SRGB vs adobe

[TreeofLifeStairs]

Do you guys switch between color spaces depending on what you're shooting or just stick with one? If you do switch, why? If you stick with one, which one?

 

[Ysarex]

Shoot raw so it's not a concern.

Otherwise recommend you shoot sRGB if you're creating camera JPEGs.

Joe

 

[unpopular]

^yep, at least until it's time to edit! then Adobe or some other wide gamut.

 

[KmH]

Yep. Raw has no color space so it's only a TIFF or JPEG concern.

The web is sRGB and not all print labs can print from Adobe RGB files.

I edit in the ProPhoto RGB color space and at output convert to a color space appropriate for how a photo will be used.

COLOR MANAGEMENT & PRINTING

Concepts

• A Background on Color Perception
• Understanding Digital Pixels: PPI, Dithering & Print Size
  Understanding the Process
• Color Management, Part 1: Concept & Overview
• Color Management, Part 2: Color Spaces
• Color Management, Part 3: Color Space Conversion
• Understanding Gamma Correction & Digital Tones
  In Practice
• How to Calibrate Your Monitor for Photography
• Soft Proofing: Matching On-Screen Photos with Prints
• Working Space Comparison: sRGB vs. Adobe RGB 1998

 

[TreeofLifeStairs]

I've read all those links before but was unaware that RAW files did not use a color space. I would imagine that even though a raw file may not use a color space like sRGB it still would have the color space limitation of the sensor, correct?

 

[Ysarex]

I've read all those links before but was unaware that RAW files did not use a color space. I would imagine that even though a raw file may not use a color space like sRGB it still would have the color space limitation of the sensor, correct?

Digital camera sensors don't technically record color; they only record tone (light/dark) variations. In front of a digital camera sensor is an RGB color filter array most commonly in a four square pattern called a Bayer array. Each photosite in the sensor has a filter then in front of it.

The color in a digital photo (the color value for each pixel) has to be interpolated from the Bayer filter array. Some variation exists between camera makers in how that filter array is laid out for example Sony has experimented with an equal number of RGB filters as opposed to the traditional 1 red, 1 blue, 2 green and Fuji uses two repeating 6 square patterns instead of the traditional 4. Also some variation exists in the quality and material used for the RGB filters, but these variations aren't major. They are all basically RGB (three color) devices with pretty similar characteristics.

Adding a fourth color filter (Sony tried that once) would be a more substantive change (see related: Quattron - Wikipedia, the free encyclopedia).

A raw file is the sensor/filter array data before the RGB values are interpolated. You have the inherent physical variations between materials used by the different manufacturers but no digital color space at that point exists. The interpolation process (called demosaicing of the array) includes the task of assigning color values relative to an RGB color space. In this case, as Unpop noted, interpolation to a wide gamut (typically Pro-photo) color space is most appropriate.

There's substantial variation in the implementation of the interpolation (demosaicing) algorithms used between the competing camera and software companies and this can result in noticeable differences in the final results. In other words Adobe LightRoom "color" can be different than Apple Aperture "color" both of which can be different than Nikon "color" all from the same camera raw file original. In this case we're seeing variation in software interpretation of the same physical source.

Joe

 

[unpopular]

Moreover, Fuji had sensors with extra pixels for high and dark ranges as well as color, and in a sense the traditional sensors (including x-trans) use a send green filter to help calculate tonality as well.

I have for a while proposed a theoretical sensor that instead of filtering the light it instead diffracts it along a specified area of the sensor. For photography you wouldn't need a relatively large area, to obtain highly detailed information about color. A ten pixel wide area would record the color information of ten filters, but not be limited to device space.

of course, spectroscopic images are made in forensic science and astronomy, but usually this combining multiple images taken in succession or arranged onto the sensor using special optics.

 

[KmH]

The interpolation process (called demosaicing of the array) includes the task of assigning color values relative to an RGB color space.

Just to be clear.
RGB is a color model (additive) that has an entire family of color spaces associated with it.
sRGB, Adobe RGB, and ProPhoto RGB are just 3 color spaces in that RGB color model family of color spaces.

RGB color model - Wikipedia, the free encyclopedia
Color space - Wikipedia, the free encyclopedia

While electronic displays use the RGB color model and use red, green, and blue lights, prints are made using the CMYK (subtractive) color model.
That's why inkjet printers have Cyan, Magenta, Yellow, and Black ink cartridges.

CMYK color model - Wikipedia, the free encyclopedia

Color science is pretty complex and worth studying if you are serious about your photography. See the External links on each of the Wikipedia links above.

 

[Browncoat]

General rule of thumb regarding color space and prints:

95%+ of all printers are set up for sRGB. A lot fewer are capable of Adobe. If you have the software, are doing prints yourself, and can manage your workflow, then Adobe might be a better choice for you.

 

[sonicbuffalo]

Adobe has a broader spectrum of colors, but as others have said, most of todays printers utilize sRGB.

 

[TreeofLifeStairs]

I think I've always saved as Adobe and had problems with getting prints done and been unimpressed with how the colors came out.

 

[unpopular]

I think that there is a TON is misunderstanding going on here.

The best way to edit a photograph for print is to soft proof on a calibrated monitor using the output device's profile but keeping the wide gamut profile. Once finished convert to the output profile (or whatever profile the print shop requests) on the print-ready artwork.

You want to edit in a larger gamut space so that you have the best second choice available to do you edits in on any out of gamut color.

 

[CameraClicker]

Do you guys switch between color spaces depending on what you're shooting or just stick with one? If you do switch, why? If you stick with one, which one?

I start with a raw file. The converter output is ProPhoto. If it is printed on my printer, it is left as ProPhoto. If it is for the web, it is converted to sRGB. If it is being printed at a lab, I ask the lab what they want. If the lab can't tell me, I shop for a different lab.

Work from a large space to a smaller space because as you convert to smaller spaces, values are squished (a technical term ) to fit the smaller space. Moving from a smaller space back to the larger space, there is no reliable way to un-squish, to get back to the original image.