Grayscale variances - different patterns, plasma calibration - Page 4

Does VT50 model has both 2-points and 10-point grayscale adjustment? You can try to use them both to fix gamma.

No you can't.

The two point is just RGB balance.

The 10 point part is RGB + Gamma.

I meant to use them both, 2 points to balance white then 10 points to flatten Gamma and balance RGB. That was what I did with my old LG, by doing this way you don't to have to make large 10-points adjustment and it easier to fix gamma than just using 10-points alone.

Thought on the VT50 that you had to use the gamma settings to effect gamma. Two and ten point adjustments readjusted itself to maintain the same Y. Is that not the case?

Ah yes, that's very correct.

The two point are the course controls and the 10 points the fine control.

You'll absolutely have an easier time of it adjusting the two point controls first. and the moving to the 10 point control.

That is correct, but in CalMAN we do show the gamma control next to the RGB controls, so it's easy to think of them as being part of the same 10 point control.

That [plasmas are 5 or 6 bits at best] would imply no more than 32 to 64 shades of red, green and blue can be displayed on a plasma. I see many more than that. In fact, the 4-year-old Samsung plasma I'm using shows 219 steps of each color (16-235) and more than that when you display a pattern with 0-255 but some of those steps blend together (as much a limitation of human vision as a limitation of the display, I wasn't interested in taking ~256 readings with a meter to see if there were 256 (or 255) discrete "levels" being displayed (the meter would see them even if I didn't).

The fact that plasma pixels can only be on or off and that they have to flicker on and off multiple times during a frame to produce intermediate values of red, green, and blue MIGHT be interpreted as lower resolution, especially at very close viewing distances. But at normal viewing angles (more precise than distance since 30-40 degree angles would be panel-size independent), you don't see the individually flickering pixels.

Plasmas use dithering of the values to overcome the bit-depth limitation. That works for many images, but not smooth ramps.

Steve

If I understand this post correctly, it is counter productive to calibrate the "Gamma Detail Adjustment" (IRE) on Panasonic Plasma display's to coincide with the selected gamma setting in the user menu.

Example:
You begin a calibration session with the gamma set to 2.2 in the user menu, your software shows that you have an average gamma reading around 2.1 with 10% and 20% being lower and of course 80% and 90% is lower. These numbers are not accurate because of the size of the windows used due to ABL and your "true" gamma is actually 2.2.

Therefore, adjusting your "IRE" in the "Gamma Detail Adjustment" menu so that the software is showing a flat 2.2, is incorrect and what you have actually done is set your gamma to a higher number than 2.2.

If this is correct, then assuming that I understand Michael's post, the correct procedure would be to leave the IRE settings alone with respect to gamma?

No that's not exactly right...

Average gamma is a BAD indicator of what the displays REAL gamma is doing. Because each point in the grayscale (except 100% and 0%) can have a gamma calculated for them. You don't want a gamma of 2.5 at 90% that falls to 1.8 at 10%... it might AVERAGE 2.2 gamma, but images are going to look like crap. You want the gamma at every grayscale step to be your target (I prefer 2.25 Gamma as the target for most "modern" displays since mastering is done on displays set to 2.2 or 2.3 gamma, no matter WHAT anybody else says, that's what is actually being used during editing/grading and evaluation). What you want is a gamma of 2.25 for every step in the grayscale... or as close to that as you can get. .05 variation in gamma is fairly difficult to see, so if the gamma stays between 2.2 and 2.3 for every grayscale step, you'll have good-looking images. A variation of .1 in gamma measurements is moderately visible so +/- 0.1 variations (total range of .2) is getting fairly obvious and will impact the appearance of images.

Bottom line is that AVERAGE gamma looks at all the gamma values and calculates an average. You could have half the grayscale at 0.6 gamma and the other half at 3.3 gamma and the average gamma would be 2.25 but images would be horrendous. You really need to focus on gamma for each grayscale step and try to minimize the gamma differences between steps.

Understood, so we want as flat a line as we can get from 10% to 90%.

However, it's how we get there is what I am trying to grasp. Michaels's post appears to be stating that raising the gamma setting in the user menu is not the way to go as others have suggested because the user menu setting is somewhat accurate, but you need a very small window to get the reading. Using a 10% window will give you a lower than actual reading and then making adjustments with gamma ire may give you the flat line but you are not calibrating to the gamma measurement that you think that you are.

Example:
1. You set the picture mode gamma to 2.2 in the display.
2. You test your gamma and it is showing 1.9
3. You go into the menu and change the gamma setting to 2.4 in the display.
4. You test gamma again and now it's showing 2.2.

You think you are good because you have the result you were looking for, but you really don't because you are calibrating with 10% windows. If you were to run the same gamma test with much smaller windows, the result would be closer to the actual gamma setting that you selected in your display.

Is this correct or am I misunderstanding what Michael is stating in post #9?

I skimmed over this thread and I didn't see you mention what meter you are using.. Just curious as some of the older low end meters drift a lot and cause this type of issue.

That's what I thought I had remembered.. but sometimes you gota go back and check the basics.. Hope you get it figure out..

Interesting, where did you get the pattern and what size windows did you use?

He said it above AVSHD790 and GCD discs. Small APL patterns

'SillySally' has reported good results using the approach we suggested using a grey ramp to set the probe distance from the screen and set the patch size.
We don't use APL patterns as they distort the ABL in a way that is not replicated with normal image scenes.

Worth a look/try...

Steve

The GCD disc has 1% windows so I am guessing that's what he used. You do get different reading with 10% windows reading a lower gamma and the 1% being about the same as whatever you have the display set to. I question whether you would use the 1% windows for a full calibration, I imagine you may run into issues with meter distance especially with plasma.

I would also like to try this, but how do I make my own patterns?

Here are the patterns for those interested.

Here is one using PC levels 0 to 255 graduant.jpg 313k .jpg file

and one with Video levels 16 to 235 graduant.jpg 310k .jpg file



Next step for most of us is to figure out if the large or small window patterns (or something inbetween) works best.
Edited by JimP - 12/10/12 at 5:28am

MichealLTV mentioned in a previous post in this thread that Panasonic and THX used a 6x5 window patterns for the 2011 panasonics.

JimP could you please photoshop a 30% and 80% ire 6 x 5 window pattern like THX uses to see how it would work....Thanks

Peter,

I'll be glad to when I get back this evening.

What unit of measure is the 6X5 window pattern suppose to be?

Think a 50% window would be handy also to verify for us OCD types.

That's the million dollar question, specifically relating to plasma. 10% seems to be the consensus, but there is a discrepancy with how accurate the gamma number is. As stated, MichaelTV mentioned that 6.5% window pattern provided an accurate number and the larger you went with window size, the lower the gamma reading. Good luck.

JimP, I know you have the 65vt50 too, if and when you figure out the right size, let me know! . I'll be too bussy getting to know my radiance mini the next couple of weeks
Edited by Wouter73 - 12/10/12 at 6:14am

Was that 6.5% window for a particular brand/model plasma? I'm specifically interested in my Panasonic 65VT50.

Panny GT30/VT30 and possibly the newer 50 series...Micheal can probably answer this better though.