sRGB vs Adobe RGB

Your image has various numbers (RGB triples) assigned to each pixel. Those numbers are interpreted as specific colors by the profile that is attached to the image.

When you assign a profile to the image, you tell the image processor to leave the numbers unchanged, but interpret them according to whatever new profile you assigned. As a result, the color appearance in your monitor will change.

When you convert a profile, a new profile is attached, but, at the same time, the numbers are changed so that whatever color the pixel was when interpreted by the old profile will be (as closely as possible) the same color as now interpreted by the new profile. As a result, the color appearance in your monitor will stay (for all intents and purposes) unchanged.

--
gollywop

-----------

PhotoFredy said:

So if I tell the lab to use no corrections they still get
"processed"? How do get my prints to match my semi-calibrated LCD (I
used adobe gamma)? What would be the benefit of a truly calibrated
monitor if most labs and browsers aren't in synch, or are they. This
is a bit confusing so sorry for all the questions.

Click to expand...

The files will still get profile conversions to meet the needs of the output printer. Even the printer/processor itself has internal conversion tables that are updated with emulsion changes and every few hours just to offset any color shifts due to chemistry and processor temperature changes.

Your lab may offer lab corrected color and economy color (uncorrected) but both go through the same software and will be at least briefly seen by a lab technician. The difference is, the lab corrected color images will have a highly trained tech making color corrections based on a lab standard for skin tones and overall color/contrast/density.

Adobe Gamma is as close to 'color calibration' as a box of crayons is to oil pastels. Similar but not the same. Proper display calibration is well beyond what the human eye can do using simple software. And that is assuming your monitor is capable of accurate color and you have perfect color perception, even if Adobe Gamma was good for monitor calibration.

Even using a hardware calibration device such as a Spyder3 or similar, the display then becomes the weakest link. Even though the hardware and software combo can calibrate the display, it is making it as good as it can. Sometimes that isn't very good and colors will still be only partly accurate. Laptop displays are notorious for bad color while high end graphic displays can be calibrated to near perfect color. (If you don't mind spending $1000+ on a monitor.)

The monitors at the lab I work in are all identical and the calibration is checked and adjusted as needed on a weekly basis so that they all match. We use high end graphic monitors considerably better than what most people are using in their own home or studio. Although we are internally well calibrated from display to display and printer to printer, this still assumes that the files you send to the lab are of proper exposure and good color and have an appropriate and valid color profile applied.

--

There are no rules for good photographs, there are only good photographs. - Ansel Adams

My photography blog, Photoshop tips, color tips and more: http://modifiedphoto.wordpress.com

gollywop said:

Your image has various numbers (RGB triples) assigned to each pixel.
Those numbers are interpreted as specific colors by the profile that
is attached to the image.

When you assign a profile to the image, you tell the image processor
to leave the numbers unchanged, but interpret them according to
whatever new profile you assigned. As a result, the color appearance
in your monitor will change.

Click to expand...

So this is basically to preview the color interpretation of another profile? This means I could revert back since teh actual values didnt change, right? Why use this?

gollywop said:

When you convert a profile, a new profile is attached, but, at the
same time, the numbers are changed so that whatever color the pixel
was when interpreted by the old profile will be (as closely as
possible) the same color as now interpreted by the new profile. As a
result, the color appearance in your monitor will stay (for all
intents and purposes) unchanged.

Click to expand...

So there is no turning back on this one right? Once you convert, you convert.

gollywop said:

--
gollywop

-----------

Click to expand...

Thanks for your answers.

PhotoArtKC said:

PhotoFredy said:

So if I tell the lab to use no corrections they still get
"processed"? How do get my prints to match my semi-calibrated LCD (I
used adobe gamma)? What would be the benefit of a truly calibrated
monitor if most labs and browsers aren't in synch, or are they. This
is a bit confusing so sorry for all the questions.

Click to expand...

The files will still get profile conversions to meet the needs of the
output printer. Even the printer/processor itself has internal
conversion tables that are updated with emulsion changes and every

few hours just to offset any color shifts due to chemistry and...................................

Click to expand...

I use a Dell "ultrasharp" display (connected via DVI) and a Raedon card. Am I better off just sticking with sRGB to begin with or should I use aRGB and convert when finished? Is the converted sRGB image technically better than if it was recorded in sRGB to begin with. Any thoguhts? I may be making this harder than it is especially since I have a farily cheap LCD monitor. It might all be for nothing at this point.

PhotoFredy said:

gollywop said:

Your image has various numbers (RGB triples) assigned to each pixel.
Those numbers are interpreted as specific colors by the profile that
is attached to the image.

When you assign a profile to the image, you tell the image processor
to leave the numbers unchanged, but interpret them according to
whatever new profile you assigned. As a result, the color appearance
in your monitor will change.

