theoretical question about display color reproduction accuracy - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 14 Old 02-19-2014, 10:07 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 745
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 30 Post(s)
Liked: 65
Question 1:

Suppose you have an 8 bit display whose characteristics match the following two conditions:

The three primaries are perfectly on target

The white point is perfectly on target for each of the 256 luminance steps.


Does it then follow that each of the 256 saturation steps for each primary will be perfectly on target?


Question 2:

Suppose that these three conditions are met:

The three primaries are perfectly on target.

The white point is perfectly on target for each of the 256 luminance steps.

All 256 saturation steps for each primary are perfectly on target.

Does it then follow that each of the 16.7 million colors will be perfectly on target?
spacediver is offline  
Sponsored Links
Advertisement
 
post #2 of 14 Old 02-19-2014, 11:28 PM
AVS Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 6,353
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 14 Post(s)
Liked: 204
1. No. Perfect white balance and gamma does not guarantee an accurate gamut.

2. No. Bad color decoding could screw up the secondaries even if the primaries are perfect.
ConnecTEDDD likes this.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #3 of 14 Old 02-19-2014, 11:37 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,582
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 163
In a theoretically perfect display that was purely additive in the way it mixed color, the answer would be yes to both questions.

When the rubber meets the road there are significant obstacles. Some that are introduced by the display producer, some that are intrinsic to display technology, that cause this to be a pipe dream as Tom points out.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #4 of 14 Old 02-19-2014, 11:54 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 745
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 30 Post(s)
Liked: 65
thanks for the replies.

The additivity issue is what I've been trying to wrap my head around.

Barring color decoding artifacts, what would be an example of a failure of additivity in any given display?

Let's take the first question

You have the correct primaries, so your basic units which you're adding together are sound.

You have your white balance perfect across all 256 luminance steps, which means that the correct amount of signal is mixed together for each of the three sub pixels when the display renders shades of gray.

I'm assuming that saturation (or more accurately, excitation purity) would be reflected by the amount of white that is added to each primary. So if you're adding the correct amount to the correct primary, how come the saturation sweeps are not guaranteed to be accurate? To make it simple, let's assume that the saturation sweep also sweeps upwards in luminance, so that you start at red, and work your way up to pure white.
spacediver is offline  
post #5 of 14 Old 02-20-2014, 01:07 AM
Advanced Member
 
Light Illusion's Avatar
 
Join Date: Aug 2010
Posts: 518
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 211
This is basically what RGB Separation in LightSpace shows...

Steve
Gordon Fraser and ConnecTEDDD like this.

Steve Shaw
LIGHT ILLUSION

Light Illusion is offline  
post #6 of 14 Old 02-20-2014, 07:55 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,427
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 75 Post(s)
Liked: 299
Quote:
Originally Posted by spacediver View Post

thanks for the replies.

I'm assuming that saturation (or more accurately, excitation purity) would be reflected by the amount of white that is added to each primary. So if you're adding the correct amount to the correct primary, how come the saturation sweeps are not guaranteed to be accurate? To make it simple, let's assume that the saturation sweep also sweeps upwards in luminance, so that you start at red, and work your way up to pure white.

Because the SPD of the two opposing primaries that are added to (or subtracted from) the primary often overlap with or have side peaks in the wavelength region of that primary -> cross-talk.
ConnecTEDDD likes this.
zoyd is offline  
post #7 of 14 Old 02-20-2014, 09:46 AM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,582
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 163
Quote:
Originally Posted by zoyd View Post

Because the SPD of the two opposing primaries that are added to (or subtracted from) the primary often overlap with or have side peaks in the wavelength region of that primary -> cross-talk.

Or the physics of the display actually causes the cross talk by exciting adjacent sub-pixels.
ConnecTEDDD likes this.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #8 of 14 Old 02-20-2014, 06:03 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 745
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 30 Post(s)
Liked: 65
Quote:
Originally Posted by zoyd View Post

Because the SPD of the two opposing primaries that are added to (or subtracted from) the primary often overlap with or have side peaks in the wavelength region of that primary -> cross-talk.

