging. That's why some knowledge of CRT-aging comes in handy.

Natural Gamma curve of each CRT.
the green CRT behaviour is closest to the natural Gamma curve of 2.2. Red is more dificult and proper correction in the driving circuits is required to achieve 2.2 gamma. Blue is worst of all, much more correction is required to achieve the 2.2 Gamma curve.

Some interesting reading about Blue roll-off at high drive level:
(This link also partially desrcibes the Gamma slope and breakpoint function)
http://www.patentstorm.us/patents/5644360-description.html

http://www.torrscientific.co.uk/phosphors.htm

http://www.engineering.uiowa.edu/~aip/Misc/ColorFAQ.html#RTFToC17

[kal]

[kal]

[r.bauer]

[kal]

[kal]

[kal]

So I've spent a few days working on my greyscale and I can get everything nice and flat except for the low end of the blue tube. Not a big deal (looks fantastic) but I'm a stickler for perfection so I'm trying to figure this one out.

I've tried so many different things that I actually think there's something slightly wrong with how the blue tube reacts.

Take a look at this:



No matter what I do, the blue drops right off in the 0-30 IRE range.
I've adjusted my G2s manually so that when 0 IRE is displayed, all 3 tubes tubes "light up at the same time". But the problem is that the light output out of the other two tubes just increases so much faster as I turn up the brightness (or display increasing patterns from 0-30 IRE) that the result turns out the same no matter how I adjust the G2's. Blue always drops off like crazy in the low end.

If I change the Blue cut-off to get more blue in the lower IRE's (ie: calibrate for the 10 IRE or 20 IRE point) it creates an enormous blue hump in the middle of the graph that makes the rest of the image too blue. My blue midlights setting is already at the factory default of zero - I can't lower it any further to reduce blue output in the mid-section.

I'd like to raise the blue in the 0-30 range but I've litterally spent 10-15 hours on this and it just doesn't seem possible.

I'm starting to think the only way is with an scaler that gives you advanced controls for stuff like this. Would be nice to have one for 72Hz output too. I'm seriously considering one.

Any hints?

Kal

[kal]

So I've spent a few days working on my greyscale and I can get everything nice and flat except for the low end of the blue tube. Not a big deal (looks fantastic) but I'm a stickler for perfection so I'm trying to figure this one out.

I've tried so many different things that I actually think there's something slightly wrong with how the blue tube reacts.

Take a look at this:



No matter what I do, the blue drops right off in the 0-30 IRE range.
I've adjusted my G2s manually so that when 0 IRE is displayed, all 3 tubes tubes "light up at the same time". But the problem is that the light output out of the other two tubes just increases so much faster as I turn up the brightness (or display increasing patterns from 0-30 IRE) that the result turns out the same no matter how I adjust the G2's. Blue always drops off like crazy in the low end.

If I change the Blue cut-off to get more blue in the lower IRE's (ie: calibrate for the 10 IRE or 20 IRE point) it creates an enormous blue hump in the middle of the graph that makes the rest of the image too blue. My blue midlights setting is already at the factory default of zero - I can't lower it any further to reduce blue output in the mid-section.

I'd like to raise the blue in the 0-30 range but I've litterally spent 10-15 hours on this and it just doesn't seem possible.

I'm starting to think the only way is with an scaler that gives you advanced controls for stuff like this. Would be nice to have one for 72Hz output too. I'm seriously considering one.

Any hints?

Kal

[perisoft]

[kal]

[kal]

[Chuchuf]

Kal,
I assume that you are taking the readings off the screen w/ your sensor (sorry I didn't read all of this thread). If so that sensor is not that sensitive when taking reflective readings. Even pointed at the PJ, below 19ire is tough.
I believe that the readings you are getting below 30ire are the fault of the sensor?? Especially at 10ire.
I generally look at trends as I approach 20-30ire to determine where the color is going. Yours are fairly flat which is good.

Terry

[WTS]

Hi Kal,

I was just wondering if you have the gamma boost circuit in the path when you're doing the greyscale. Of course you need to include it during cal but have you tried it without?

[Person99]

[Person99]

[kal]

[kal]

[kal]

[kal]

[kal]

[quote="Person99"]

[kal]

[quote="Person99"]

[Person99]

[kal]

[kal]

[MikeEby]

Forgive my ignorance to these kinds of thing, but when you take your readings does the sensor repeat when going from 0 to 100 IRE then back from 100 to 0 IRE? In other words do you see hysteresis in your readings?

Mike