Click to expand...

So this is basically to preview the color interpretation of another
profile? This means I could revert back since teh actual values
didnt change, right? Why use this?

Click to expand...

I've used this for a couple of purposes. Once, I knew that an untagged file actually had sRGB numbers, and I wanted to embed that profile, so I assigned it sRGB.

On other occasions, I've been given untagged files with unknown numbers. Of course, when you bring such a file into your image processor, it has to give it some color-space interpretation, so, typically, it assigns it your RGB working-space (this is definitely what PS does). The resulting image may look horrible, because that profile may not be appropriate to the numbers. In this case you can assign various profiles until the colors look right. Then you'll know what profile those numbers correspond to, and you can assign that profile.

You may want the image to have another profile, so, in the case above, you would first assign the profile that makes things look right, then you would convert it to the new profile. It would then still look right (since conversion doesn't change the looks), but would have a different profile (and different numbers corresponding to that profile.)

PhotoFredy said:

PhotoFredy said:

When you convert a profile, a new profile is attached, but, at the
same time, the numbers are changed so that whatever color the pixel
was when interpreted by the old profile will be (as closely as
possible) the same color as now interpreted by the new profile. As a
result, the color appearance in your monitor will stay (for all
intents and purposes) unchanged.

Click to expand...

So there is no turning back on this one right? Once you convert, you
convert.

Click to expand...

Well, you can always convert back. There could be a very slight loss that results because certain colors in one color space may not be exactly representable in another color space (particularly with 8-bit files) and because mathematical rounding and truncations errors take place as the computer does its calculations. But, for most images, you will not notice anything on this account.

Also, if you are working in PS (and likely with other image processors as well), and you still have the history for the processing you've done, you can go back to the point in history just before the conversion with no loss at all.

PhotoFredy said:

Thanks for your answers.

Click to expand...

You're welcome.

--
gollywop

-----------

Amy,

there is also another difficulty: most of the users don't have LCD monitors that have such a wide gamut to correctly display aRGB and, therefore, ProphotoRGB too.

The last word is always to the HW, in this case the monitor, and not the SW (i.e. the browser).

The color space has nothing to do with whether you edit it in PS or not.

AdobeRGB has a wider gamut. If you are shooting raw, use adobeRGB. Unless I need all the speed my D300s can give me (sports), I shoot 14-bit compressed NEFs and adobeRGB. This way the RAW file has as much of the image data collected by the sensor as possible.

Always convert to sRGB prior to posting online. IE and Firefox only support sRGB, so other color spaces will get rendered poorly. Safari, however, supports most common color spaces.

I would recommend sRGB is if one is shooting in-camera JPGs and expects to primarily post online with little or no post processing.

If your print shop only supports sRGB, find somewhere else to print.

--
-Dan Rode
http://rodephoto.com

All pro printers may well be able to use sRGB but any "pro" print shop can also use many other color spaces, including adobeRGB, ProPhoto, CYMK, even L.A.B..

The perfect color space would include all of the visible colors. This color space does not exist. I believe one called the "Wide Gamut" color space is capable of +70%. I think you even need special monitors for this color space.

sRGB can render 35% of all visible colors.
AdobeRGB can render over 50%
That's 40% more distinct colors

ProPhoto (used by Lightroom) is even higher than adobeRGB.

When you convert to sRGB, you lose some of the fine variations of color. The colors in adobeRGB that are not contained in sRGB (out of gamut) are converted to a similar color that sRGB can display.

Why would anyone want to throw away detail when printing?

LuvLatins said:

saw a video that said all pro printers take sRGB without the need to convert so that is what I use.

Click to expand...

--
-Dan Rode
http://rodephoto.com

It makes a big difference. Once sRGB is assigned, all of the colors available to adobeRGB are gone, never to return. You have thrown away fine color detail and cannot recover it.

a 14-bit NEF processed to 16-bit AdobeRAW file has as much color detail as my camera is capable of storing. I want all of the data, all fo the details, all of the fine gradations of color.

One may be happy with the results from sRGB. The web is full of great looking images displayed in sRGB. Just realize that there is a difference.

Robert Peters said:

doesn't matter whether you set the camera to sRGB or AdobeRGB other than the appearance on the lcd screen and the various histogram displays.

--

Click to expand...

--
-Dan Rode
http://rodephoto.com

Try reading http://www.luminous-landscape.com/tutorials/prophoto-rgb.shtml

It may help explain why working in wider gamut (larger) color spaces is beneficial.

--
-Dan Rode
http://rodephoto.com

Unless you are running Apple's Safari. It supports adobeRGB among others. Not sure about the Windows version, tho. It's nice but since most of the world can only view sRGB, it doesn't help much.

