Does the selection of sRGB or Adobe RGB in camera when shooting RAW ever matter?

Question

In talking to a colleague of mine yesterday, we got onto the subject of color spaces in digital photography. He's a Nikon buff, I use Canon (specifically, an EOS 50D); I don't know if that is relevant to my question or not, but mention it for completeness.

I shoot RAW essentially 100% of the time, and have the in-camera color space setting at AdobeRGB (sRGB is also offered). The raw converter in turn is set to produce sRGB output files. As I understand the process, the general flow is for raw shooting: raw sensor data readout (including A/D conversion), compression, write to a raw file, then in post-processing rather than in-camera apply color space/color temperature/etc; for JPEG shooting: raw sensor data readout, apply color space/color temperature/etc, JPEG compression, write to memory card (in camera).

Hence, when he said that I should leave the camera set to sRGB for best/consistent results despite using the camera's raw format, it didn't really make much sense to me. I have mulled it over for a while now and it still doesn't really make any sense.

One of the answers to What is RAW, technically? seems to support my view of the differences in the data recorded in RAW and JPEG modes, respectively.

I did find What's the difference between Adobe RGB and sRGB and which should I set in my camera?, but the answers there seem to focus more on the difference between AdobeRGB and sRGB. My question is, when shooting RAW, does it ever matter (except for the image preview) whether the camera is set to sRGB or AdobeRGB? If it does, then why, and in what situations? Is my understanding of the "raw" format processing chain flawed?

Asked by Michael Kjörling

Answer

The answer is basically the same as this one on white balance and raw. You're right, it doesn't matter at all in how the image is recorded or stored.

As you note, the selected color space applies to the preview image and to the histogram. The camera also may make metering decisions intended to avoid clipping (overexposing to the full saturation point) in particular channels, and that may be affected by choice of color space in-camera. I think this is less likely to be a real problem than having the white balance far off might be, though.

But I'd also take a careful look at the preview image in both modes. The rear LCD screen isn't something you can calibrate on most cameras. I think it's more likely that sRGB is correctly represented, and the screen might not cover the Adobe RGB gamut very well. (And really, I'd not be surprised if many cameras don't really handle preview display of Adobe RGB correctly at all — that's a lot more complicated than simply having the mechanism to save in that space.) I'd do some test shots and see which looks most like what your final printed output does, and then use that — that'll make the review screen more useful. It may very well turn out that an accurate display of sRGB is more like your final results than a bad rendition of Adobe RGB (even if you ultimately do use Adobe RGB in the end).

Answered by mattdm

Check more discussion of this question.

Should I perform the final color space conversion before or after final sharpening of an image?

Question

When preparing an image for the web (or other media), the last two steps is usually to resize the image to the desired resolution and sharpen it at that resolution. However, when working in a larger/different color space than the final result is going to be exported in, when should I convert to the target color space relative to these two operations?

Answer

My last steps, in order, are resize, sharpen, convert color space from ProRGB to sRGB, convert to 8-bit.

The only really critical orders are sharpening after resizing and changing to 8-bit after changing the color space.

In fact, I was curious, so I just tried. One copy, I did my above steps. On the other copy, I converted color space, converted to 8-bit and then resized and sharpend. I put the images over each other and changed the blending mode to difference.

There was absolutely no pixel level difference between the two. I expect to see a few minor differences at 100% that would never show up on the print, but in the case of this one particular photo, there were absolutely no differences at all.

Check more discussion of this question.

How much of a difference do different color spaces make?

Question

How much of a difference does it make to use one of the extended colorspaces vs. the normal sRGB? And what difference does it make?

Are there situations where it is not important, and conversely are there situations where using an extended colorspace is essential?

Answer

Color spaces, as ysap stated, can be a confusing issue. There isn't a single correct answer to this question, and what you intend to do with the "final copies" of your images will really dictate what color spaces you use and when you convert from one to the other.

While I think it is getting a bit dated, sRGB is still the "safest" color space these days. Many professional printing services require sRGB for print (although that is changing these days, as many professional printing services now support AdobeRGB or even ProPhoto RGB). If you are publishing your images to the web, sRGB is also the safest color space to convert your final images to, as many web browsers do not support proper ICM, and default to rendering in the sRGB space.

What color space you use during your processing workflow, on the other hand, is a more complex story. First off, camera's see far more into the realm of green color than computers or printers can generally render. If you work in RAW, and want to preserve as much original color accuracy as possible throughout your workflow, it is best to keep your images in the widest possible gamut. If you are working with muted scenes, or scenes with limited color, particularly lower saturation, a narrower gamut will be ideal. Use of a wide gamut when processing RAW is usually done for you if you use a tool like Lightroom or Aperture, as those tools make the default assumption that RAW images will utilize a considerable amount of the ProPhoto RGB gamut...the only gamut that covers nearly the entire LAB space that represents human color perception. Converting from RAW to TIFF will, then by default, save TIFF images with ProPhoto RGB unless you choose otherwise. (When working directly in RAW, gamut is really not taken into account, as RAW images are generally not tagged with a gamut at all.)

