Can Accurate Displays Still Be Different From One Another?

1. black level

2. max light output (while still producing a reference image)

3. viewing angle

4. motion resolution

5. input lag

6. contrast ratio (on/off and ansi)

7. gamma response (how linear and range: 2.2 to 2.5)

8. display tech

9. product features like number and type of inputs, internet access/services, etc.

In addition to the list above, things I can think of off the top of my head:


power efficiency

display structure visibility

resolution/MTF

longevity and wear

controllability (serial, IP, etc)

bezel size

thickness & weight

capabilities in ambient light/reflection rejection esp. w/plasmas

motion interpolation (which I consider Satanic)

motion handling, including black-frame insertion

flicker and decay issues

rainbow artifact issues

bit-depth/banding issues

color posterization (temporal bit-depth)

environmental ruggedness (outside, etc)

Net and app features, like netflix, pandora, etc

audible noise/buzz when the display is on

presence of CMS

scaling/deinterlacing capabilities

pro/digital signage features

Chris thought of a whole pile of things that I hadn't considered as I'd taken the view that the question was regarding image quality. I'd thought about things like whether the colour saturation is linear at all points, but then figured that if it was an 'accurate display' then that would be a given. Likewise adding CMS shouldn't be necessary if the picture is already accurate as it would only serve to make it less accurate if adjusted.


I'd take the view that the three main calibration factors of colour temperature, gamma and colour gamut are all measured with zero delta E. So therefore I think things like contrast ratio improvements (on/off and ANSI types) and maintaining full resolution with movement, plus the peripheral items like internet TV features, cosmetic and ergonomic would still differentiate displays.


I think I won't be holding my breath waiting for such a device though.



Quote:
Originally Posted by GeorgeAB /forum/post/19440913


If all video displays could be built to produce an accurate picture, would there still be plenty of ways for manufacturers to make models that could be differentiated from one another? If so, how?

Click to expand...

I believe that all depends on your definition of "an accurate picture". To some, "accurate picture" means "accurate color". For instance, PlasmaPZ80U mentions viewing angle. Well, if you wanted to be really harsh, any deviation experienced when viewing the picture from an angle, would be inaccurate.


Also, I believe a lot of peple underestimate how complex it actually is to say that you have completely accurate color. You may measure gamut and greyscale, and note that two displays measure exactly the same, while still looking very different color-wise. This is because different sets have different behaviour, wrt. different stimulus. Say you had two sets that measured theoretically perfect color-wise, in ANY area (i.e. identical regardless of stimulus), and had identical gamma and similar contrast, they would indeed look very similar. Other areas of performance would seem fairly negligible in comparison.


My take non this is that in reality, most sets differ at least as much in very simple, basic areas of performance, i.e. gamma and color linearity, as they do in other areas. I personally think that very often people are misplacing the responsibility of good or bad performance, based on whatever technology is being touted on that particular display. Take for instance Pioneer Kuro's. They have (had) very good black level, and, when properly calibrated, very good color performance (though not perfect). A lot of people notice the good black level, and the good color, and conclude that good black level is needed for good color - which is false. These kinds of assumptions lead to a lot of misinformation. It is not very often that you have a chance to see exactly what a difference in color linearity, and ONLY in color linearity, actually does to the picture. It's usually a "package deal", where the difference in color fidelity is "hidden".


Anyway, long story short, yes there would be other areas of performance where you would see differences, but basically, if you had linear and accurate color, linear gamma, decent black level, and no _really_ obvious artifacts in the picture, you would have a great picture.


Personally, I think people are just trying too hard, when it comes to delivering good pictures. Seriously, get the basics right, and don't create artifacts that stand out to the naked eye, and you have a winner. It's not rocket science - or rather, it shouldn't be.



Quote:
Originally Posted by Kelvin1965S /forum/post/19445100


So therefore I think things like contrast ratio improvements (on/off and ANSI types) and maintaining full resolution with movement, plus the peripheral items like internet TV features, cosmetic and ergonomic would still differentiate displays.

Click to expand...

