Preference question: sRGB or Adobe RGB (1998)?

[ho_co]

Advertisement (gone after registration)

Thanks for a lot of helpful information--the subject is far deeper than I imagined.

 

But this raises one additional question:

 

Since JPG allows by definition only 3 * 2^8 colors, I don't see how it can offer a larger or a smaller color space.

 

On the other hand, most dSLRs offer 3 * 2^12 colors in RAW, while DMR and M8 offer 3 * 2^16 colors in RAW. Both cases are a far larger gamut than offered in JPG.

 

So shouldn't the choice of color space be reflected in the RAW output?

 

I've got a lot of homework to do on this topic!

 

--HC

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[pascal_meheut]

Since JPG allows by definition only 3 * 2^8 colors, I don't see how it can offer a larger or a smaller color space.

--HC

 

You are making a confusion between the gamut and the bit depth. They are not directly correlated.

Let's say the gamut is a map of the colors. Some camera can map colors over a larger space.

Bit depth is how precise each coordinate is. You can map a huge color space with only 8 bit depth but of course, your grid will be loose.

And you can map a small gamut with 16 bits: then you may have extra precision, being able to separate even more colors than the human eye.

[Rolo]

Forgive me if I've missed this question above.

 

I'm finding that commercial printers here are using sRGB only and are converting 1998 images to that colour space. Was chatting last week to a photographer who had done some searching and he said he'd only found 2 in the UK that use the Adobe 1998 space. That may be true, or not, but the two labs I use are sRGB.

 

As a consequence, I changed my Photoshop Preference to sRGB as I only print colour via labs. Was that the wrong thing to do?

 

Secondly, I'm sure that the web only shows sRGB and that 1998 is converted to that space and it's best to convert web images to sRGB to be sure of what the viewer see. That contradicts the post above that states that you can't see the difference in colour space on a monitor. can someone confirm this.

 

Thirdly, a poster showed some fab colour space diagrams of the M8 showing how much better the M8 space was to everything else. Great diagrams, fantastic colour space, BUT will the benefit of those graphs get to the print, or is it limited by sRGB, 1998 or some other bottleneck. My car has a speed limiter set at 154 mph, but I never go there. Is it the same with colour space?

 

Regards

 

Rolo

[pascal_meheut]

I'm finding that commercial printers here are using sRGB only and are converting 1998 images to that colour space. Was chatting last week to a photographer who had done some searching and he said he'd only found 2 in the UK that use the Adobe 1998 space. That may be true, or not, but the two labs I use are sRGB.

 

As a consequence, I changed my Photoshop Preference to sRGB as I only print colour via labs. Was that the wrong thing to do?

 

No, you were right. Of course, you can prefer to stay in AdobeRGB for archiving purpose, should you one day have access to a better lab/printer and convert just before sending to a sRGB lab.

 

Secondly, I'm sure that the web only shows sRGB and that 1998 is converted to that space and it's best to convert web images to sRGB to be sure of what the viewer see. That contradicts the post above that states that you can't see the difference in colour space on a monitor. can someone confirm this.

 

The web does not show sRGB, it shows the monitor native space because the browsers do not use color management. As most of the monitors are very close to sRGB however, it is common use to publish pictures on the web in this color space.

 

Thirdly, a poster showed some fab colour space diagrams of the M8 showing how much better the M8 space was to everything else. Great diagrams, fantastic colour space, BUT will the benefit of those graphs get to the print, or is it limited by sRGB, 1998 or some other bottleneck. My car has a speed limiter set at 154 mph, but I never go there. Is it the same with colour space?

 

You should try to drive faster sometime And yes, this is the same for color space as it is for resolution: most of us do not need 10 Mpix but we are glad to get it.

It is the same with color space for many reasons.

[ho_co]

You are making a confusion between the gamut and the bit depth. They are not directly correlated.

Let's say the gamut is a map of the colors. Some camera can map colors over a larger space.

Bit depth is how precise each coordinate is. You can map a huge color space with only 8 bit depth but of course, your grid will be loose.

And you can map a small gamut with 16 bits: then you may have extra precision, being able to separate even more colors than the human eye.

Pascal--

Excellent explanation! I think I'm beginning to find my legs (ready for tomorrow's lurch off-kilter ). Also your response to Rolo's excellent questions was quite helpful!

 

--HC

[jrgeoffrion]

...I'm finding that commercial printers here are using sRGB only and are converting 1998 images to that colour space. Was chatting last week to a photographer who had done some searching and he said he'd only found 2 in the UK that use the Adobe 1998 space. That may be true, or not, but the two labs I use are sRGB.

 

As a consequence, I changed my Photoshop Preference to sRGB as I only print colour via labs. Was that the wrong thing to do?

 

Secondly, I'm sure that the web only shows sRGB and that 1998 is converted to that space and it's best to convert web images to sRGB to be sure of what the viewer see. That contradicts the post above that states that you can't see the difference in colour space on a monitor. can someone confirm this.

 

Thirdly, a poster showed some fab colour space diagrams of the M8 showing how much better the M8 space was to everything else. Great diagrams, fantastic colour space, BUT will the benefit of those graphs get to the print, or is it limited by sRGB, 1998 or some other bottleneck. My car has a speed limiter set at 154 mph, but I never go there. Is it the same with colour space?

