contrast setting and white clipping pattern - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 6 Old 12-25-2012, 08:46 PM - Thread Starter
Senior Member
 
milehighou's Avatar
 
Join Date: Jun 2008
Posts: 420
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 15
I'm using my PS3 and did a calibration with the AVS709HD disc on my Samsung LN52B750. The common contrast setting on this TV seems to be in the high 90s. When using the white clipping pattern with the flashing bars (see below)

http://cache.gawkerassets.com/assets/images/17/2011/11/0c200cfa46cebd5305fcdaf5e27e2c26.jpg

I can set contrast to 100 and all bars remain visible up to 253, but there is a definite pink tint. As the contrast increases much past 92 or so, the red seems to run out. I loaded a 109% white window and kept raising the contrast to determine this. My question is, how important is it if the super white signal (the white above 235) looks pink in that pattern? If that maximum white level in a movie is 235, and it looks ok, does it really matter if the whites higher than 235 look pink?

I finally settled on a contrast setting of 95. On the flashing bar pattern, super white still looks slightly pink, but when I look at a grayscale ramp, I don't see pink (just black, gray and white). Also, the gray step pattern looks good. If I want the super white to lose the pinkish hue in flashing pattern, I have to turn contrast down to 90, and the picture is too dim even with the backlight at 4 (it's about 26 ftL). I'd really prefer not to turn the backlight up past 4 because black levels would probably suffer. With the contrast at 95, and the backlight at 4, I get right around 30 ftL. If I look at the RGB color clipping pattern with contrast at 95, red does clip a little, but it doesn't occur until about 241.

Here's some screen caps from HCFR


milehighou is offline  
Sponsored Links
Advertisement
 
post #2 of 6 Old 12-25-2012, 10:19 PM
Member
 
dominickwok's Avatar
 
Join Date: Jul 2009
Posts: 188
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
When the red starts running out at say 100% level, the above white will exhibit a bluish tint (as measured by your meter); but when you look at the white flashing bars at various stimulus levels at the same time, your brain will quickly adapt to "treat" the bluish-white as the "pure white"; thus making all the other flashing bars look "reddish" (even though your meter tells you that the white at 100% is at D65). When it comes to judge D65 white, you can't trust your eyes.
dominickwok is offline  
post #3 of 6 Old 12-26-2012, 05:00 AM - Thread Starter
Senior Member
 
milehighou's Avatar
 
Join Date: Jun 2008
Posts: 420
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 15
Quote:
Originally Posted by dominickwok View Post

When the red starts running out at say 100% level, the above white will exhibit a bluish tint (as measured by your meter); but when you look at the white flashing bars at various stimulus levels at the same time, your brain will quickly adapt to "treat" the bluish-white as the "pure white"; thus making all the other flashing bars look "reddish" (even though your meter tells you that the white at 100% is at D65). When it comes to judge D65 white, you can't trust your eyes.

So it sounds like I'm OK if that white flashing bar patterns seems pinkish to my eyes? It's the only test pattern that has that tint; the rest look pretty decent. My 100% white has about 99% red and the delta E is ~3.9.

Also, do you think I could get a better contrast ratio if I upped the contrast all the way to 100, and then lowered the backlight to 3. Contrast 95 and backlight 4 has me at about 30 ftL. With contrast at 98, backlight 4, I measured almost 2000:1 ratio. Now, it's about 1700:1.
milehighou is offline  
post #4 of 6 Old 12-26-2012, 06:44 AM - Thread Starter
Senior Member
 
milehighou's Avatar
 
Join Date: Jun 2008
Posts: 420
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 15
Found this in a older thread:
Quote:
Originally Posted by PlasmaPZ80U View Post