But if we buy into the principle of superposition (Grassman's third law), then adding two SPD's will result in the same tristimulus values as adding the individual tristimulus values of each SPD.
Quote:
Originally Posted by sotti View Post

Or the physics of the display actually causes the cross talk by exciting adjacent sub-pixels.

I tried doing some calculations to see whether subpixel leakage would result in a "non additive" display, but my brain broke halfway through, so for now I'll take your word for it smile.gif
spacediver is offline  
post #9 of 14 Old 02-21-2014, 11:05 AM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,427
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 75 Post(s)
Liked: 299
Quote:
Originally Posted by spacediver View Post

But if we buy into the principle of superposition (Grassman's third law), then adding two SPD's will result in the same tristimulus values as adding the individual tristimulus values of each SPD.

I don't see how that's contradictory, the extra energy from the overlap is present in all the SPDs so that will be propagated whether you measure and add or add and measure.
zoyd is offline  
post #10 of 14 Old 02-21-2014, 12:07 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 745
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 30 Post(s)
Liked: 65
maybe I'm missing your point.

Any SPD, measured within the visible region, has an associated chromaticity.

Any two SPDs, added together, will produce a chromaticity equal to adding the individual chromaticities together. This happens regardless of how much overlap there is in the SPDs.

I don't follow what you mean by crosstalk here.

Are you saying that in some situations, adding together two SPDs produces a chromaticity that is not the same as adding together the two chromaticities?

Does your point have something to do with some of the limitations of the standard observer functions?
spacediver is offline  
post #11 of 14 Old 02-21-2014, 01:29 PM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,427
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 75 Post(s)
Liked: 299
No, I'm saying that regardless of whether measured separately or after addition the chromatic scaling will be the same and not correct if there is cross-talk.
zoyd is offline  
post #12 of 14 Old 02-21-2014, 01:37 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 745
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 30 Post(s)
Liked: 65
Ah, so if there is "leakage" between sub pixels, then crosstalk will occur in the case that SPD components overlap. I think that makes sense to me.

I thought you meant that the overlapping SPD components was itself a form of cross talk, even if there was no subpixel leakage.


So, to be clear, this crosstalk wouldn't occur if you were just combining two light sources together into a the same spot in space, correct? (similar to how the original color matching experiments were done).
spacediver is offline  
post #13 of 14 Old 02-21-2014, 01:55 PM
AVS Special Member
 
zoyd's Avatar
 
Join Date: Sep 2006
Location: Planet Dog
Posts: 4,427
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 75 Post(s)
Liked: 299
Quote:
Originally Posted by spacediver View Post


I thought you meant that the overlapping SPD components was itself a form of cross talk, even if there was no subpixel leakage.

That's exactly what I'm saying. It doesn't matter what the source is, subpixel leakage or a green phospor emission that contains a red primary side lobe, if the primaries are not independent (orthogonal) the chromaticity will not scale properly. It's not just that there is red emission in the green primary, it's that the red emission in a green primary is correlated with the red primary, that's what I mean by cross-talk.
zoyd is offline  
post #14 of 14 Old 02-21-2014, 02:31 PM - Thread Starter
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 745
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 30 Post(s)
Liked: 65
What you are saying seems to be in direct contradiction to (my understanding of) the principle of superposition (grassman's third law), a principle that doesn't include independence as a condition for linearity.

My understanding is that if you take any two points on the CIE xy space (e.g. two points that represent two primaries), the mid points between those two points will represent the combination of equal amounts of those primaries. Crucially, this holds for all metamers associated with those two primaries.

Anyway, I went and did some matlab experiments, and found that when I averaged two randomly generated SPDs together, their chromaticities didn't match the average of the two individual chromaticities. The Y was exactly the average, but the x and y weren't exactly the average (although they were close), so maybe this reflects what you were saying.
spacediver 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