 

For tidbits, it should be pointed out that the sRGB color space was developed based on the spectral and colorimetric characteristics of HDTV phosphors -- which is very close to that of CRT (predominant when sRGB was created).

 

With regards to whether or not one will see benefits from a larger gamut in prints, the answer is: 99% of the users will most likely not benefit from a larger color space. Why? Because the gamut of the output device will most likely be smaller than that of 'bigger M8 space'. For simplicity reasons, unfortunately, most labs process images using sRGB as their input space (the law of the lowest common denominator). Hence, everything larger than sRGB ends up being re-mapped to the color space. Those that print with their own professionally calibrated printer, using a color management workflow, may see a difference.

[Rolo]

Advertisement (gone after registration)

Good answer pascal. many thanks.

 

JR " Hence, everything larger than sRGB ends up being re-mapped to the color space."

 

And presumeably that means your not in control (ish) of what the Lab produces if you don't supply sRGB to them.

 

I use a Gretag One-Eye to calibrate my screen colours (must check if it is with 1998 or sRGB !!) and then generally import the lab scans with a proprietary space and was changing it to 1998 when opening, but since last week set the Preference to sRGB.

 

Sometimes the Lab print output has a slight green cast, whereas the screen didn't. Could this be the 1998 to sRGB converson the Lab was having to do ?

 

Am I best using sRGB throughout my system? I'm working on an Apple G4 with a 23" Cinema screen.

 

Larger part of my input (500 images a week) is from Lab scanned 35mm Fuji Superia and Kodak Portra plus a little CN B&W. Recent Labs scans from Superia have been quite unnaceptable, but they are changing machine next week, back to a Noritsu, and I can't wait.

 

A further complication is that I have a Nikon 8000 MF scanner that I use for larger prints or MF, and that is set to Nikon Colour space. What should that be set to?

[albertwang]

Ah, no wonder! I've been posting my Abode RGB images on the web without conversion to sRGB... hmm, that may explain a few of the problems I've been having..

[jrgeoffrion]

Ah, no wonder! I've been posting my Abode RGB images on the web without conversion to sRGB... hmm, that may explain a few of the problems I've been having..

 

Yes, that is probably your issue. In addition, some web browser are ICC profile, which explains why some people see different things.

[jrgeoffrion]

And presumeably that means your not in control (ish) of what the Lab produces if you don't supply sRGB to them,

 

Yes and no. What it means, is that, if you supply an sRGB file to your lab, they won't [really] be able to do 'better' in terms of colors. On the other hand, if you do not supply them with sRGB color AND they don't notice it and converts to their lab profile, you will get a certain color cast and color shift to your prints.

 

Sometimes the Lab print output has a slight green cast, whereas the screen didn't. Could this be the 1998 to sRGB converson the Lab was having to do ?

 

I can't answer this without knowing exactly what you did. BUT, if the print your sent was a black and white file in sRGB (which means that the R, G, and B values where all equal in your image), then the problem is that your lab printer profile was not perfect. As a matter of fact, I know of very few labs whose profiles are accurate. Some labs offer 'true black-and-white' digital paper. I recommend you look into this. Since the paper is black-and-white, there will be no color cast.

 

Am I best using sRGB throughout my system? I'm working on an Apple G4 with a 23" Cinema screen.

 

Larger part of my input (500 images a week) is from Lab scanned 35mm Fuji Superia and Kodak Portra plus a little CN B&W. Recent Labs scans from Superia have been quite unnaceptable, but they are changing machine next week, back to a Noritsu, and I can't wait.

 

A further complication is that I have a Nikon 8000 MF scanner that I use for larger prints or MF, and that is set to Nikon Colour space. What should that be set to?

 

If your lab accepts only sRGB images, then, you would be best to keep everything in sRGB. HOWEVER, if ever you found a lab that would accept a larger color space and you would like to take advantage of it, you would need to re-scan the files to a larger space. It really depends on your output and what you are doing and planning to do with them in the future.

[augenschmaus]

Yes, that is probably your issue. In addition, some web browser are ICC profile, which explains why some people see different things.

As far as I know, only Apple's Safari browser natively supports color management, i.e. it recognizes and processes attached profiles. The same is true for MS Internet Explorer and OmniWeb for Mac, if the browser preferences are set accordingly. There are currently no PC Windows browsers supporting color management.

[wparsonsgisnet]

Ah, no wonder! I've been posting my Abode RGB images on the web without conversion to sRGB... hmm, that may explain a few of the problems I've been having..

 

OK, when I send pictures to the web (or to others who may view them or attach them to emails), I use PSCS2's conversion "For Web." Are those the images converted to sRGB?

 

Thanks.

[ho_co]

ECI-RGB is from the European Color Initiative, and its members are listed here:

ECI - ECI |::| Members

Jonathan--thanks!

Looks as if producing in ECI would simply reduce the number of steps needed for publication through a certain set of European publishers.

 

There is probably also such a standard in the US that I simply haven't heard of. Good to see Leica reach out this way, as they did with DNG.

 

I appreciate the link!

 

--HC

[augenschmaus]

OK, when I send pictures to the web (or to others who may view them or attach them to emails), I use PSCS2's conversion "For Web." Are those the images converted to sRGB?

 

Thanks.

 

No, the "Save for web" dialogue doesn't adjust the profile automatically - before exporting the file, you have to go to Image -> Mode -> Convert to Profile and choose sRGB.