I think CR improvements are decreasingly important. What we need is "good enough", then one should focus on other areas of performance, that REALLY need it. Very few higher-end displays have bigger issues with black level, than they do with color.


Re: movement resolution. I think this, too, is overblown. I think that 95% of movement issues arise from signal processing, not panel deficiencies. If you had _only_ the movement issues of the panel itself to worry about, the issue would be so negligible that noone would care much about it. It has become an issue, becuase the panels have been blamed for issues caused by the signal processing, or indeed the source material itself. One thing that has created this stir, is frame interpolation. This has improved movement, while introducing artifacts, to the point where people have started to dream about a frame interpolated picture that has the advantages but not the disadvantages of FI. But this will just never happen. Because of this, people are starting to see movement issues, where there really are none. They are comparing to an imaginary quality that is just not going to happen, with the kinds of source material we have today.


One example: Black frame insertion. I bet a lot of people is reading these forums and going "oh, if just we could get black frame insertion, we would have perfectly smooth motion". No, we won't. We would get motion that is a little better than what we had before, but still not perfect - simply because the source material doesn't allow it to be perfect. That's what it all comes down to: The source material AND (not least) the individual set's ability to deliver the source material as-is, is the real limitation of our picture quality, not the display technology. We don't actually need improvements to our display technologies, we need to use the ones we have accurately! THAT'S where the holy grail to better pictures is. The difference between the potential of a current, high-quality display (regardless of tech) and a theoretically perfect display, is actually fairly SMALL, compared to the difference between the actual performance of displays today, and the potential performance they COULD deliver if you designed them properly.



Quote:
Originally Posted by Otto J /forum/post/19489372


I think CR improvements are decreasingly important. What we need is "good enough", then one should focus on other areas of performance, that REALLY need it. Very few higher-end displays have bigger issues with black level, than they do with color.

Click to expand...

I'm coming to this from a front projection point of view I suppose. I don't care much for TVs as long as it's good enough for day to day viewing of mostly standard definition viewing...having said that I've calibrated mine as well as they will allow (some even quite close even without CMS).


For front projection there is still along way to go in terms of contrast ratio, both in ANSI type and on/off IMHO. Although I only have a lowly JVC HD350, with a VideoEQ Pro for CMS, the calibrated image has 'perfect' colours (as far as the i1Pro and Chromapure tell me after adjustment). However, I can see that the blacks aren't totally black and that even in a near black pit there could be further improvement on the ANSI type contrast side. I've seen more upmarket versions in very good demo rooms and there is still room for improvement on those (I'm also seeing the new X7 later this month and I'll put money that it still isn't 'good enough'). In front projection terms I consider 'good enough' is when it gets to a point that screen masking for 2.35:1 becomes unneccesary, but maybe that's a bit unrealistic as the room would have to be made of some totally light absorbing black substance. In that case, just being able to count to 5 in a fade to black and still not 'see' the screen would work for me, along with at least 12-14fL at peak white.


However, for TV I have to say I agree regarding the contrast of TVs. Whatever the on/off contrast on my spare room TV is, then I think it's 'good enough' even though it's only an LED lit Sony LCD model, so for me I guess it depends on the size and the importance of the viewing content.

I don't agree: I had a calibrated Panasonic AE3000 and now have a calibrated HD350. Even allowing for the fact that I'm now not viewing them in the same room side my side, I know that the HD350 looks 'different'. This difference is to do with native contrast in this case and better optics (my AE3000 was especially soft in this regard). Motion of the two images (which calibration does nothing to change AFAIK) is different: The AE3000 could be made smoother using frame creation, but with it turned off there was more judder than the HD350 which has no frame creation mode to turn off (I did demo them side by side before upgrading to the HD350).


There's a few differences for you that are outside the usually applied values for calibration of greyscale, gamma and gamut. Yet if I went purely by your metric they should have both looked identical.



Quote:
Originally Posted by Otto J /forum/post/19489357


I believe that all depends on your definition of "an accurate picture".

Click to expand...

