or Connect
AVS › AVS Forum › Display Devices › Display Calibration › Color saturation without equipement?
New Posts  All Forums:Forum Nav:

Color saturation without equipement? - Page 2

post #31 of 46
I am not a pro like Doug B. but I can tell you of the display I have calibrated using HDMI input, most of the time the Default for Color and Tint is where it needs to be when the white balance is set correctly . Using the filter gets you one setting, then checking with meter gets you another, then watching content send you back to the defaults. As has been said more than once, no equipment to do WB, use a disk to, Set Brightness, Contrasts, Sharpness, and overscan, Watch content and adjust color/tint till you are happy.. but default is probably correct. NOTE: leave WB and CMS at the factory defaults, and choose the lowest color temp. Yes, it will look red compared to every TV you see in a store, it should! ;-)
post #32 of 46
Quote:
Originally Posted by delphiplasma View Post

Point taken...not sure where the 98% statistics come from? Filters maybe a bit hit and miss due to the chromaticity mis-match to display phosphor. However, in a well designed display the blue only method can acheive fairly good results. In theory the blue only mode should work.

98% comes from the same place that even a broken clock is right twice a day.

It's not that they will never be right, but the instances that they are right is pure chance and not any form of engineering.
post #33 of 46
I used "98%" because sometimes when the stars align and everything else is just perfect, you can get a good setting from a blue filter.

Blue-only modes fail ALMOST as often because the TV's Blue may not be accurate. If the TV's blue is not "right on" the HDTV standard for blue, there will be an error using a Blue-only mode. Ditto for Red-only modes and Green-only modes.

Theory is fact. We are dealing with hypotheticals when it comes to filters and single-color modes... both CAN work, but FILTERS rarely work right (because of uncontrollable variables like inaccuracy in the filter, aging of the filter causing it to change, or the filter not being a perfect match for the light source in any given TV (phosphor, CCFL, LED, projection lamp, etc.). And single-color modes, while they work more often than FILTERS, they are still subject to not producing an ideal result, especially if the color point in question on the display in question does not match the HDTV spec. So what you're really dealing with is the fact that filters or blue-only modes CAN work, hypothetically. But the reality is that there are large variables that compromise results so much that more often than not, you're in "don't bother" territory. Which means your only alternative (without software, meter, and test pattern source) is long-term viewing (over days) after tweaking Color or Tint or Color Temperature trying to find the ideal settings.

And when I say single-color modes don't work reliably because of errors in the color being used (i.e. blue-only or red-only, or green-only modes available in some TVs), I'm referring to both xy errors (essentially hue and saturation) AND luminance errors. All 3 have to be "right" for a single-color mode to work right (assuming you are using a standardized test pattern that assumes that xy and Y are correct before using the pattern).
post #34 of 46
Quote:
Originally Posted by Doug Blackburn View Post

I used "98%" because sometimes when the stars align and everything else is just perfect, you can get a good setting from a blue filter.

I guess the question is are we trying to set the color *decoding* or are we trying to coax "exact" REC709/REC601 chromaticity out of a reluctant display? The two things aren't necessarily the same. smile.gif

The "blue filter (or BRG filters) method" had a time and a purpose ... that time and purpose has been superseded by the availability of relatively low-cost and relatively accurate colorimeters and spectros. We now have the ability to go directly to the "correct" answer(s) - as long as we are willing to accept some inevitable loss of color space "linearity" as saturation tends toward 100% on displays that don't/can't display exact REC709/REC601 primaries (which would be most of them.) These abilities did not exist back in the day.

Furthermore, at some point, Joel et al. will have probably perfected automated calibrations and 3D LUT generation to the point that calibration will be as simple as running Audyssey (or the like) ... and then perhaps we'll no longer have need for "opinions." wink.gif
post #35 of 46
Filters never had a time or place. They never worked better than they work today because there have always been variations in phosphors, the filters themselves (including changes due to age, temperature, and humidity), and the electronics used in the video displays themselves. There was NEVER a time where you could use a filter and nothing but a filter to set the color or tint controls. There has ALWAYS been the caveat that if the display doesn't look right after using 1 or 3 filters, you made adjustments to the color and tint controls to make the images appear "normal" (by eye).

