New CiC Tutorial Added: Monitor Calibration

[McQ]

This one is not visible on the main page yet, but I thought I'd share it here first:

Tutorial: Monitor Calibration for Photographers

I still need to add perhaps one more section and do some more proofing, but it is mostly done. If you have anything which you think should be included based on your own experience, please let me know. As usual, if there are any typos (there's likely still several), or if something is unclear or misleading, please let me know as well.

Thanks!

[Dave Humphries]

Hi Sean,

Typos:
"Iif" in first word of second sentence.
Only found that one.

A fairly minor point, but in (1) Calibration, the text example of green=50 > 70 differs from that shown in the table below, which says 78. I think it would be better if they concurred (assuming they should).

Is it wise to use "(1)" and "(2)" twice?

Overall, and remember I don't do this myself, I found the tutorial understandable.

I thought the graphic of Wide to Narrow colour gamut in (2) Profiling worked well to show the effects we typically see with Adobe RGB being displayed as sRGB.

When the LUT is first mentioned in the Monitor Calibration Devices, I wondered where it was; in some new s/w that came with the calib. device that forever more sat between programmes and screen, the video card or the monitor itself. From later text, I gather it will be the one in the video card - correct?

Speaking personally, one thing I have often wondered about is the standard gamma 2.2 thing; I knew of this from my TV production days and how that gamma is applied in the TV camera. However, I am still unsure if it is applied in the digital camera, or during jpg creation (camera or RAW processing), or not at all (but just in the display), and does this explain why I see more in the blacks on screen than when printed - or is this just my very badly mis-profiled CRT at work? I don't print very often at all, so it's not a huge issue for me, and if something comes out dark, I tweak it and print again (I can almost feel Colin cringing!)

The standard elements Levels dialog has an apparent 1:1 default gamma, which can be adjusted to give results like the 4 pics of the chap in the flat cap, so is this before or after the 2.2 gamma is applied?

Both my ancient CRT and laptop's LCD can show all the steps in Shadow and Highlight Detail, so I have assumed I don't have a problem. However, neither show your 3 box calibration as the text in Mid-tones implies (I noted this to be under review).

One other thing I was keen to understand was how the LUT bit depth affected things, I think I got it from the 3 graphs ok. Although I did wonder if the central graph for "2 output shades" should have had less horizontal grey lines than the left hand one? (since the right hand one has more)

I think most of the above shows my ignorance as much as anything, but it may help you understand where misconceptions arise. (well, mine anyway )

Cheers,

[Colin Southern]

" all the time spent image editing and post-processing" -> " all the time spent on image editing and post-processing" Perhaps?

" is to just adjust" -> " is to simply adjust " Perhaps? (too many "justs")

On a small side note, I have my monitor set for 200 Cd/m2 - it's several years old and still doing just fine

Overall I think it's really good - it's a hellishly complex area that I think you've explained really well; I'll bet it took more than 10 minutes to write that one!

[McQ]

Colin, Dave: Thanks a ton! Everything's been incorporated and it's now accessible from the main page.

Originally Posted by Dave Humphries

When the LUT is first mentioned in the Monitor Calibration Devices, I wondered where it was; in some new s/w that came with the calib. device that forever more sat between programmes and screen, the video card or the monitor itself. From later text, I gather it will be the one in the video card - correct?

Yes, the LUT that you adjust when you perform calibration is usually in the video card -- unless you have a special monitor that supports changing its own internal LUT (much less common). In the latter case, there's still an LUT in the video card that applies gamma, but the monitor LUT is the one that performs the corrections from calibrations.

Originally Posted by Dave Humphries

Speaking personally, one thing I have often wondered about is the standard gamma 2.2 thing; I knew of this from my TV production days and how that gamma is applied in the TV camera. However, I am still unsure if it is applied in the digital camera, or during jpg creation (camera or RAW processing), or not at all (but just in the display), and does this explain why I see more in the blacks on screen than when printed - or is this just my very badly mis-profiled CRT at work? I don't print very often at all, so it's not a huge issue for me, and if something comes out dark, I tweak it and print again (I can almost feel Colin cringing!)

I will get back to you on this one...

Originally Posted by Dave Humphries

The standard elements Levels dialog has an apparent 1:1 default gamma, which can be adjusted to give results like the 4 pics of the chap in the flat cap, so is this before or after the 2.2 gamma is applied?