I have no personal definition of "an accurate picture." I rely on video industry standards from international standards bodies to define what the reference for "accurate" is. Any other definition would be strictly subjective, unique to the individual viewer making up their own standards, and of no particular relevance to any other viewer, display owner, or manufacturer. It's my conclusion that much of the confusion in the discussion of video display performance comes from an ignorance of, and/or avoidance of, fundamental video industry standards, engineering guidelines, and recommended practices.


Do these video industry display performance standards cover every aspect of every type of video display behavior and function? No. That's the point of this thread. Even a "perfectly" calibrated display can have unique characteristics, while still producing "an accurate picture." Isn't it bizarre that so many manufacturers devote enormous effort and cost to make displays that include so many options, features, characteristics, etc., but fail to provide the most fundamental performance characteristic- a correct video picture?


Best regards and beautiful pictures,

G. Alan Brown, President

CinemaQuest, Inc.

A Lion AV Consultants Affiliate


"Advancing the art and science of electronic imaging"



Quote:
Originally Posted by Mr.D /forum/post/19490969


I was assuming we didn't have to bother discussing displays that are already hamstrung.

Click to expand...

I don't follow: In my example both displays were calibrated to very similar (ie under 'perceptable' delta Es in each case) greyscale, gamma and gamut results using an outboard (LUT type) CMS in the manner that George implies in his above post. I don't see what was 'hamstrung' in my example, just that there is a difference in the image as the JVC had a better native contrast for example.


Does your Plasma/Projector example use some kind of LUT device or perhaps a video processor so that both displays have the same colour gamut? I'm very aware of how wide the JVC colour gamut is, though I'm not sure if the 42PHD8 is similar or even if it has a CMS built in, so unless you're applying some external correction to the JVC to match the 42PHD8 then I'm struggling to see how both could have a 'constant overall look to the image'.



Quote:
Originally Posted by Mr.D /forum/post/19490515


You profile the displays. You generate correctional LUTs and apply them in a CMS. All the displays should then look exactly the same. If you have any that can't deliver the required brightness or black level you replace them with ones that can.


I've had rooms with over 50 calibrated displays in there ...they all looked the same. Even some less than ideal displays that were technically not passing calibration woul usually be in the ballpark close enough to look pretty much the same as the others.

Click to expand...

A CMS cannot address:


on/off CR

ANSI CR

viewing angle

MTF

display structure visibility

motion handling

temporal artifacts

banding artifacts

quality of any DI implementation

quality of zone lighting implementation


etc.


You can't just stick a CMS on any display with oversaturated physical primaries and just turn it into a reference display or make it look the same as anything else.



Quote:
Originally Posted by ChrisWiggles /forum/post/19491636


A CMS cannot address:


on/off CR

ANSI CR

Click to expand...

It can as long as the "better" displays are matched to the "lesser" displays and both still have enough precision and gamut to successfully map a desired target colorspace into. Although generally when it comes to white point I let it scale to whatever the display is capable of producing without clipping to avoid running into precision problems especially with LCDs.


Critical color displays I usually lock to a target white point , print for example.

Quote:
Originally Posted by ChrisWiggles /forum/post/19491636



viewing angle

MTF

display structure visibility

motion handling

temporal artifacts

banding artifacts

quality of any DI implementation

quality of zone lighting implementation

Click to expand...

This would fall into the category of inadequate tolerance for your desired purpose ...ie chuck it and get a new display if any of these things are a limiting factor. You cannot for example use a CMS to turn a monochrome display into a color one at the risk of sounding facile.


etc.

Quote:
Originally Posted by ChrisWiggles /forum/post/19491636



You can't just stick a CMS on any display with oversaturated physical primaries and just turn it into a reference display or make it look the same as anything else.

Click to expand...

Well you can as long as you are going from a large pot into a small pot with all your necessary tolerances. Its exactly how you turn a video projector into something you can grade film on or an LCD or a CRT (or a plasma but thats cutting it fine in some areas...nice client displays though).


All that aside you can have a room full of different displays and they all look the same and they all meter the same ...its kinda important if your client is walking about and your workstation displays are not all over the place.