• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

Seeing black on black

Doug Kerr

Active member
Some recent lovely work by Chris Calohan included "very dark shadows" on a background that was "very dark gray".

As the image was rendered on my system, I could not distinguish on-screen the "very dark shadows" (about RGB 3,4,5) from the "very dark gray" background (about RGB 15,15,16). Apparently others here had the same experience.

This may lead to a discussion of the rendering of the bottom end of the gray scale by way of browsers onto typical display chains.

To help give a quick idea of what we each get in this regard, I have just "spruced up" a grayscale test image I have used in the past. Here it is presented through the forum in the usual way, from a JPEG file:


Then color space here is sRGB. Each "block" (I have not put visible boundaries between the blocks) has a color of R=G=B= the number mentioned. Those values range from 0 to (full black) through 64*, in steps of 4 units of RGB.

*Ideally relative luminance 0.0513, "94.9% gray", 4.29 stops down from maximum luminance.​

Across the top is a bar of RGB=0,0,0 (full black). This allows us to distinguish how well (if at all) can one of the gray colors be distinguished from black in our display chain and viewing environment in an "adjacent" juxtaposition.

The image file itself can be downloaded here:

As a JPEG file:

http://dougkerr.net/images/test/sRGB_step_wedge_22.jpg

As a TIFF file:

http://dougkerr.net/images/test/sRGB_step_wedge_22.tif

I can readily make other test images in this same vein if there is anything one of you would like to have.

As a matter of interest, on my current system, in its current state, in my current viewing environment, in my current state, with the file displayed from its presentation on the forum, in this very idealized juxtaposition, I can just barely distinguish the RGB 24 block from the RGB 0 block.

I would say, rather arbitrarily (this is of course so very subjective), that in this situation, the smallest distinction against black that could be exploited in an image (assuming a uniform "shadow" with a distinct outline, against a uniform RGB 0 background) would be with about RGB 32.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Doug,

On my wife's iMac, (see, I disown it), I first see some distinction at 30, LOL!

.............at night, it might be different!

Asher
 

Doug Kerr

Active member
Hi, Michael,

Hi guys,

I can see the boundary line between 8 and 12 if I draw the shades. Shade up it is 12/16.
Nice to hear from you, and what a good report. Sounds like you are well tooled up (as we would certainly expect).

Thanks.

Best regards,

Doug
 

Maggie Terlecki

Active member
Not all of us have a "Spider" or high end monitor that is constantly calibrated and may find this useful:

Several years ago I found this one website by a brilliant guy called Han-Kwang Nienhuys. (learn more about him here: http://en.wikipedia.org/wiki/User:Hankwang

It helped me calibrate mine well enough to trust what I would get as a print. I just recently changed monitors and need to do it again, but this site has several different tests you can do, both for black to white but also contrast, gamma, gradient banding, etc., etc., It is also a very interesting read. I do think it can help those of us with lower end monitors. http://www.lagom.nl/lcd-test/

Maggie
 

Asher Kelman

OPF Owner/Editor-in-Chief
Not all of us have a "Spider" or high end monitor that is constantly calibrated and may find this useful:

Several years ago I found this one website by a brilliant guy called Han-Kwang Nienhuys. (learn more about him here: http://en.wikipedia.org/wiki/User:Hankwang

It helped me calibrate mine well enough to trust what I would get as a print. I just recently changed monitors and need to do it again, but this site has several different tests you can do, both for black to white but also contrast, gamma, gradient banding, etc., etc., It is also a very interesting read. I do think it can help those of us with lower end monitors. http://www.lagom.nl/lcd-test/
Maggie

Thanks for the link!

This is a very sophisticated routine. I presume it's accurate and wonder what if any of these parameters the Spyder or other hardware-software combo might miss?

Asher
 

Chris Calohan

Active member
8/12 requires a good stare but I can just see it. 12/16 and up is far more distinct and I did recently Spider calibrate. In PS, I can distinctly see 4/8 and a good stare reveals 0/ - I think, but as has already been pointed out, more likely because I know it is there. The actual image show 0-2 in the blackest areas and 8ish on the BG so it should be more distinct. I am posting a segment with Doug's scale to see if I can get a better representation of what I see in PS as opposed to what translates via internet interpolation. The posting is slightly larger than Asher generally likes to see but I am looking for special shading and think I need the extra space to see it.

 

Asher Kelman

OPF Owner/Editor-in-Chief
I have re-set my wife's iMac and now can distinguish readily 12-16. I can, with staring imagine 8-12, but, only at some odd angle.

What do you think about trying lightening your blacks so that the 0-2 is 8-12, just as an alternative presentation for us here?

Asher
 

Doug Kerr

Active member
Hi, Chris,

I am posting a segment with Doug's scale to see if I can get a better representation of what I see in PS as opposed to what translates via internet interpolation. The posting is slightly larger than Asher generally likes to see but I am looking for special shading and think I need the extra space to see it
Yes, I made my chart wider than is normally handsome for that same reason.

For what it's worth, here are the RGB colors of the patches on my chart as they appear in that composite image (both as captured from my browser and as directly downloaded):

Patch RGB

0 0
4 3
8 8
12 21
16 32
20 41
24 50
28 56

I'm not sure how you combined the two images.

