AVS HD 709 - Blu-ray & MP4 Calibration - Page 99

I found gamma to be stable on my display with APL (video level) > 15-20% so I use the larger APL windows and set peak white below where the brightness limiter kicks in for this patterns.

Thanks zoyd. I subscribe to the AVS Samsung PNXXD8000 calibration thread and will search and if necessary ask you there how you figure out what level the brightness limiter kicks in at (although I assume it is that the gamma curve drops off at higher IREs, the question is using what window to measure it.)

Thanks for the clarification, so is there any difference between:
a) everything at video(16-235) levels (TV, graphic card, renderer)
b) everything at PC(0-255) levels

One more question: BluRays are video levels(16-235), are MKV's ripped to video or PC levels?

Thank you for all your help, alot is clearer now
So, conclusion is, if HTPC is primarily for video reproduction, best is to stick with video levels (16-235).

Who would thought that things that seem so simple, are in reality so complicated. Or I'm just stupid

Thank you for the thoughtful response. So if I understand correctly, in addition to all the factors about how gamma is, or should be, defined, measured and calculated by these various calibration software products, there is an additional variable with some of these plasma sets which have some sort of built in "floating gamma", and all we are doing is trying to find the right stimulus to measure and adjustments to fudge so that at least this quirk of these sets is negated as much as possible. And currently the best stimulus appears to be the large APL windows (but with a depressingly low contrast setting). That about right?

I agree, the avoidance of generating a flat gamma when ABL is active may not be relevant to others. I run my display in a dimly lit environment at what many consider to be low peak white values (28-35 ftL). The majority of the time I will never hit ABL picture levels so it just makes sense to calibrate using a "typical" scene, ie one where ABL is not active. For others who through preference or due to bright rooms that want to run peak whites much higher have two options:

1. Flatten gamma using a lower peak white which does not activate the ABL and then raise contrast to the desired level afterwards
2. Flatten gamma using the desired final peak white values.

The first case will generate a too low gamma for very bright scenes (ABL active) while the second case will generate a too high gamma for low to mid brightness scenes (ABL not active). I have no idea whether one would be able to distinguish the difference between the two cases.

OK, I'm continuing with plasma TV calibration.
Primarily, I'm using this tutorial:
http://www.curtpalme.com/forum/viewtopic.php?t=10457

It says that light output on plasma displays on 100% IRE pattern should be 30-40 ftl.

I'm calibrating for night viewing mode, so I calibrated to 32ftl (I've seen lot of people calibrate to that value in LG xxPK550 owners thread (that's my set).

Second, in tutorial is said that brightness should be calibrated to 0,65% of desired light output at 10% IRE pattern.

So, at 32ftl, I have Y = 110.76, which gives me 0.72ftl at 10% IRE pattern.
But, I have 2 issues here:

1.When I calibrate contrast to 32ftl, I see all white bars flashing in white clipping pattern
2. If I set brightness to 0,65% of desired ftl, my picture is too bright, and I see all black bars flashing in black clipping pattern

So, are these formulas correct? Or there is another preffered way to calibrate contrast and brigtness.

Also, I found that Limited RGB in ATI drivers + Limited RGB in renderer (MPC-HC EVR CP) gives me grey black (it seems that signal is double compressed).

So, these are my settings:
I'm using Eye One Display 2 with HCFR for display calibration.

1. Low black at TV set (16-235)
2. Full RGB in ATI driver (0-255)
3. Limited RGB in renderer (16-235)
4. Contrast calibrated to 32ftl at 100% IRE pattern
5. Brightness calibrated so I can't see bars 0-16 flashing in Black clipping pattern
6. Using LG 20 IRE point calibration method

Below are my calibration results (didn't do CMS yet). What it concers me, that gamma setting is too low (below 2,2). How can I correct this:

1.Luminance


2.Gamma


3. RGB Levels


4. Color temperature

One can assume scenes with dark shades will have low APL and scenes with bright shades will have high APL but scenes with very high contrast will be rare. And it more or less can be simulated with window patterns.

But my main concern regarding the APL patterns (either small or big) was thet my TV has a relatively big and opaque OSD box to control the calibration settings.
How could I avoid the fluctuating APL if the OSD slider and it's background box covers a lot of the screen area?
Yah, I could exit from the menu from time to time, but it would be very time-consuming.

It would be nice if Samsung TVs would be remote controllable with disabled OSD like the new Panasonics.

EVR-CP doesn't support limited RGB output with proper whiter-than-white because the mixer output is strictly full range RGB with clipped WTW and the limited output is only a result of a post-process shader code.

madVR (it's a third-party renderer which works with MPCHC and some other players) can output limited range RGB with BTB and WTW information.

The ATI driver pixel-format options are all fakes. The limited RGB is only a post-process on the full range RGB framebuffers (like the shader in MPCHC but the results aren't dithered back to 8/10-bit as nicely as with MPCHC) and the YCC mode is "double-fake" (the framebuffer is still RGB full and don't expect HQ RGB->YCC conversion with advanced dithering, it's more like a compatibility but not a quality feature...)

