Advertisement from Google AdSense

QUESTION: Which would give the best colors? Shooting aRGB or sRGB? And what bit depth should I use in After Effects?

First off, I'd like to mention that there is no such thing as aRGB. There is sRGB, and there is AdobeRGB, but Adobe RGB should never be written as aRGB.

Among other things there is obvious potential confusion (not to mention typos) relative to sRGB, not even mentioning that "Adobe" is a proper brand name and should be capitalized. 

"ARGB" is ALSO wrong, as it could lead to confusion with RGBA (red/green/blue/alpha), but WORSE, Windows uses ARGB as the method for storing colors in windows (alpha/red/green/blue), but Windows uses sRGB as the target COLORSPACE. ARGB in Windows is NOT a color space.

So you see, when you use aRGB, you are being ambiguous and will create confusion.

Adobe RGB can be referred to as "Adobe 98" or "Adobe RGB". I suppose if you really must desperately conserve characters because your fingers are too tired to type out all 9 characters, you could type "AdbRGB". But never aRGB. (/End of my rant.)

IMAGE CAPTURE in THE smallest space with no clipping:

If you are shooting/capturing in 8 bit (such as most dSLRs) then you want to use the smallest colorspace where you have no clipping. If you are shooting ultra-saturated bright colors, then maybe Adobe RGB would be a useful space - but for most subjects, you'll get the best image fidelity shooting in sRGB - this is because the primaries in sRGB are closer together, so there are fewer delta E errors. A delta E error means there is not a data point (due to the limited data size of 8 bit) to correctly identify a particular color.

Larger gamuts/color spaces should be avoided unless you are using a data size greater than 8 bit. 10 bit, 12 bit, or 16 bit for instance usually work fine with the larger colorspaces like P3 and Adobe RGB. Even so, the super big spaces like ProPhoto and ACES really need 16 bits.

The point is, if your subject needs a wider gamut/larger colorspace, then you want to be shooting with a camera that supports that both in profile and bit depth. If you are using an 8 bit camera then in nearly all cases you want to stick with Rec709 (or sRGB for dSLRs), and then work to get your lighting to fit within that 8 bit space.


Bit Depth in After Effects 

I'm not sure what you are going to be using AE for specifically, but for compositing and most effects I prefer and recommend working in 32 bit mode with a linearized color space. When you do, additive operations act like light in the real world, and come across as more natural. 

When you are using 32 bit mode in AE, you can use pretty much any colorspace you want to as a working space. 8 bit mode is never recommended, unless you are trying to achieve 8 bit artifacts on purpose.

More on camera and export colorspaces:

With dSLRs you are going to be creating 8 bit compressed camera originals. 8 bits means that you will have 256 levels of red, 256 levels of green and 256 levels of blue. This is regardless of what profile you use - Adobe RGB or sRGB.

The difference between sRGB and Adobe RGB is the *distance between colors*. If you increase the distance between colors, you are actually increasing the potential errors (i.e. delta E errors). Since Adobe RGB is larger than sRGB, the distance between colors is greater, and you get greater delta E errors as a result.

The min reason for using a larger colorspace is when the smaller colorspace is clipping colors. But when you are in 8 Bit, the larger color space will as a consequence be more prone to *banding*.

The standard for HDTV is Rec709, and Rec 709 and sRGB share the same primaries, and thus have very close to the same gamut size (the difference between Rec709 and sRGB is in the gamma curve).

If your final output is for HD, at Rec709, then you will probably not derive any real benefit from shooting in a larger space like AdobeRGB - in fact you may find additional *problems* due to delta E types of errors, and lower image fidelity in your final output.

Instead of using AdobeRGB to shoot, consider using sRGB, and set the contrast LOW and also possibly reduce the saturation slightly so that you get no clipping of colors, but then you will have an image that can be "expanded" back to a higher contrast one, retaining more detail in blacks, and with less clipping on highlights and saturated colors. (Be careful of going too far with this idea, as you can introduce more noise into the image if not careful. As always, do tests!!)

Obviously, tests should be done with your intended subject matter to find the right workflow along these lines. Note that if you are going out to actual FILM, or DCI, then Adobe RGB *may* be a better choice (but not necessarily).

As a "rule", use the smallest colorspace that does not result in clipping, i.e. "out of gamut" errors. If you do not get out of gamut errors (clipping) in sRGB, then you will gain no benefit from using Adobe RGB.

LoCon Profiles for dSLRs:

ALSO: if you are using a Canon dSLR, go to the Technicolor website and look at the CineStyle picture profile: Technicolor CineStyle Plugin Download

If you use Nikon, you can download Cineflat by Photographio.

If you are shooting video and want to end up with a final result that looks more like a feature film, then these profiles (Cinestyle or Cineflat) are a good start.

This article was originally written circa 2012, but was revised July 2018

Advertisement from Google AdSense

Help | Legal | Contact Us

Copyright © 2012-2018. All Rights Reserved.