Modern professional photography hardware, both computer screens and printers, have all moved towards AdobeRGB as the base or reference gamut. Top of the line, and even middle grade, LCD screens like the Apple CinemaDisplay, NEC, Eizo ColorEdge, or LaCie 700 RGB-LED series monitors all support 98-123% of the AdobeRGB gamut. Modern printers from Epson and Canon, particularly the prosumer models, but now also the professional/commercial lines, also support 98% or more of the AdobeRGB gamut. Epson printers tend to cover more of the blues and violets beyond the realm of sRGB and AdobeRGB, while Canon covers more of the reds and greens beyond the realm of sRGB and AdobeRGB. Therefor, if your intention is to process your photos on a professional grade wide-gamut LCD screen and print your photos on a prosumer Epson or Canon printer, the AdobeRGB gamut is the ideal choice for your final images.

Converting from a wide gamut to a smaller gamut can be a critical step in your workflow. There are a variety of rendering intents when converting between color spaces. The two most common are Relative Colorimetric intent, which aims to preserve original color value accuracy at the cost of perceptual accuracy, and Perceptual intent, which aims to preserve perceptual accuracy at the cost of original color value accuracy. Converting from one gamut to another should be done as infrequently as possible. Ideally, you would work in RAW until the moment you generate a "final copy" image for a specific medium, at which time you would convert that final copy to the appropriate gamut. If you are saving for web, the best gamut is sRGB. If you are saving for print, I would choose AdobeRGB unless you are sending off to a print lab for print, and they require sRGB.

When converting, it is important that you have your white and black points, curves, and contrast set correctly for. If you are converting to sRGB for display on the web, white and black points are far less important. For print, it is best to try soft proofing your image and tweak white and black points first before converting to a final gamut. You should convert and compare the original image to the converted image, and make sure you have not lost vibrancy or saturation in key areas. When converting from a wider gamut to a narrower gamut, the key areas where you may lose color will be in the greens and deep blues. If the difference between a source gamut and destination gamut color is broad, especially if that particular color is found in gradients, you may encounter some posterization or clipping. In such cases, you may want to adjust saturation, white and black points to reduce the range of color you are utilizing in your original image. This will lessen the amount of compression of color into the destination color space, and mitigate or eliminate posterization and clipping.

For a few general rules of thumb:

• Cameras "see" a much wider gamut with their sensor than computer screens or printers can render. Particularly greens.
• If you shoot vibrant, saturated scenes, a wider gamut will preserve colors better.
• If you shoot muted or desaturated scenes, a narrower gamut will preserve colors better.
• Most professional photography computer screens and printers support the AdobeRGB gamut.
  • Computer screens cover 98% - 123% or so of AdobeRGB
  • Printers cover around 98% or so of AdobeRGB, with different brands covering extended saturation into blues, violets, oranges, reds, and greens.
• Saturated green is one of the primary colors lost when converting to a narrow gamut, with blue and red following.
• The key difference between wide and narrow gamuts is their chromaticity.
  • Chromaticity generally refers to the hue and overall saturation of a color...its purity
  • Wider gamuts reach greater chromaticity
  • Narrower gamuts reach lesser chromaticity

Check more discussion of this question.

What's the difference between Adobe RGB and sRGB and which should I set in my camera?

Question

In my Canon 60D, there's a setting between Adobe RGB and sRGB. What's the difference and what should I prefer when shooting to RAW ?

Answer

sRGB is the most common color-space used anywhere.

AdobeRGB is a wider color-space which can represent more colors but with less precision when looking at the colors which overlap sRGB.

Neither color-space really matters when shooting RAW.

The embedded thumbnail or preview within a RAW file may be affected by the choice of color-space though, so keeping sRGB selected is usually the most sensible thing to do.

Check more discussion of this question.

What's the difference between Adobe RGB and sRGB and which should I set in my camera?

Question

In my Canon 60D, there's a setting between Adobe RGB and sRGB. What's the difference and what should I prefer when shooting to RAW ?

Answer

sRGB is the most common color-space used anywhere.

AdobeRGB is a wider color-space which can represent more colors but with less precision when looking at the colors which overlap sRGB.

Neither color-space really matters when shooting RAW.

The embedded thumbnail or preview within a RAW file may be affected by the choice of color-space though, so keeping sRGB selected is usually the most sensible thing to do.

Check more discussion of this question.