Amypics said:

While I agree that my point isn't relevant to the wider discussion, Suntan wanted to know if you had any pictures you could post which illustrated why ProPhoto RGB was "better" than aRGB. When you said it wasn't possible to post, I just wanted to clarify that technically it IS possible to show an example, but for general web use it's not advisable.

Amy

Click to expand...

--
-Dan Rode
http://rodephoto.com

Yes and no. The original with the aRGB profile should not change. One typically creates a jpg copy that is converted to sRGB for display on the web. It DOES have less colors that the aRGB version. the conversion process will map colors in aRGB to similar colors supported by sRGB. For example, to very similar blues in aRGB might be mapped to a single slightly different blue in sRGB. AdobeRGB has 40% more colors than sRGB. A darn big portion of you colors will change slightly .
--
-Dan Rode
http://rodephoto.com

DRode said:

Unless you are running Apple's Safari. It supports adobeRGB among others. Not sure about the Windows version, tho. It's nice but since most of the world can only view sRGB, it doesn't help much.

Click to expand...

Question?? when making a capture, who is the final entity to satisfy. In my case it is me, not the rest of the world. Therefore I am always striving to get the most/best out of my equipment with the least amount of actual work. I shoot aRGB.

DRode said:

--
-Dan Rode
http://rodephoto.com

Click to expand...

Amypics said:

Dejan80501 said:

I shoot with a D300 primarily sports & outdoor using RAW with in
camera Adobe RGB.

Click to expand...

If you're shooting raw, then it doesn't matter what you set your colour mode to. It's all the same raw data, and a colourspace is only applied once you save your image as a non-raw file. Your raw processor or image editing application controls all this, not the camera. If you're using Capture NX to process your raw files, then it will by default respect whatever colourspace was set in camera when you took the photograph, but overriding it is as simple as selecting a different default colourspace in the preferences dialogue.

Amy

Click to expand...

Amy you are absolutely right. Something to note is that although people don't calibrate their monitors, if you store the ICC profile then the browser itself will read the ICC profile if it's programmed to do so...the trend is that Firefox and Explorer are on their way (i believe you can set it yourself at the moment) to making it standard. In other words never shoot sRGB because your doing the quality of your work a disservice. Second, export to web in sRGB mode while exporting the ICC profile along with it. Problem solved. So....

1. Shoot in Raw or Adobe RGB (wider gamut of greens in Adobe RGB)
2. ACR into photoshop
3. Save for web with ICC sRGB profile

Any questions?

--
Shoot for pleasure and the rest will come...

http://www.saykowski.com

Dan said:

It makes a big difference. Once sRGB is assigned, all of the colors available to adobeRGB are gone, never to return. You have thrown away fine color detail and cannot recover it.

Click to expand...

No, Robert's post is correct. But I am not sure that you're talking about the same thing. Robert's post is about RAW (see the subject of his answer), and is true for that format. You may be talking about JPEG: you are then correct as far as JPEG is concerned.

The raw data represents what is captured by the sensor. RAW is converted to JPEG through several steps: white balance, gamma correction/demosaicing/various tonal curves/mapping of the RGB values according to the chosen color space + a few more like sharpening etc.. In any case shooting Adobe RGB or sRGB does not affect the values stored in raw. The image does record a "tag" which specifies your preference for the output color space, but that's all there is to it (in raw).

On a different subject: there is actually more "fine detail" in JPEG in sRGB than there is in Adobe RBG for the same data storage requirement (say 8-bit per channel). What you're loosing in sRGB compared to Adobe RGB is gamut (i.e. some "extreme" colors get clipped). On the other hand the price to pay for a wider gamut space at a given storage requirement is always less color detail.

Which begs the question: do those details really matter? Well not really. Adobe RGB is not that large. So in practice Adobe RGB keeps the advantage of larger gamut without a penalty in color detail which we should care about (some would disagree with that statement).
--
Thierry

Dejan80501 said:

I shoot with a D300 primarily sports & outdoor using RAW with in camera Adobe RGB.

Click to expand...

For RAW, the camera settings don't mean much. They only affect the JPG preview. Capture NX may use those setting and thus be affected. Photoshop won't.

Dejan80501 said:

I generally convert to JPEG and post my images on my website for download or purchase and periodically print in-house.

Click to expand...

Ah, you should be asking three questions: What color space should I use for editing, for the web, and for printing.

For editing, I'd suggest ProPhoto. sRGB for the web. Printing - that depends on the printer.

What you want to avoid is data loss due to rounding in the edit process. Lots of edits require mathematical operations that cause numbers to get rounded. Both sRGB and aRBB are 8-bit values, and you may see banding in the sky, or other such effects. ProPhoto is a 16-bit values, and the rounding effects are largely hidden. So editing in ProPhoto, then converting to sRGB JPG will nearly always give you the best quality.