I have a question about the White Clipping Pattern. I calibrated my Samsung LCD TV via my PS3 using the AVSHD disc and set video output format to YCC with super-white off. When I turn on super white, the white background in the pattern gets brighter and shifts in color quite noticeably while the flashing bars turn from neutral gray to a very obvious pink. When doing the same with the 10% grayscale windows or full fields, the 100% pattern stays neutral white regardless of super-white being on or off, while the 109% (254) pattern does the exact same thing described above with the white background of the White Clipping Pattern. So, is the background of the White Clipping Pattern 100% (235) or actually 109% (254)? Why do the bars from 230-234 look neutral gray with super-white off and then turn pink with super-white on? My contrast is maxed out at 100 from the default 95 and so there's likely a color shift above white, but 100% measures very close to D65 (CIELUV dE of 1.3). I don't see why the 230-234 looks right with super-white off but plain wrong with it on. Perhaps the background is not actually white (100% or 235) with super white on and so the bars from 230-234 just look pink relative to that background.

Quote:
Originally Posted by alluringreality View Post


The background is 254. If I decide to release a new version the APL pattern will be switched to 235. My current impression is that there's limited, if any, above-white material on the majority of commercial video. Generally my thought is to calibrate to white.

Depending on the setting either 254 or 235 is likely sent, so probably the display has a color-shift for above-white.

Anything above 235 probably goes to 235 with super-white off.


Quote:
Originally Posted by thomasl View Post


This most likely means that the display is compressing/clipping red with the current contrast setting. I have a number of Samsung LCD HDTVs and they all do this if I crank the contrast up above a certain point. The reason why the 230-234 bars appear pink is that there is a lack of red in the 254 encoded background thus it brings out the red in the lower encoded bars thus making them appear pinkish. When you switch off Super-White, the PS3 is clipping anything above 235 thus you should not be able to see any flashing bars above 235 and they should all blend into the background and the lack of red at the highest end is hidden from you.


Whether or not you want to preserve above white information and/or avoid clipping/compressing of red/green/blue channels above 235 is really a personal preference decision. For me, I've tried to keep contrast below the point I see color compressing/clipping beginning to occur even if it's above 235. For most of my Samsungs that is usually around a contrast setting of 95.


hope this helps,



--tom
milehighou is offline  
post #5 of 6 Old 12-28-2012, 05:21 AM
Advanced Member
 
hungro's Avatar
 
Join Date: Mar 2007
Posts: 589
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 56
Hello. What I noticed on my samsung plasma is that before I got a meter my greyscale was very red. Including the white which appeared a little pink. After calibration this red/pinkish tint is completely gone. I also did not lower my contrast. On my set contrast of 95 and cell light at 20 I have 32ft lamberts max output. The point I am trying to make is that initially I thought I had contrast too high because of how the white clipping pattern appeard to my eyes. But this was a case where the Red gain was way too high. Default 25 Redgain, calibrated 0 Red Gain. Without a meter I would not have known how far to lower the Red Gain. Without a meter you would not be able to make these corrections.



hungro is offline  
post #6 of 6 Old 12-28-2012, 11:17 AM - Thread Starter
Senior Member
 
milehighou's Avatar
 
Join Date: Jun 2008
Posts: 420
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 15
That's interesting that you had so much red all across the board. Mine was almost the opposite—too low. The again, mine is an LCD vs your plasma. I recently got the i1 Display 3, so going by my results from that, I know my 100% is good (RGB fairly even). As my contrast approaches 100%, the red begins to run out, but it happens in the whiter-than-white signal. Even though that flashing bar pattern looks pinkish, I decided to stay at 95 on the contrast. From what I've read, and like Dominick said above, this is like an optical illusion. Since the background of that pattern is 254, my eyes see that as white (even though it's too blue) so the other whites look pink.

Anyway, thanks for the info. I think I'm good to go for now. With contrast at 95 and backlight at 4, I get right around 30 fL. My red does clip on the AVSHD color clipping pattern, but it doesn't happen until around 243 or 244, so I don't think this is of any significance. Between 16-235, things look good. If you look at that first post, I've got a slight blue hump, and a little red dip, but that's about the best I can do without ruining the low end. It looks like I should bump up the red offset a notch to help with the lower midrange, but if I do that, 20% gray looks too red to me.

Here's my before:

milehighou 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