People keep saying filters are "better than nothing" but that's a HUGE stretch. 98% of the time, nothing and filters are exactly equivalent in their results. So 2% of the time filters are (maybe) better than nothing. Not much of a confidence builder for using filters.
post #36 of 46
Quote:
Originally Posted by Doug Blackburn View Post

When you have no meter and test patterns, you can only set controls by guessing at correct settings by looking at images. You cannot do this for 1 minute or 1 hour or 1 day... you have to choose a setting and view content for 2 or 3 or 5 days and decide if the setting is too much or too little. Then make a change and view TV programs or movies for 2 or 3 or 5 days again before making another change.
When I change a setting, I make only half the change that seems to produce the best result, then look at a variety of sources and try changing other settings, looking for an improvement, before considering any further change to the first setting. The reason for being cautious and tentative is that, while I'm confident I can choose a setting change that makes an image improvement, I can't predict the effects of changing to a different source or interactions with other settings. Proceeding without instruments, I don't think there is a guarantee that I can find good settings, but I think my odds are better if I take my time and do a lot of experimenting.
post #37 of 46
Quote:
Originally Posted by Doug Blackburn View Post

Filters never had a time or place.

Sure they did ... my point was that there has been a significant change of *philosophy* in the "art" of calibration since the "blue filter" method was devised. That change has rendered the method "obsolete" on modern digital sets with at least a rudimentary CMS.

Back in the the blue filter's day, you lived with the phosphors you had, not the ones you wished you had (i.e. SMPTE-C.) Maintaining linearity was considered more important than reaching specific x,y points. Furthermore, in most cases, there wasn't anything you could do beyond getting the green-yellow-red side of the gamut "correct" - a task that the blue filter was specifically designed to do. I've never seen a set that "failed" this task (even really bad ones,) provided you knew what the filter method was meant to do, how to use it *and* how to cross check it with red and green. I never had to go back and "re-adjust anything by eye," that appears to be something you've appended to the method ... I've never seen that in *any* instruction material.

It's really only been in the past five(ish) years or so, that we've had the tools and, most importantly, the display controls that allow us to go beyond what was the common practice for the first 45-50 years of (NTSC) color TV. (OK, actually for a lot of that time the common practice was to tell viewers to adjust their (Color/Tint) controls until the color bars looked right. The "BFM" was an attempt to remove the subjective "looks right" element and replace it with an objective method. Perhaps you've melded the two methodologies in your memory? )

PS: I'm all for squeezing as much accuracy as we can out of our modern digital displays, but I don't think we really need to shun the past in the process; on the contrary, if more people understood what the SMPTE bar chart and "filter" method represents mathematically, they might better understand what they're doing when they start playing around with their newfangled CMS controls ...

Calibration wasn't always about being perfect (and anything less just won't do,) it was about getting the best performance you could out of less than perfect displays ... while sometimes having to use stone knives and bear-skins to do it. I think in our obsessions with trying to achieve higher and higher levels of "perfection," we may have lost sight of that.
Edited by HDTVChallenged - 4/20/13 at 1:40am
post #38 of 46
If the blue filter is not an exact match for the blue (CRT) phosphor, there will be errors and it is damn near impossible to have a perfect match of filter and phosphor (for CRT or any modern display) because the filters are typically "free" and as such are not partlcularly well-controlled tolerance wise. And once the blue filter exists, it begins changing due to age so even if it was accurate when new, it doesn't take much time for the filter to become inaccurate and that just gets worse over time.

I've been an imaging systems engineer since 1972. I'm well aware of the capabilities and "usefullness" of the blue filter method. I have NEVER seen a set of instructions for using filters to set color or tint controls that did NOT include the disclaimer that if the images don't look right after using the filter, that you will have to re-adjust the color and tint controls by eye to further improve the image (an acknowledgement that the blue filter method is not reliable). If you have never seen blue filter instructions include that caveat, you have only seen the worst-written blue filter instructions in existence.
post #39 of 46
Yep, filters only get ballpark at best. True you need to tweak afterwards if blue filter doesn't match the display phosphor. However, the blue only mode always worked, this being applied to grade 1 broadcast monitors.

I guess it is similar to colorimeters with filters. In fact, caution needs to be applied with the compatibility of a meter with the display type.