If printing direct from your application (Photoshop?), there is no need to change the color space. But if you need to export, generally I'd pick sRGB.

To illustrate, here is an exercise: Take the numbers 1 through 10. Divide by 2, round it, then multiply by 2. Repeat using the number 3. Look at the values. You used to have 10 distinct values, now you have only three values (3, 6 and 9). If the 1 through 10 values represented 10 shades of blue in the sky, you can see the rounding process would create three bands of blue in the sky.

Now do the same thing, but use a 10x larger color space by multiplying 10 times the 1 through 10 values (10, 20,.....90, 100). After you perform the same math process as above ( 2, round, x2, 3, round, x3), you'll also see rounding. But now divide the results by 10 (round to whole numbers) and you'll see you are back to 10 separate values.

In conclusion - edit in the largest space you can and leave the final color space reduction as the last step.

--
Ken Elliott
Equipment in profile.

Dan said:

Yes and no. The original with the aRGB profile should not change. One typically creates a jpg copy that is converted to sRGB for display on the web. It DOES have less colors that the aRGB version. the conversion process will map colors in aRGB to similar colors supported by sRGB. For example, to very similar blues in aRGB might be mapped to a single slightly different blue in sRGB. AdobeRGB has 40% more colors than sRGB. A darn big portion of you colors will change slightly .
--

Click to expand...

Another incorrect post. You can't answer in all cases. There are 2 situations:

• The picture has colors out of the sRGB gamut . Then several things can happen depending on the conversion is performed, what PhotoShop calls "rendering intent". The choices are either clipping the out of gamut colors to their closest equivalent in the final color space (sRGB in this discussion), or "compressing" the colors from the original images so that they all fit in the gamut of the final images. This is done whille trying to maintain specific perceptual relationshiips between the colors.

• The picture does not have out of gamut colors in the final color space. In which case the errors are limited to round-up errors which occur during the conversions. These are very small (in particular in 16-bit images).

In any case, one can see which colors will be clipped in the conversion. See the following explanation:

http://help.adobe.com/en_US/Photoshop/11.0/WSfd1234e1c4b69f30ea53e41001031ab64-7679a.html
--
Thierry

You are making the argument that a wider gamut (more colors) is superfluous; that the extra colors add no value in storing, processing or printing? That position defies even basic common sense.

Let's use a simplified example.

You have 500 unique colors in an adobeRGB image. You convert it to sRGB which can only represent 350 colors. Ignoring the fact that each color will possibly (likely) change slightly, what happens to the remaining 150 colors? 350

Moreover, when processing the image, it is better to have more possible colors represented or less?

If not, why don't we just use 256 color or 10,000 color spaces instead? The camera is capable of capturing more color variation that AdobeRGB or even ProPhoto. Why collect them if they are of no use?

--
-Dan Rode
http://rodephoto.com

I am wrong on this point. The only way the colors are lost is if the rendered image is assigned a specific color profile. The RAW file contains as many colors as the camera can record and store.

DRode said:

It makes a big difference. Once sRGB is assigned, all of the colors available to adobeRGB are gone, never to return. You have thrown away fine color detail and cannot recover it.

a 14-bit NEF processed to 16-bit AdobeRAW file has as much color detail as my camera is capable of storing. I want all of the data, all fo the details, all of the fine gradations of color.

One may be happy with the results from sRGB. The web is full of great looking images displayed in sRGB. Just realize that there is a difference.

Robert Peters said:

doesn't matter whether you set the camera to sRGB or AdobeRGB other than the appearance on the lcd screen and the various histogram displays.

--

Click to expand...

--
-Dan Rode
http://rodephoto.com

Click to expand...

--
-Dan Rode
http://rodephoto.com

MasterOfGoingFaster said:

What you want to avoid is data loss due to rounding in the edit process. Lots of edits require mathematical operations that cause numbers to get rounded. Both sRGB and aRBB are 8-bit values, and you may see banding in the sky, or other such effects. ProPhoto is a 16-bit values, and the rounding effects are largely hidden. So editing in ProPhoto, then converting to sRGB JPG will nearly always give you the best quality.

Click to expand...

This in nonsense. I can make 16-bit sRBG files and 8-bit ProPhotoRBG files in a second. How many did you want to order? ;-)

More seriously, you have touched upon some important points and so, let me say this:

1. JPEG files are always 8-bit. TIFF's can be either 8 or 16 bit. RAW's are, well, defined by the camera.

2. Do not, except in dire emergency, perform any color space conversions on 8-bit files.

So, per rule 2, shooting an AdobeRGB JPEG is not a good idea if you're subsequently going to have to convert that file to sRGB.