Holy C&^$ there is a lot to learn-- questions about color space and image quality.

[Helen B]

I shot some comparisons using both RGB and Adobe in both MPEG fine and the same two color spaces in NEF (RAW) 14 bit.

A Raw file has the colour space of the sensor. The colour space setting on the camera only affects the JPEG files, not the Raw files.

Best,
Helen

 

[expat42451]

Hi Helen--

Interesting. If thats true then what would determine the color space of the sensor in relation to the color space requirements of the "rest of the chain" so to speak I.E. monitor, printer &c. Another question. In EXIF data for the test shots when I am shooting RAW it tells me either I am using sRGB or Adobe RGB depending on what color space I have set the camera on. I am under the possibly mistaken impression that color space data gets saved no matter whether in RAW or JPEG.Also that either setting determines the depth of color space.....

I am not trying to be perverse or convoluted, only trying to understand. Thanks.

Expat

 

[clanthar]

Consider that there are two kinds of color spaces: analog and digital. A digital color space is a specific number mapping of a range of colors. An analog color space is sans-numbers but nonetheless physically limited. The color space of your camera sensor is it's physical recording capability. RAW files do not typically assign any further restrictions. A specific numbered (and named) color space is assigned by the RAW converter software. Whether the sensor actually was capable of recording more or less content than the numbered color space is an intriguing question. If the sensor physically records more, then data is lost in the RAW converter. If the sensor physically records less then we can assume that's all you get.

Joe

 

[The_Traveler]

Let me say something in print that I'm certain everyone else responding knows but I want to fill in this bit of assumed info for others.

When you open a jpg (compressed) in PS (or any equivalent software) that 3 or 4 M file becomes 12-20 M as the compression is reversed.

Some of that information in the now-large file is actually created by the decompressing algorithm, because some of the original real information detail has been discarded as part of creating the jpg by processing and compressing the raw data. (thus a jpg is a lossy file format)

If you edit information and resave, the compression algorithm may indeed squeeze up some of the originally compressed parts but in all likelihood, additional other areas will be compressed (i.e. more original detail will be discarded to make the data stream shorter)

So each time you open (uncompress), edit and save (recompress) you are probably deleting more of the original data.

Thus a jpg file that has been opened, edited and saved multiple times will have less original detail than the original jpg.

 

[The_Traveler]

And yes, there is a lot of crap to learn - but it all become baseline knowledge that is incorporated in the way you do photography.

Technical issues are not important until they hurt the photograph.
If a photograph is great, no one gives a damn about any of the technical things; OTOH, the technical issues became a handy metric to define where an image fails.

 

[expat42451]

Hi Traveler How are you amigo??

I understand what you are saying about JPEG being lossy. Thing to do is take good shots and not have to do much to them after the fact. However if you could see how many of my images "fail".... and I am trying to understand this new medium. I have thanks to this thread and a lot of help from a lot of you more knowledgeable than am I , at least a better idea now of the medium....but a lot more to learn. One thing I have done is sent a query to Nikon about whether "color space" whether in sRGB or AdobeRGB gets "saved" in some form in RAW like it obviously does in JPEG. Really this is part of my effort to understand and will have not a lot of direct bearing on the end result of what I shoot. Helen B was very nice to point out a couple of posts ago that I may well have been straining at a fly and she may well be right. In any case my curiosity bothers me when I wonder things and I figure this is the best place to ask...... I will share what Nikon tells me here of course.

Regards
Expat.

 

[JimCoventry]

I have not read ALL the threads herein, but amazed to find so much misconception. Having been down these paths a million times and in teaching my employees, clients and a lot of people in workshops I hold, I'll tell you all some things I know to be true - for my work flow. It is not expensive but I have tried many other flows as I see you all are (by these threads).