In this composite, your "flower shadows" are typically RGB 2, and the background RGB 12.

All very interesting, wot?

Best regards,

Doug
 

Doug Kerr

Active member
Hi, Asher, Maggie,

This is a very sophisticated routine. I presume it's accurate and wonder what if any of these parameters the Spyder or other hardware-software combo might miss?
We have to be very cautious about these "works without any instruments and plays through your browser" tools. They are a little bit like things you can do at home with kitchen ingredients to find out if you have Worfernheimer's Disease, and at what stage.

For example, the visual gamma testing image apparently suffers somehow in the trail to and through my browser. It indicates gamma values that are at considerable variance with the results via colorimetry with my Spyder3 system.
This uses a clever scheme of visual gamma checking that has been around for years, but it is susceptible to misbehavior from various causes. I have often had bogus results from that general scheme.​

But if I take the gamma test image file as downloaded from the Lagom site (they are all in a handy ZIP archive) and open it in a respectable viewer, the gamma indications it then gives comport well with the Spyder3 colorimetric values for my display system.

The "grayscale" wedges as seen on my browser also do not follow the intended plan (the discrepancy is not large).

This is probably mischief conducted by my browser (perhaps some color space handling), but in any case it screws up the testing to some extent.

I would suggest that anyone interested in using Niehuys' instrument-less test protocol download the set of images and work from those.

Meanwhile, I'm thinking of seeing a phrenologist, perhaps in Albuquerque, about my Worfernheimer's Disease.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Meanwhile, I'm thinking of seeing a phrenologist, perhaps in Albuquerque, about my Worfernheimer's Disease.
Doug,

The phrenologists in Texas are a secretive lot. If you get to find them, beware that they, following custom, like to clean the skull before making any measurements.

Asher
 

Doug Kerr

Active member
Hi, Asher,

Doug,

The phrenologists in Texas are a secretive lot. If you get to find them, beware that they, following custom, like to clean the skull before making any measurements.
I'm sure.

I'll try and stay out of Texas.

Thanks.

Best regards,

Doug
 

Doug Kerr

Active member
Hi, Asher,

This is a very sophisticated routine. I presume it's accurate and wonder what if any of these parameters the Spyder or other hardware-software combo might miss?
The Lagom test suite (I inadvertently referred to it as the "Lacom" suite in some of my earlier writings - sorry) and a monitor "calibration and profiling" system such as the Spyder have quiet different, albeit somewhat overlapping, missions.

The primary tasks of the Spyder system are to:

A. Calibrate the display system ("monitor" and its driver) so that, to the degree feasible, it will, by itself, support some "target" color space (often sRGB). This is done by Spyder writing appropriate values in the "look up tables" (LUT) in the monitor driver.

B. Determine the residual discrepancies between the calibrated system and the "target" color space, and generate and save in a file a profile that a profile-aware application can use to "precompensate" for those discrepancies to yield a high-precision rendering.

Having done that, the Spyder system ordinarily does not report to us anything (other than it is finished). If we have one of the more advanced forms of the software package, however, it will display for us via various tables and graphic curves what it found, what it did, and (in the most important aspects) how close is the calibrated monitor's behavior to the desired behavior.

The basic task of the Lagom test suite protocol is to let us learn many properties of our display system. But this allows us to somewhat "calibrate" our monitor to the extent that there are user-accessible controls for key factors, such as brightness and contrast (and maybe some others). This is very nice, but falls far short of item A, above.

For example, in many monitors, we can "set the gamma". But that usually means choosing between 1.8 and what is called 2.2 (but isn't that simple). The Lagom suite allows us (if we take the proper steps!) to ascertain just what gamma our monitor exhibits at three different luminance levels. But if those are not the correct ones to suggest that the sRGB transfer curve is being properly followed, the user choice between 1.8 and "2.2" would not let us improve on that.

The Lagom suite allows us to ascertain a number of things about our monitor that the Spyder suite (at least Sypder 3, with which I am familiar) can't or doesn't deal with. For example, the Lagom suite provides some nice test images that let us see how the color rendition of our display varies with angle of view. It allows us to learn of certain timing discrepancies that might occur in an analog connection to the monitor. (And some advanced monitors in fact have provision for "tweaking" this.)

The Lagom suite also will let us ascertain the pattern of R, G, and B dots in each pixel cluster in our monitor (several different schemes are used). We need this if we wish to advise the Windows ClearType sub-pixel aliasing system of that detail so it can best do that voodoo that it do so sort-of well. (I doubt if many of us have used the Windows ClearType Tuner, the control panel where this voodoo can be tweaked, or if one does not believe in it, suppressed. You may not even have that in your Windows installation.)

So the Lagom suite is not a substitute for calibration and subsequent profiling of the display chain with an instrument-based system. It will, if we don't have such a system, allow us to make some important basic settings in the interest of best monitor performance under our situation.

And of course it will tell us a lot of interesting, and perhaps useful, things that a system such as Spyder does not deal with.

Best regards,

Doug
 
Top