Home · Register · Search · View Winners · Software · Hosting · Software · Join Upload & Sell

Moderated by: Fred Miranda
Username   Password

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Post-processing & Printing | Join Upload & Sell

1      
2
       end
  

Archive 2011 · Why Color Profile?
  
 
Peter Figen
Offline
• • • •
Upload & Sell: On
p.2 #1 · Why Color Profile?


"If your monitor is correctly calibrated the card should also look gray on screen."

Having R=G=B neutral gray numbers looking gray on screen does nothing to guarantee that the colors your monitor displays are calibrated. Neutral RGB is only one component and if that's what you are relying on primarily, as Chuck so often says, you would only be fooling yourself. The reality is, is that most people make adjustments based on what they see on screen, so that monitor calibration, and it's accuracy is very important. You can only go so far with the numbers, which are great for telling you if a gray is neutral or setting endpoints, but can tell you very little about things like overall contrast.



Dec 19, 2011 at 04:37 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.2 #2 · Why Color Profile?


Hrow wrote:
Thank you. I understand now and appreciate all of the thoughtful and lengthy responses to my question. . The component that I was missing was the relativity of the process to variable real world conditions as it is usually presented / sold as a "scientific" necessity. To me, a scientific process implies repeatable, constant results and that was the error of my ways. I did not factor in that "repeatable" and "constant" was also fluid.



you can nuy a thermometer than can repeatably measure a given temp to 0.0001F and yet on any given day, if you put it outside it may deliver quite a different number to you than it had on some other day

with monitor calibration it's pretty much constant all the time (as far as probes are concerned, ignoring metamerism here).... other than brightness/backlight setting (sometimes people chose a slightly different tone curve for pitch black vs sunny hdtv viewing though too)



Dec 20, 2011 at 01:58 AM
theSuede
Offline
• • • •
Upload & Sell: Off
p.2 #3 · Why Color Profile?


Yes, but the main problem here as I see it is to NOT put people off calibrating their screens... That's why I'm quite disappointed in many of the softwares you get with many of the cheaper profiling solutions.

The screens should NOT end up as being "very different", that's hrow's main point, and in this he is absolutely right.

MY main point is that the most important (first step, basic level CM) things are that your screens:
1) Have (about) the same WB temp
2) Are about as bright as each other, and about as bright as the recommended value - with small adjustments for ambient light and preference
3) Have the correct tone curve - this is maybe the most important thing in a "normal gamut" screen
4) Have a reasonably correct profile, so that color intensitys are shown correctly, with the right hue angle.

Point (1) and (2) are pretty easy with most heads/softwares and screens. Most softs have an option saying "manual adjustment" where you fiddle around with the physical controls on your screen until some kind of measurement indicator says "yeah, keep it here". Most screens use the "contrast" denominator for setting maximum screen luminance.
Point (3) is a BIT more complicated, since it often includes playing around with the other screen lightness option, "brightness". This will screw up measurement (2), so often you have to do this over again, like:
Do adjustment (1)+(2), and then adjustment (3). Now (2) will have changed slightly - start over.

Most of the time this needs to be repeated at least once, but it only takes a minute or so... Or 30s if you have done it before.

After this your screens should be showing similar "whitepoint brightness", "color temperature" and "total contrast". And this is as far as you can take the screen with CALIBRATION, which is the more physical part of the setup, fiddling around with controls on the screen. Their gamma curves (the midtone brightness) might still differ a bit, but you need a screen with programmable LUT's or continually adjustable gamma to make this right. Most people don't, so on to the PROFILING.

Here you HAVE to tell the software to let the screen be "as it were", and don't let the software fiddle around with absolute whitepoint or whitebalance. The ONLY two things a profile on a cheaper screen should do are:
1) Correct the tone curve so that you get a smooth, correct slope gamma
2) Inform the computer about the maximum saturation and hue of the R, G and B channel outputs.
........

After the "calibration step", which WILL be a manual labor for most people (we often forget that most ppl don't have tax deductable premium range programmable hardware screens...) the screens should be pretty similar, except for the midtone brightnesses and absolute color accuracy (hue and saturation).

After the profiling step (that again, should NOT adjust maximum brightness and WB!) and the application of the screen profile in the system, midtone brightness curves and a color not on the edge of what the screens are capable of should show as almost identical.