Color is a very difficult process to acheive The only way i know my calibration is good is by comparing the calibrated display with a broadcast grade 1 monitor. Believe me, even a well calibrated consumer display can look considerably well off the mark when compared to an industry display.
post #40 of 46
If I remember correct the blue filter thing you hold worked better on the CRT. I find it does not help much on this plasma.
post #41 of 46
Quote:
Originally Posted by Doug Blackburn View Post

I've been an imaging systems engineer since 1972. I'm well aware of the capabilities and "usefullness" of the blue filter method. I have NEVER seen a set of instructions for using filters to set color or tint controls that did NOT include the disclaimer that if the images don't look right after using the filter, that you will have to re-adjust the color and tint controls by eye to further improve the image (an acknowledgement that the blue filter method is not reliable). If you have never seen blue filter instructions include that caveat, you have only seen the worst-written blue filter instructions in existence.

Then you and I were "taught" differently. In "my world," the only reason to go back and re-adjust would be if the results of the BFM left one of the *other* primaries over-cooked (a legitimate red-push, for example.) To "fix" the issue, you would use a red (or green) filter to reduce the Color control to the point where the offending, overcooked color matched the gray/white reference bar. Done. No "naked eye" tweeking involved.

Eventually, I discovered that it made more sense to just do the BFM on all three primary color channels and take the average of the three resulting Color and Tint values.

Back in the day, we tried to force displays to fit the math ... Today we try to make the math fit the display ... We also have modern tools that allow us to do that. We also have displays with native gamuts that far exceed SMPTE-C or REC709 ... We no longer don't know if we are measuring native display primaries with a 100%Sat/100%Stim target or something that's been blended down from a much wider native RGB (or even RGBY) gamut.

So YES ... Today, the BFM is problematic ... OTOH, we still have folks that think setting the luminance of 100% red to a fixed percentage of the luminance of 100% White is "correct," because they don't understand what the SMPTE bar chart w/ grey reference represents mathematically ...

The truth isn't always a simple black and white, either/or choice. wink.gif
post #42 of 46
Quote:
Originally Posted by HDTVChallenged View Post

we still have folks that think setting the luminance of 100% red to a fixed percentage of the luminance of 100% White is "correct," because they don't understand what the SMPTE bar chart w/ grey reference represents mathematically ...

can you elaborate on this?
post #43 of 46
Doug,

I think part of the problem here is semantics. Upon downing about a gallon of coffee yesterday afternoon, I realized that your concern is with the actual film or gel based filters not the method itself.

Over time, I've conflated the term "blue filter method" with what would more accurately described as the 'primary isolation method,' much in the same way that the community has come to use the name of a variable (gamma) to describe a mathematical function. ... So if that's the case, as Roseann Roseanadana would say, "Nevermind." smile.gif

I've considered the issue of "filter" shortcomings to have been dead and buried for at least a decade, so when the topic came up I saw it in terms of "old-school methodology" vs. "new school (if ya ain't using a spectro ...)" mentality.

For the record, I've only had to use filters on one of my displays ... on the others, I could achieve full isolation either by turning off individual guns (direct views) or covering them with a magazine (RPTV.)

So with that being said, where ever I said "blue filter method" or "BFM" above think "Blue Isolation Method" or "Primary Isolation Method," instead. That should get us back on the same page.
post #44 of 46
Quote:
Originally Posted by PlasmaPZ80U View Post

can you elaborate on this?

Can I? Yes ... Will I? Maybe later.

It's one of those things that's simple, yet a bit confusing at the same time. ... smile.gif

The good news is that if you're using Calman or Chromapure, you probably don't have to worry about it ... HCFR (at least the original version) is another story.

Edit #2: The short answer is that D65 white isn't always (and, in fact, is almost never) made up of 21.26% Red, 71.25% Green and 7.22% Blue on a specific display.
Edited by HDTVChallenged - 4/22/13 at 12:54am
post #45 of 46
Hi. Sorry if this has been covered elsewhere. How do you use the basic color/tint controls in conjunction with the individual RGB contols within the CMS options?
post #46 of 46
Quote:
Originally Posted by Charlie Kocis View Post

Hi. Sorry if this has been covered elsewhere. How do you use the basic color/tint controls in conjunction with the individual RGB contols within the CMS options?

Probably leave them at neutral, either that or move them to their most centered position as far as error goes for the different CMS controls.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › Color saturation without equipement?