First, there are a lot of folks who are already good photographers and usually get their lighting nailed in the first frame of two by first planning the shoot then shooting the plan. They are proponents of shooting JPEG. However, I am in that group and since there are now so many options in post production and getting different results on the fly I shoot only RAW images. Raw images allow me so much latitude I would not ever go back to JPEG. Just as an example, highlights that in jpg get blown out, can never be brought back - in RAW they often can.

RAW is not really an image file it is like a negative. Your computer processes the RAW file at your will and settings. That all can be done pretty quickly and automatically.

You have to remember, jpg is basically a compression routine to make files smaller. It automatically decides which pixcels to delete EVERYTIME YOU SAMVE THE FILE. The major actor in this case is the Discrete Cosine Transforamtion (or DCT) which divides the image into blocks (usually 8×8 pixels) and determines what can be “safely” thrown away because it is less perceivable (the higher the compression ration/lower quality JPEG, the more is thrown away during this step). And when the image is put back together a row of 24 pixels that had 24 different tones might now only have 4 or 5. That information is forever lost without the raw data from the sensor recorded in a Raw file.

ALL prints are made using SRGB color space because Adobe RGB, for instance is too big for the print paper and or inks, and or dyes. Ink jet printers can use the Adobe RGB color space but you need to be very careful to be consistent. SRGB has a smaller (more narrow) gamut and it's the "best guess" in approximating different computer display devices.

Also, I work 16 bit out of Lightroom and into Photoshop all set to SRGB. I first edit in Lightroom for both organizing files and averaging color across any given shoot. I then export out of LR to PSD files at the largest possible size, in 16bit at 500 dpi. That said, that's how I work on images. When I send them to a lab, I have built actions that automatically convert to 8bit and 300dpi to which ever size needs to be ordered all saved to jpg files.

As to monitor accuracy: I am Mac based and calibrate my monitor using SPYDER II, but there are other devices that are good to. I insure accuracy by adding to my monitor - software called GAMMA CONTROL. There are other PC versions but they are more obsure (harder to find and make work) sorry about that PC lovers. Gamma Control allows me to "fingerprint" my monitor(s) to any printing device, lab or printing press. I simply get sample print outs of the same image from the target device and use Gamma Control to adjust my screen to match that print out. Then I save the results in a file which can be recalled each time I elect to use that device. This allows me to GUARANTEE my color. I make such guarantees to my commercial clients who depend on having accurate color in their marketing materials and web space. SPYDER simply puts the monitor in it's best state to start using Gamma Control.

Sorry to those I've confused. I hope I have helped. Write back if I can help.

 

[expat42451]

Mr. Coventry

Thanks very much for outlining how a professional workflow happens. Actually for me no confusion created, particularly in understanding how JPEG compression and decompression works. How I started this thread was in a couple of coupled questions about JPEG, RAW and color space. You and the other people here have done a lot to help me understand (since I am new to digital) how things happen.

Expat

 

[clanthar]

I insure accuracy by adding to my monitor - software called GAMMA CONTROL. There are other PC versions but they are more obsure (harder to find and make work) sorry about that PC lovers. Gamma Control allows me to "fingerprint" my monitor(s) to any printing device, lab or printing press. I simply get sample print outs of the same image from the target device and use Gamma Control to adjust my screen to match that print out. Then I save the results in a file which can be recalled each time I elect to use that device.

I think I'm reading this correctly; it sounds like you alter the appearance of your display to better match various output devices. It also sounds like you arrive at a specific "fingerprint" setting on your display by using software to alter the display visually. Wow! I'm glad that works for you, but it sounds really difficult and shaky to me.

I would want my display to be an anchor point of standardization. Once I had it there I sure wouldn't want to see my images change in the process of prepping them to various output devices.

What I do and what everybody I've met who understands color management does, is calibrate and profile my display to industry standard conditions. Then when it comes time to print for example, I use an ICC control profile for the printer (output device) that reconciles any unique characteristics for example the paper stock. I use both paper manufacturer supplied profiles as I find them acceptable or make my own custom profiles which I can then edit to fine tune as needed. This way my image on my display once edited is "anchored" as accurate and completed. I do not expect to see it altered thereafter for any reason.