If not, then something is WRONG - and you might need brand-specific support. Get it, in the end it's worth it. At least, "get close", and verify by measurement.
But don't obsess over absolute accuracy - that's for people in the high volume production business, where accuracy is needed to be able to guarantee customers consistent results. Here it can be a 100k$ loss to get a print run botched, an then you have to show hard proof to the customer that you're "within contract specification" unless you want to waste a few tons of printed material.



Dec 20, 2011 at 10:28 AM
cgardner
Offline
• • • • •
Upload & Sell: Off
p.2 #4 · Why Color Profile?


@ Peter:

You've quote me out of context below, implying things I never suggested. The context was discussing process control STARTING AT THE CAMERA CAPTURE.

When Custom WB off a gray card is used and the camera captures a full range of tone with detail the engineering of the camera produces a "nominal" result, which in photography is an image that is accepted as being "real" with seen in facsimile. When those two basic conditions for correct reproduction are met AT CAPTURE, the file out of camera can become an OBJECTIVE BASELINE for judging the color calibration state of the monitor.

That's not to say there aren't other important factors in calibrating a monitor, simply saying that if you start with a nominal file out of camera (i.e., one with known correct gray balance) and you detect a color cast when it's viewed on the monitor what would be suspect is the calibration of the monitor.



Peter Figen wrote:
"If your monitor is correctly calibrated the card should also look gray on screen."

Having R=G=B neutral gray numbers looking gray on screen does nothing to guarantee that the colors your monitor displays are calibrated. Neutral RGB is only one component and if that's what you are relying on primarily, as Chuck so often says, you would only be fooling yourself. The reality is, is that most people make adjustments based on what they see on screen, so that monitor calibration, and it's accuracy is very important. You can only go so far with the numbers, which are great for
...Show more



Dec 20, 2011 at 01:28 PM
 

Search in Used Dept. 



RustyBug
Offline
• • • • • •
Upload & Sell: On
p.2 #5 · Why Color Profile?


+1 @ theSuede makes a good point at differing monitors / calibration methods affecting LUT vs. non-LUT. This is somewhat akin to my point that my two different monitor types CANNOT achieve the same level of luminance (CRT vs. LCD) ... i.e. non-LUT vs. LUT, the former CANNOT achieve what the latter can.

While I realize mine are mixed types and nobody else would be doing that, it points out that it matters a bit as to how the monitors were designed by the mfr and not all monitors have the same capability/method for establishing control ... i.e. some allow for making LUT adjustments ... others, only brightness controls. If a monitor is limited to manual adjustment of brightness/contrast controls, then (imo) expensive calibration equipment is of lesser value. If instead LUT adjustment is available, then the better calibration equipment is of much greater value.

Using a precision measuring device such as a micrometer makes good sense for precision work ... but not so much for a process that can't take advantage of its precision. I routinely use a vernier caliper for measuring hardwood dimension when cutting pieces for geometric cutting board designs. But for general construction / rough carpentry work, a good tape measure is all you need ... the process for dimensioning 2x4's simply doesn't have the same level of precision that hand planing hardwoods does, so even if you do use a vernier caliper to measure a 2x4 and find out it isn't square within .001" ... "ain't nuttin' you're gonna do 'bout it" ... you're simply going to use it. 2x4's are diminshed "good enough" for the task they are intended for @ "nominal' dimensions ... not precision tasks, so measuring 2x4's with a vernier caliper is a waste of time, money & effort (unless you have the ability to re-dimension them to such precision).

The difference between LUT vs. non-LUT is going to be the dividing line between what your expectations can be ... and the value of the differing calibration products from 'basic' to "professionally advanced". If your system/monitor won't harness (allow for LUT adjustment) the precision of advanced 'scientific' calibration equipment ... there is a degree of "overkill' with the more advanced calibration devices. Conversely, if your system has the capability for such precise adjustments, a "basic" calibration device won't get the job done properly.

LUT vs. non-LUT ... first order of business in assessing what your calibration approach/expectations should be relative to your monitor/system ... imo. Until that is established/defined ... it is easy to see why such "mixed" opinions exist as to "critical" vs. "unnecessary" ... likely some are assuming LUT, others not.

BTW ... I'm "clueless" @ Mac on the LUT matter.




