Decoding Gamma Function - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 7 Old 05-12-2013, 01:48 PM - Thread Starter
Member
 
fluxo's Avatar
 
Join Date: Oct 2012
Posts: 186
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 56
I know calibration software allows you to correct the gamma decoding of a TV. But what is the gamma decoding function that is actually being targetted?

In theory you could use the inverse of the Rec. 709 OECF. In practice, you can't, because most TVs are incapable of producing true black (0 cd/m^2). That means the bottom section of the function is not practically invertible. So what function is actually being targetted by calibrators?

Cheers.
fluxo is offline  
Sponsored Links
Advertisement
 
post #2 of 7 Old 05-13-2013, 12:35 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,565
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 144
Kind of a Calibration for Beginners' question. smile.gif

99.9% of the time (gamma) calibrations target the basic power law function (or a slight variation thereof.)

Recently (within the past 1.5 yrs,) there has been some (glacially paced) movement toward BT1886 on displays that are capable of approximating it without any nasty side-effects ...
HDTVChallenged is offline  
post #3 of 7 Old 05-13-2013, 12:27 PM - Thread Starter
Member
 
fluxo's Avatar
 
Join Date: Oct 2012
Posts: 186
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 56
Quote:
Originally Posted by HDTVChallenged View Post

Kind of a Calibration for Beginners' question. smile.gif

Yes. Perhaps I should have read around here more before posting. If I had, I would have found this interesting discussion:

http://www.avsforum.com/t/1409045/how-power-law-gamma-calibration-can-lead-to-crushed-blacks

Quote:
99.9% of the time (gamma) calibrations target the basic power law function (or a slight variation thereof.)

Recently (within the past 1.5 yrs,) there has been some (glacially paced) movement toward BT1886 on displays that are capable of approximating it without any nasty side-effects ...

That is helpful, thank you. I do wonder about this basic power law function. Here's a plot of a power function:



The blue points represent calibration points for a TV with a poor MLL. What I mean by that is that they represent a certain TV's response at stimulus levels of 0%, 10% and 20% after calibration.

The first point could be moved upward by adjusting the TV's brightness control, but it can't be moved down below the MLL. The second and third points could be moved by, say, using 10-pt gamma controls. Clearly there is a problem in that the dashed section of the gamma curve cannot be accurately targetted with this TV. I would imagine this is a well-known problem.

The other thing that is bothering me is this: most TVs have quite coarse-grained calibration controls. So the gamma response might be adjusted at, say, 10 points. But that only matches those particular points to the target curve. How do TVs deal with all the intermediate points? Do they reproduce what is essentially a piecewise linear function? I.e., something like the green curve below?



I have to emphasize that the graph is for a TV with an exceedingly poor MLL. Which is in fact the MLL (0.24 cd/m^2) that David Katzmaier measured on the DT50, so it's perhaps not impossible to find an example that is that awful. For a better TV, such as one of the recent Panasonic plasmas, a much closer match to the target curve would be possible.

My apologies for being a complete beginner and any abuse of terminology that might cause :--)

Thank you for your patience.

[EDIT] Perhaps I should say in passing that my labelling of the power function is somewhat misleading. What I've actually plotted is a graph of 100(x/100)^2.4. 100 cd/m^2 is the target peak luminance.
fluxo is offline  
post #4 of 7 Old 05-13-2013, 02:05 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,639
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 43 Post(s)
Liked: 175
Hi Fluxo,

I agree with and understand everything you pointed out.

But the truth of it is, almost everyone simply uses a power curve at 2.2 with no offset for black level.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #5 of 7 Old 05-14-2013, 11:12 AM - Thread Starter
Member
 
fluxo's Avatar
 
Join Date: Oct 2012
Posts: 186
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 56
Quote:
Originally Posted by sotti View Post

Hi Fluxo,

I agree with and understand everything you pointed out.

But the truth of it is, almost everyone simply uses a power curve at 2.2 with no offset for black level.

Thank you very much.

Could I ask, do we know what the various TVs do between the calibration points? Let me put it this way: suppose a display is calibrating using a 10pt gamma control system at stimulus levels 10%-100%. Do we then have an idea what is happening between 20% and 30% etc.? Have people plotted the measured gamma response in (say) 1% increments post-calibration?

Graphically, suppose this is the target gamma response:



Is the following the sort of approximate response a TV is likely to produce?



Cheers.
fluxo is offline  
post #6 of 7 Old 05-14-2013, 11:40 AM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,639
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 43 Post(s)
Liked: 175
Quote:
Originally Posted by fluxo View Post

Thank you very much.

Could I ask, do we know what the various TVs do between the calibration points? Let me put it this way: suppose a display is calibrating using a 10pt gamma control system at stimulus levels 10%-100%. Do we then have an idea what is happening between 20% and 30% etc.? Have people plotted the measured gamma response in (say) 1% increments post-calibration?

Graphically, suppose this is the target gamma response:



Is the following the sort of approximate response a TV is likely to produce?



Cheers.

It all depends on the TV. The only way to really know is to validate.

In CalMAN you can measure an arbitrary number of points so you can see in between a 10 point step to find out what is going on. I know the Lumagen as a processor does piecewise linear in between calibration points, but that's layered on top of the displays native curve, so the in between points certainly aren't linear.

Every manufacturer does something different with their internal CMS, and none of them make their algorithms known, even to us. Panasonic doesn't tell us how a VT50 goes from 10% to 20% to 30%.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #7 of 7 Old 05-14-2013, 01:07 PM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 228
Just because a display has only 2-point or 10-point grayscale controls doesn't mean you can't measure more than 2 or 10 points. CalMAN allows you to measure 21 points easily. So if you have only 2 or 10 adjustment points, you can still see how 10 or 11 or 21 (or any other number you want to setup as a custom measure) looks when you measure additional points. Of course you have to have a test pattern source with the 20 or 21 or more points available... if you're relying on a disc with 0%-100% patterns in 10% steps, you'll only be able to measure 11 points. If you have a video processor or test pattern generator with programmable pattern generation, you could, theoretically measure 128 points or more if you wanted to and were willing to wait for all those measurements to complete.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
Reply Display Calibration

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