Photoshop is designed to support the method I just described (View/Proof). If I'm concerned that a printer for example won't be able to physically handle the color gamut of my image I can soft-proof that in Photoshop by loading the ICC profile for that printing device. At that point I can if necessary try and modify the image to adjust for the inadequacy of the printer or decide to switch printers. Commercial printers likewise support this methodology as they are typically happy to supply ICC profiles on request.

Joe

 

[Helen B]

The Raw file can contain your choice of colour space in its metadata, but that does not affect the image data itself. It is raw data from the ADC (analog to digital converter) that is usually unprocessed, though Nikon cameras do some noise reduction, so an NEF file is not truly raw data from the ADC.

The Raw processing software converts the Raw file into a high-bit, large gamut colour space for internal processing. All colour managed space conversions require the source space (the native color profile of the camera in the case of a Raw file) and a destination space. Conversion between spaces goes via Lab colour. (The image data is also converted for display using a monitor profile no matter what colour space the image is in.)

The Raw file includes embedded metadata that acts as a set of instructions to the Raw conversion software. You can change those instructions by adjusting settings in the Raw converter, and your instructions are saved either in the metadata or in a linked xmp file. The original image data is not changed in a Raw file, only the instructions on how to handle it. When you export to another format (PSD, JPG, TIF etc) the image is converted from the internal colour space the conversion software uses, via Lab, to the chosen output colour space, such as sRGB or Adobe RGB or even one of the CMYK profiles if that is what is required.

I use a variety of colour management tools including ProfileMaker 5 for creating and editing device profiles (camera, monitor and printer) and CoCa for camera profiles. Both of these make ICC camera profiles, usable with Raw Developer, DxO Optics Pro and other Raw software but not Lightroom so I also use the X-Rite Passport system for Lightroom/ACR profiles though it is not as good as the PM5/CoCa - which may not be important to you. Chromix ColorThink is good for examining profiles and comparing them with the gamut actually used by a particular image. This, for example, is a diagram of the colours in an Adobe RGB landscape image (the dots) at a mid-range brightness compared to the printer profile, sRGB and Adobe RGB (the two similar shapes) all at the same brightness:

You can see that the printer has problems handling the saturated blues and purples in both sRGB and Adobe RGB, but that using Adobe RGB makes better use of the printers gamut in cyan/green/yellow.

Best,
Helen

 

[expat42451]

Helen

Thanks very much for the information. Still not sure I completely understand-- the color space that you "save"then when shooting RAW, would be part of the metadata that the destination space uses to know how to handle the source data i.e the actual image data then?? Or would the metadata be totally separate and something different i.e. color space settings on the camera only apply when shooting JPEG?

Something else fascinating here that you are alluding to--- you seem to indicate that you are able to actually able to change quite a bit in the way that the camera converts the output of the A/D to raw. Thats pretty fascinating! I assume thats done by loading the changes you create either with ProfileMaker or Coca into the camera via one of the memory sticks (loading a profile or setting) or using the camera in tethered mode...... I had no idea there was so much flexibility in this....its almost like being able to change the properties of film emulsion at will.........

Thanks very much for the diagram and for the time you have taken to try to help me understand this....what it demonstrates at that luminance point is interesting...again I go back to what I have said before, that I did not realize how much actual science there is in making the transformation from the output of the sensor on the front end--to what either appears on your screen or what appears on a magazine cover as the final product.

Expat

 

[Helen B]

The ICC camera profiles made by PM5, CoCa (and ColorEyes and others) can only be applied by some, but not all, Raw conversion software. They don't work in-camera. As I said, the Raw file coming out of the camera has minimal processing applied in camera. That's the whole point of Raw - it moves almost all the processing to a later stage. You do indeed have the potential to change the look of the image. The colour space set in camera only affects the JPEGs, although it can be recorded in the metadata (or it could be read from the accompanying JPEG file, which should be tagged with whichever profile was used to make it).