Edited on Dec 20, 2011 at 05:10 PM · View previous versions



Dec 20, 2011 at 02:33 PM
RustyBug
Offline
• • • • • •
Upload & Sell: On
p.2 #6 · Why Color Profile?


One thing that should probably be expanded upon (others better than I) is the difference between monitor calibration vs. profiling.

Knowing that Y-X vs. X-Z exists, the adjustment that Chuck makes seems to be one easy "grass roots" way of roughly contending with it (likely better than my dual monitor approach). This is of course "less precise" than using a professional profiling system that evaluates the print output, but I think it does an ok job of corroborating the point @ Y-X vs. X-Z and the unrealistic expectation that the two should match without some form of compensation.

Whether that be from a scientific profile, visual manual adjustment or a mental estimate of the how much darker the print will be ... or providing additional illumination for the print ... imo, the acceptance that there must be some form of compensation between the monitor vs. print serves to relieve much angst from the faulty notion that the two could ever be equal without such compensation.

Monitor calibration is one issue ... print profiling is a different issue.

cgardner wrote:
This is baseline nominal file on screen, but it will print darker...
http://super.nova.org/TP/TowelGary.jpg
Knowing that I will make a Levels tweek before printing to open the midtones to compensate for the "dot gain"...
http://super.nova.org/TP/TowelGaryLighter.jpg
The net result is that the lighter screen file when printed is a closer match to the "nominal" screen file.




Dec 20, 2011 at 03:18 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.2 #7 · Why Color Profile?


theSuede wrote:
Yes, but the main problem here as I see it is to NOT put people off calibrating their screens... That's why I'm quite disappointed in many of the softwares you get with many of the cheaper profiling solutions.

The screens should NOT end up as being "very different", that's hrow's main point, and in this he is absolutely right.

MY main point is that the most important (first step, basic level CM) things are that your screens:
1) Have (about) the same WB temp
2) Are about as bright as each other, and about as bright as the recommended value - with small adjustments
...Show more

Most screens have very, very poor controls for tone response curve (some HDTV now have 10pt controls and such but many still don't and few computer monitors seems to have anything of the sort and all the presets and contrat/brightness settings will leave many screens with very different TRCs) so I'm not sure you often can get them to look reasonably similar in that regard until after the probe calibrates the TRC to the monitor in hte graphics card's LUT. Some probably do let you at least get close something like gamma 2.2 but many do not and hardly any to something like sRGB TRC.

One thing to watch for is that many monitor have very cheap color engines so trying to adjust primaries in them just makes for a cracked and twisted profiles and many behave worse if you move WB and brightness away from defaults, it depends. You have to go through the entire process adjusting as much as you can ahead through calibration with controls and then try leaving some controls at default instead. Some controls on some monitors do more damage than good and getting them closer through calibration can actually lead to a worse end result. It really varies tremendously screen to screen though.




Their gamma curves (the midtone brightness) might still differ a bit,


shadows too, many screen have the curve take a dive at the low end to make blacks appear to be deeper and richer and the screen more contrasty.




Here you HAVE to tell the software to let the screen be "as it were", and don't let the software fiddle around with absolute whitepoint or whitebalance. The ONLY two things a profile on a cheaper screen should do are:
1) Correct the tone curve so that you get a smooth, correct slope gamma
2) Inform the computer about the maximum saturation and hue of the R, G and B channel outputs.
........


some build a complex LUT though and report more than just the xyY of the primaries

some screens have very poor saturation tracking so they may be close to ideal at say 100% luminance but then be well undersaturated or well over at say 60% or 20% etc. a simple matrix based profile would offer zero correction for this at all, sometimes a fancy LUT profile can make up for this a bit but sometimes thing can still be a bit off



Dec 20, 2011 at 07:52 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.2 #8 · Why Color Profile?


theSuede wrote:
Yes, but the main problem here as I see it is to NOT put people off calibrating their screens...


but this is the main point, calibration is very useful and so long as the probe isn't way off (although some cheap ones can be at times) it almost always makes things a good deal more consistent no matter the screen (although some can be made more consistent than others)



Dec 20, 2011 at 07:54 PM
1      
2
       end




FM Forums | Post-processing & Printing | Join Upload & Sell

1      
2
       end
    
 

You are not logged in. Login or Register

Username   Password    Retrive password