2.2 gamma

[Megalith]

I recently tuned my display to a gamma of 2.2 but can't shake the impression that the black levels are higher than they should be. I am pretty sure that the calibration/profiling is correct, so maybe I'm just too used to seeing crushed blacks. I also watch a lot of movies, which are mastered at 2.4, so that might have something to do with it.

Anyone else feel the same after calibration?

 

[Panmaster]

2.4 gamma is better for anything made after 2005ish?

 

[spacediver]

the situation is unfortunately far more untidy than many think. There was no explicit standard for the luminance function (of which gamma is a form) up until 3 years ago with the introduction of BT.1886. The idea was that most CRTs had a natural gamma of around 2.4, however this could vary from unit to unit. When other display technologies took over in the studios, many of them were calibrated to a gamma of around 2.2.

Even nowadays, some studios use 2.2 although there is a big shift starting to happen where people are adopting BT.1886, which has an exponent of 2.4.

One should also keep in mind that the perceptual appropriateness of any given luminance function interacts with viewing conditions. With brighter conditions, material that was mastered at 2.4 in a very dimly lit environment will need to be viewed at a lower gamma to create a perceptual match.

My advice is to calibrate, if possible, to BT.1886 and also have a preset for 2.2 for different viewing conditions/differently encoded material.

 

[Sycraft]

I dunno I've become accustomed to an ever "brighter" gamma, that being L*. It is more perceptually uniform which ends up meaning a bit higher levels in the dark colours. I love the way it looks. Shadow detail is clear, brights aren't blasted out, etc. It is by far how I prefer to work and game.

It can depend on what you like, particularly with computers where to the extent there are any standards, they are not well followed at all. If you like things darker/brighter it can be a lot of personal preference.

It can also depend on your gear. Some monitors are fiddly with their dark steps and correction can make banding and such apparent on them. Also some monitors (older VA monitors like some of the original BenQ VAs) need a pretty high brightness level before they stop crushing the blacks and trying to calibrate around that at lower brightness doesn't help.

You can fix the problem with high end monitors with hardware calibratable LUTs, of course, but then those cost a ton and most people aren't interested in spending that much.

So if this is just for you, then do what makes you happy. I'd say give the new gamma a bit of time and see if your opinion changes, but if not, then go with whatever looks best. After all what matters is that you are happy with the way it looks. So long as it does well for you for surfing/gaming then you are good to go. You should be able to tell the calibration program to use a different gamma target, if you like. There is no "right" way to game, it is all about what looks/feels good to YOU. You make yourself happy, don't worry about anything else.

If this is for something professional, like video editing, then stick with whatever standard curve is specified for that work. If you work in a profession where there's standards, best to follow them

 

[spacediver]

I dunno I've become accustomed to an ever "brighter" gamma, that being L*. It is more perceptually uniform which ends up meaning a bit higher levels in the dark colours.

L* is a function that approximates our incremental sensitivity to changes in luminance. The exponent that best describes this function is about 0.42. The inverse of L* is what should be targeted if you're going for perceptual uniformity, and it just so happens that 2.4 is pretty damn close to a perceptually uniform function, when you're dealing with deep blacks and a peak white around 100 nits. 2.4 has great depth and contrast if done correctly, however many users calibrate incorrectly when dealing with raised black levels (see this thread).

It can depend on what you like, particularly with computers where to the extent there are any standards, they are not well followed at all. If you like things darker/brighter it can be a lot of personal preference.

Agreed, but see point at bottom of post.

You can fix the problem with high end monitors with hardware calibratable LUTs, of course, but then those cost a ton and most people aren't interested in spending that much.

A decent display (i.e. not 6 bit + 2 frc) should be able to handle LUT adjustments without quantization artifacts, and with software like Argyll you can create incredibly accurate LUTs for free.

If this is for something professional, like video editing, then stick with whatever standard curve is specified for that work. If you work in a profession where there's standards, best to follow them

The whole point of encoding standards is that consumers can enjoy material with artistic integrity. This requires the viewing conditions to match, as closely as possible, those in the mastering studio. When it comes to modern blu-ray material, for example, that has been mastered at BT.1886, then viewing that material at home with an inappropriate luminance function will result in a washed out or crushed image. But I agree that for web surfing and gaming this is hardly an issue.

 

[Bluesun311]

Was some info in this thread I needed to know about.

My calibrations are getting better and better as I learn, and it's a slow process because I do take my time to "get used" to each new setup I try.

But I still come back to being pretty mystified by Windows and the best way to manage calibration settings. Currently I'm letting DispCalGUI handle the profile I made with an ID3 pro, I never did figure out how to get a .LUT file as was suggested elsewhere and use the command line to load it. Never found a .LUT file at all, just .icc/.icm and .cals.

 

[spacediver]

I'm not familiar with dispcalgui, but I am somewhat familiar with the underlying "backend", which is argyll. Dispcal.exe is one of the functions of argyll and it creates .cal files which are basically 1d luts, and can be loaded with dispwin. I'm pretty sure that if you use dispcalgui you can use that software to load the .cal file it creates.

 

[Megalith]

So what basically happened was that I assumed sRGB gamma was equivalent to 2.2, but this is not the case. I just finished re-calibrating and creating a new profile, and the blacks are definitely deeper this time around.

 

[spacediver]

Sorry, I misspoke earlier - when I said that up no luminance function was specified until BT.1886, I was referring to video standards.

sRGB gamma actually is 2.2, but that standard was created with the assumption that many people would be using displays in brightly lit environments. As such, many find that using 2.4 in a dimly lit environment is appropriate for sRGB encoded material.

 

[XoR]

for web, games and movies at night you should use normal power 2.4
most monitors have 2.2 but not sRGB. Actually sRGB is not very good for content consumption and is kinda artificial standard that doesn't look very good with most content and most monitors doesn't use anything like it.

2.3 is also good choice which you might like. I actually use 2.3 when calibrating my monitors

 

[spacediver]

XoR, have you tried out BT.1886?

 

[Megalith]

I read that the sRGB gamma curve is similar but not identical to the standard 2.2 gamma curve. The sRGB gamma curve supposedly verges away from 2.2 as it gets closer to RGB 0, which would explain the elevated blacks.

 

[spacediver]

sRGB isn't very well specified in the encoding standards, but in terms of decoding, it's a pure 2.2 function with a veiling glare term, which I believe allows the function to rise perceptually smoothly with displays that have raised black levels (although I'm not sure about this). I'd just stick with BT.1886 - with argyll you can use the BT.1886 type function with any exponent, if you're dead set on 2.2