With some cameras, like the D3, you can upload your own tone curve (gamma curve, contrast curve) but that does not affect the Raw file.

Best,
Helen

 

[expat42451]

Helen B

Thanks very much! Question explained.

Still given the headroom if one will of the sensor, it really is almost like being able to "change emulsion properties" to an extent using RAW data--as you said, not within the camera, but later. Thats pretty awesome!

In the D7000 (and I am not completely sure of this) I believe that it might be possible to upload what Nikon calls "picture control" settings. The Picture settings are divided into Standard, Neutral, Vivid, Monochrome, Portrait and Landscape. I have already changed the Standard settings (what I shoot on so far all the time) to increase the sharpness -to my eye the D7000--and it may be a combination of things and may also even be "kit glass"- seems at times a little "soft" ...but one is able under each of the aforementioned settings to adjust Sharpening, Contrast, Brightness, Saturation and Hue. I assume you are doing something similar on the D 90 but I also assume that changing these settings would only affect the JPEG output and not the RAW output of the D7000 like the D 90.



Expat

 

[clanthar]

Helen B

Thanks very much! Question explained.

Still given the headroom if one will of the sensor, it really is almost like being able to "change emulsion properties" to an extent using RAW data--as you said, not within the camera, but later. Thats pretty awesome!

In the D7000 (and I am not completely sure of this) I believe that it might be possible to upload what Nikon calls "picture control" settings. The Picture settings are divided into Standard, Neutral, Vivid, Monochrome, Portrait and Landscape. I have already changed the Standard settings (what I shoot on so far all the time) to increase the sharpness -to my eye the D7000--and it may be a combination of things and may also even be "kit glass"- seems at times a little "soft" ...but one is able under each of the aforementioned settings to adjust Sharpening, Contrast, Brightness, Saturation and Hue. I assume you are doing something similar on the D 90 but I also assume that changing these settings would only affect the JPEG output and not the RAW output of the D7000 like the D 90.

Expat

I shoot Canon rather than Nikon, but I'm going to assume similar behavior in RAW processing from both. What you're describing is also standard with Canon's DPP (Digital Photo Professional) RAW converter. The array of image adjustment choices available from the camera menu all show up in the RAW converter software. As such they can be applied at will after the fact and the understanding is that whatever you may have set on camera is NOT applied in-camera to the RAW files. Those variables only apply when the camera's image processor is producing the camera JPEGs. Shooting RAW gives you the option to apply or not those options selectively after the fact. The basic assumption is that a camera that saves RAW files will take a hands-off approach to that RAW data and extend to the photographer as much control as possible. In practice I think we find some corner cutting and maybe a little funny business going on. Next you have the question of the RAW conversion software. A disconcerting exercise is to open your RAW file in 1/2 a dozen different RAW converters and then realize that some degree of "English" is being applied up front before you get your shot.

In your latest post (Expat) you noted sharpness -- a note: It's assumed that some degree of software sharpening will be necessary for digital images. I think you will find that best practice is to resist sharpening your images either in camera or during initial processing since the degree and method of sharpening is output dependent. My cameras are set for hands-off sharpening (the JPEG only cameras) and I resist sharpening my processed "edited" photos. However I sharpen every image I print or post on the web. I sharpen at output appropriate to the output.

This thread started with color space questions. Color space then is selected during the RAW conversion process. You'll notice that when color space is brought up it's in reference to RGB. Helen's posts noted the use of Lab in the conversion process. We don't have different Lab color spaces. Lab is a recognized international standard. RGB color spaces exists because we've got different "user groups" who have adopted different numerical mappings for color in the RGB structural model -- you need a color space if you need an RGB image.