Anything displayed on your screen is already using gamma 2.2, so any gamma adjustments in levels are *in addition* to this native gamma. In order to create the levels diagrams I therefore had to figure out what curves adjustment is needed to offset or add to the gamma 2.2 to create gamma 1.0, 1.8 and 4.0.

Originally Posted by Dave Humphries

Both my ancient CRT and laptop's LCD can show all the steps in Shadow and Highlight Detail, so I have assumed I don't have a problem. However, neither show your 3 box calibration as the text in Mid-tones implies (I noted this to be under review).

The images themselves are correct, I just wanted to rework the wording (which is done now). Viewing angle is pretty important with LCD's, so just make sure this isn't affecting things. Otherwise calibration is something to consider for both monitors...

Originally Posted by Dave Humphries

One other thing I was keen to understand was how the LUT bit depth affected things, I think I got it from the 3 graphs ok. Although I did wonder if the central graph for "2 output shades" should have had less horizontal grey lines than the left hand one? (since the right hand one has more)

The LUT bit depth only prevents degradation if it has a higher bit depth than the input. This way rounding does not necessarily decrease the number of tones/values. This is why the central graph and the leftmost graph are shown with the same number of horizontal lines. I could have shown the central graph with fewer horizontal lines, but that would apply to a much smaller subset of people's displays (probably 5% or fewer).

In the example graphs, even though the central "low bit depth" LUT has the same bit depth as the input, the number of shades is decreased because they have to round to the nearest integer after the tonal curve is applied.

This concept is identical to how applying lots of tonal curves to an 8-bit image will gradually reduce the number of total colors in the image. Each time a curves or levels adjustment is applied, everything has to round to the nearest 8-bit color value. This is why people generally recommend editing using adjustment layers when using 8-bit originals in Photoshop, since all curves are effectively applied at once instead of in succession.

Originally Posted by Dave Humphries

I think most of the above shows my ignorance as much as anything, but it may help you understand where misconceptions arise. (well, mine anyway )

All very helpful, thank you! As colin said in his post, it can be a hellishly complex topic

[McQ]

Originally Posted by Colin Southern

Overall I think it's really good - it's a hellishly complex area that I think you've explained really well; I'll bet it took more than 10 minutes to write that one!

Just a wee bit... Luckily a lot of images from previous explanations were still applicable to this concept...

[Colin Southern]

Originally Posted by McQ

Just a wee bit... Luckily a lot of images from previous explanations were still applicable to this concept...

I guess that on the plus side though, all the research + thinking + documenting that goes into the tutorials must really help to solidify all of these areas into your head (What I'll call the "walking / talking encyclopedia of photography" syndrome).

Just don't enjoy it too much or you won't want to do your "day job" anymore!

[McQ]

Originally Posted by Dave Humphries

Speaking personally, one thing I have often wondered about is the standard gamma 2.2 thing; I knew of this from my TV production days and how that gamma is applied in the TV camera. However, I am still unsure if it is applied in the digital camera, or during jpg creation (camera or RAW processing), or not at all (but just in the display), and does this explain why I see more in the blacks on screen than when printed - or is this just my very badly mis-profiled CRT at work? I don't print very often at all, so it's not a huge issue for me, and if something comes out dark, I tweak it and print again (I can almost feel Colin cringing!)

OK, I said I'd get back to you on this one. It's a very confusing topic so I wanted to make sure and get my terminology correct.

The short answer: gamma is applied at both steps -- to the file, and by your video card.

However, Gamma can actually be applied at any of three stages: on the file data, in the video card, or in the monitor itself. The net result of the last two gammas is called "display gamma." This is usually what people are talking about when they use the term gamma. We'll therefore assume that there's only the following two steps:

(1) GAMMA ENCODING: A RAW file has a linear Gamma (1.0), but when it is "developed," this file has gamma applied to the data itself (usually 2.2). The converted TIFF or JPEG therefore already has a native gamma applied to it. This makes sense because it means that all of the bits describing tones are distributed in a way that is perceptually uniform (ie, not all the tonal precision is bunched up in the highlights, with little left for the shadows).

(2) GAMMA DECODING: However, this is not the end. When this file needs to be displayed on your monitor, it has to have a gamma correction applied to it. Otherwise the image would appear way too dark (gamma correction=1.0). However, this gamma is the opposite of the gamma that had been applied on the file. The problem is that most places use the values for gamma and gamma correction interchangeably.