But it's very simple: The OSD (partially) covers the patches on the screen. And the patches move on the screen, so different set of patches can be covered with different arrangements.
Take a look at the attached images (sorry, I just made those with my phone, I didn't fix the exposition time but it doesn't really matter now).



I know. I already did it. Well, it wasn't a nicely constructed test and I didn't documented anything but I figured out that I must use small window patterns because I ran out of the limits of the 10p grayscale adjustment sliders with the standard window (I set -10,-10,-10 on the [-10;+10] scale and I wasn't even close to my luminance target with IREs ~60-90 while small window measurements suggested reasonable adjustments).

I never tried to calibrate with the APL patterns though. And I think I want to lower my peak luminance anyway (I had a smaller LCD when I used to the 120 cd/m^2 target), so I guess I will re-calibrate my TV today and give the APL patterns (large or small) a shot.

I ended up using the Large APL Patterns. The small APL windows gave a significantly lower 100% IRE fl vs the large windows. My display is a Samsung 55 inch LED backlit LCD. You may want to check this before you get too far along.

Are you sure about this? I'm seeing BTB (0-16 bars on black clipping pattern) and WTW (235-255 on white clipping pattern) with EVR-CP set on Limited RGB (my TV also expects limited RGB, ATI outputs FULL RGB).

I know about madVR, but my Zacate (AMD Fusion E-350) isn't powerful enough to drive 1080p movies, only 720p.

Hey Guys,

I tried (really did) to understand the gamma dilemma with the pattern size and whether or not the pattern is APL. If anyone finds the time to explain this in a thread I made, I bet it would make a lot of confused ppl happy

Well, it depends on the actual MPC-HC version.

The latest trunk and experimental test builds will always work with Full Range RGB mixer output because our tests implied that this is the only output format which is correctly handled by the EVR mixer. For some reason, it clips the Cb and Cr levels of the YCC output formats.
It's Microsoft's fault, the custom renderer in MPC-HC depends on the EVR mixer which is part of the Windows OS.
I plan to install a Windows 8 preview build to check out if anything changed yet.

Hi everyone. I was just wondering if anyone has tried using this DVD on an LG LH7000 HDTV? I have one, and would like to cross reference my findings with others who have the same set. Many thanks in Advance.

EDIT: Out of all the preset modes on an LG TV, what is the best starting point for "tweaking" the picture? (Such as Game, Cinema etc.)#

EDIT EDIT!: What are the best settings for gamma etc when using this DVD? I normally just default the preset, then go from there (Whilst turning off 100hz technology etc.)

do you have 10 point IRE settings on that tv?

Edit: i see that you have 20 point settings?
thats great
well i think i will have the nobel prize for this
i just discoverd that you can adjust the gamma exactly the same way that you adjust the grayscale
The 10 point IRE settings is also connected to the gamma curve

in your case you should lower the brightness for IRE 10 to 60
higher the brightness for IRE 70 so that you get down to 2.2
lower some brightness for IRE 80 and more for IRE 90

in chroma pure you have the calculated Y target values for the gamma so its easy to adjust it
i dont remember if you have those targets in HCFR

as for the adjustment:
check the Y value at the same time as the whitebalance
start to higher/lower Red/green/blue for each IRE at The same amount

if you put some work to it you could also get this Reference gamma curve/line
http://privat.bahnhof.se/wb192876/LG...ryDetailed.pdf

so how is it going with your crappy gamma curve?
speed up man
you cant watch your tv with this ****** settings

Sry, I had too much work in the last two weeks.
Will try to recalibrate this week If i find some time.

Thanks for the tip

Checking your report i'm seeing that you have used 100% Color Gamut Patterns.

Generally 75% Color Patterns are recommended and you will have better results if you re-calibrate your Display using them.

Doing that you will have similar dE errors but the image will be better.

Luminance of 75% Color Patterns has about 50% of the luminance of 100%, so your adjustment point is at the center of the color luminance range.

Using 100% Color Patterns your adjustment point is at the highest end of the luminance range but you will propably have larger errors at lower luminances.

You can check how your display is tracking at different saturation levels using ChromaPure Saturations Module + AVSHD Saturation Patterns.

This will give you accurate results if your display is tracking 2.22 to 25%,50%,75%,100% of Grayscale because ChromaPure Targets are calculated from a 2.22 Gamma & AVSHD Saturation Patterns RGB Triplet Values used are generated for a 2.22 Gamma Only.

I gave the Large APL patterns a shot but it lead to a disastrous result with my "hacked" D550. It was hard to hit the target until the charts eventually looked good and the actual colorimetric performance was obviously poor with real movie contents (the picture reminded my to my old IPS LCD in it's uncalibrated state). However, the Small APL patterns proved to be a bit better than my reduced 1/4 size window patterns.

Funny; on my D550 LCD, large APL patterns produce the best result and full field patterns are a very close second. Both windows and small APL patterns result in gamma readings that are way out of range (very high point gamma readings). Based on this, you must have the D550 plasma.

Regarding the APL Large and Small Gray Patterns, if you were to quantify the APL that each of these holds constant, what would that be? Basically, what would the equivalent percent stimulus for a full field pattern for each? I hope my question makes sense as I'm not sure exactly how to phrase it.