This whole thread starts to ask the question RAW versus in-camera processed, and even more directly where do you draw the technology line? It's a hard question. I have two students in class this semester who are starting out in the wedding/portrait business. They came to class shooting in-camera JPEGs. They really don't want to adopt the extra work to set up a RAW to JPEG work flow. When I asked them if they knew their cameras saved RAW files they said, "what?" I can understand their reservations. They want to believe that the camera can do the work. Many photographers in their line of work have adopted camera JPEG only work flows. It all gets back to your original observations -- how many layers of the technology do you want to peal back? Helen talked about creating and using camera ICC profiles. I have the hardware to do that and I ran through the process some time back. I need to be able to walk through it in class, but personally I don't bother. If I shot in controlled studio light -- maybe, but I'm a field photographer. That much technology really doesn't help me. On the other hand I couldn't live without my ability to create and edit printer/paper profiles. So it's going to be interesting to see where you end up?

Joe

 

[expat42451]

Clanthar

Thanks very much for the additional light shed on what happens....again more that I did not know. And again the additional explanation leads to a few more questions--

"It's assumed that some degree of software sharpening will be necessary for digital images. I think you will find that best practice is to resist sharpening your images either in camera or during initial processing since the degree and method of sharpening is output dependent."

Output dependent meaning--what type of format (i.e.JPEG,TIFF &C) that you plan to save the images in final form after processing or for what purpose they are destined?

Additionally in paragraph 1 you talk about RAW conversion software. Again is this what you use to convert to the final image for whatever use you intend it? For example in ViewNX2 I use this to download the RAW images from the camera.ViewNX2 sounds pretty crude compared to what you and Helen B and a lot of others here are doing.... I save as as RAW images on the computer. I can convert them to JPEG's after processing if I so desire, or barring that simply leave them RAW in my library. Simply leaving them RAW....but maybe not when considering storage media and file size. In part of my travels on this machine earlier this evening I ended up formatting the second 250GB hard drive (where Susue Linux, no longer used much resided) for use as bulk storage for photos. I dont need it yet but at this rate I certainly will considering that the first 250 GB drive is about 60% full and I ve only had the camera for lets see 3 weeks more or less...........

Also you said " I sharpen at output appropriate to the output." I assume you mean here whatever destination (i.e. web posting or printed for a photograph or printed for a page in a magazine) the image is headed for?

Understand about the need to setup a profile for a printer. Again this is fascinating and a far cry from when I would shoot 400 ASA(ISO) film , process the transparencies myself and then either have someone print them or borrow a buddys' darkroom to make prints....unreal.

Observation-- maybe this is not true-- but you are when using JPEG images, actually placing the responsibility for the RAW conversion and a lot of the image post processing with the camera. Agreed that the new cameras are sophisticated as hell....but arent you almost always going to get a lot better results shooting RAW and doing conversion on an editing machine i.e. a computer with enough power and memory to avoid things like truncation in long number chains &c? Aside from the question, what might be arguable is that depending on the camera and the shooter and the glass and where its viewed, it might not be possible for the average person, when placed before 2 images, 1 JPEG fine and 1 RAW, to tell much difference if any? I am not so sure that I would be able to because I have not done that much with the D 7000, same shot in comparison. I have the option to save both RAW and JPEG and may take some shots tomorrow and try to look closely and see what is discernible. I have never tried that. As an aside I do not know whether Nikons' CaptureNX2 (the companion to ViewNX2 which is bundled with the camera) allows one to apply in camera set parameters to RAW images in post processing like Canon's DPP that you mention but that seems like a very powerful tool.

Dont get me wrong. I want to learn as much of the science and technology behind this as I can because being an old geek its pretty fascinating. I also want to learn to take really great pictures, something that will touch people who see them.

You are absolutely right peeling back the onion. Where will I end up with all of this. Right now I have no idea other than keep learning and trying to improve composition and technique. Thanks for the continued interest and help!

Expat