ISO 12646 & 3664 conditions

[newbiephotoguy SourceForge]

Is it possibe to make “screen uniformity report” in CIEDE2000 so we can check agains ISO requirements (<= 4dE00 from center)?
Same applies to moving from a 5×3 grid to 5×5 grid.

If not possible, is there an easy calculator to transform to deltaE2000 your “deltaC + brightness%” data?

Thanks in advance

• This topic was modified on 2015-07-15 07:59:35 by newbiephotoguy.

[Florian Höch]

Hi,

Is it possibe to make “screen uniformity report” in CIEDE2000 so we can check agains ISO requirements (<= 4dE00 from center)?
Same applies to moving from a 5×3 grid to 5×5 grid.

It would require some changes and a program update, but that’s certainly possible.

[newbiephotoguy SourceForge]

Would you add this feature to next DispcalGUI release?

In meantime, is there a basic of simple formula to “translate” (with OS calculator) current uniformity report to what that ISO conditions need to check?

[Florian Höch]

Would you add this feature to next DispcalGUI release?

Yes.

In meantime, is there a basic of simple formula to “translate” (with OS calculator) current uniformity report to what that ISO conditions need to check?

No, but you can make the deltaE change until then. Edit the file report/uniformity.functions.js and make the following changes:

Line 67:
delta_C.push(delta["E"]);

Line 74 and line 91:
Change ΔC*00 to ΔE*00

• This reply was modified on 2015-07-21 21:44:12 by fhoech.

[newbiephotoguy SourceForge]

Is L-star so important when measuring uniformity?
I mean, I get -15% of luminance on one corner (coord 1,5), others corners are -10%. “Coloration” is not “too bad” all cells fall below 2dC, most of them 1dC or less, but when taking acount of L* + color in the way deltaE00 computes color difference formula, it goes to 4.3dE00 (not aceptable to ISO conditions).

Would you add a funcionality, like in DispcalGUI measurement report, to mark with two green ticks for “OK” values (ISO) and one green tick of “aceptable” values (your understanding of minimum uniformity)?
The second case (non ISO conformity) is important because to my eyes, when uniformity reports hightest dC values, these are the screen patches where non uniformities are easily spotted (even if this color deviation form center are aceptable “by sight”)… without taking account to the loss of luminance in [-15%, -10%] range on the four corners.

Is there public & free documentation avaliable regarding how uniform a screen “should” be?

• This reply was modified on 2015-07-24 07:55:29 by newbiephotoguy.

[Florian Höch]

Is L-star so important when measuring uniformity?

It really depends on how much deviation across the screen you’re wiling to accept.

Would you add a funcionality, like in DispcalGUI measurement report, to mark with two green ticks for “OK” values (ISO) and one green tick of “aceptable” values (your understanding of minimum uniformity)?

That’s one way to approach it, yes.

Is there public & free documentation avaliable regarding how uniform a screen “should” be?

Not that I know of.

[newbiephotoguy SourceForge]

I tried 3.0.3.4 snapshot with requested feature, thanks.
But I find less informative because deltaC information is gone, it is not displayed with dE.
I mean that clearly noticeable “less white” zones in the screen get double check pass in deltaE (1.9), but beacause its brightness uniformity was pretty good (-4%), the relative “huge” coloration that reported deltaC value is gone and gets diluted in overal good dE. I mean that in this case almost all dE error was “color” error, and its pretty more noticeable “by sight” than similar ~2dE patches were color and luminance errors are mixed.
Is it possible to include deltaC and deltaE at the same time in the report? Or at least to mark with a warn symbol those patches which could be easily understood as “hey, watch out here: almost all error is in color!”

It’s helpful to validate against ISO conditions for critical color enviroment (PASS/NO PASS) but for “general purpose” image displays I find more useful deltaC when thinking if to return unit for refund (>2dC). This kind of “common” displays are not likely to pass ISO conditions in huge numbers, so its evaluation should focus on “not awful spots”. For MY eyes, “awful to sight” spots in screen are usually related to high dC rather than -17% luminance. Previous DCG releases are more helpful to perform that task.

Maybe I wrote wrong my previous request not impliying that both dC and dE values were needed.

[Florian Höch]

You can switch between delta E, C, H and L by clicking the label in the report.

[newbiephotoguy SourceForge]

Sorry, I didn’t notice it. Thanks again

[Author]

Posts