New Year's Resolution Number Three

Everyone I read on the web has their own flavour of workflow. On the surface, it sounds easy enough - take your picture in raw, develop it with a raw converter, tidy it up in a photo editor and either publish it on the web or print it out. What could be easier than that?

Well, peel back the first skin of the onion and you are confronted with a gazillion choices and pitfalls. Here are just some of the questions you have to answer:

• Should I bother taking photos in raw?
  
• Do I keep my raw file in its original camera format or convert it to a standard format (DNG)?
• What raw conversion software shall I use?
  
• How do I organize my directories to track all these files?
• Do I keep my raw files whizzing around on my hard disk or store them off-line on DVD?
• What colour space should I use in my camera?
• What colour space should I convert to?
• How do I calibrate my camera, monitor and printer so that all the colours match?

I like to contrast the views of two experts in colour management. The first is Tim Grey, who is a raw purist and has written excellent books on workflow and colour management. Tim argues that you should keep several versions of each file, make sure each of your editing adjustments are done in separate layers, save a Photoshop file containing all your layers so that you can go back and adjust them etc. He is also meticulous about using the colour space with the widest gamut and retaining full file resolution to the very end of the process before publishing or printing. I highly recommend Tim's books and his daily DDQ e-mail (you can sign up here).

The opposite extreme is Ken Rockwell, who insists that shooting raw is a waste of time and that you should do your whole workflow in the sRGB colour space. Ken is no dummy - he was a digital hardware engineer at one point in his life and developed colour space conversion chips, so he knows his stuff. His site is refreshingly candid and well-written - worth a visit as an antidote to the raw workflow purists.

I'm going to adopt a new workflow this year that comes down somewhere in between. Here is my new digital workflow with the rationale for it:

• I'm going to continue to capture photos in raw and in the Adobe RGB colour space. My lovely wife gave me a 4GB flash card for Christmas, so space is no object. I also am not an experienced enough photographer to be able to a) remember to change the white balance before I shoot and b) to be able to set the white balance properly. Shooting in raw allows me to see the white balance chosen by the camera and play around with a bunch of other values until I get something that looks right to my eye. I'm going to choose Adobe RGB to capture the widest gamut as possible in my image.
  
• I'm going to store all my raw files in DNG format on my hard drive so that I can go back to them later if my editing technique improves or if I need to develop the files for a different use or format. DNG has enough support (e.g. Hasselblad) to the point where where longetivity of the standard is not an issue.
  
• Adobe Photoshop CS3 Bridge will be used for raw conversion. I've tried lots of others and the new raw conversion dialog in Bridge is the best out there so far. I like Phase One, but they haven't delivered on DNG support. I loved Pixmantec Pro, but the product is going by way of the dodo bird now that Adobe has scooped the development team. Lightroom is a total waste of time in my opinion. The file handling is far to complex and the good bits, like the sliders that control the curve shape and the Pixmantec vibrance setting are replicated in CS3 Bridge without the pain of Lightroom. I also question the thinking behind Lightroom in the first place. Being able to handle a high volume workflow in a tool separate from Photoshop sounds good in theory, but nearly every photo requires some action that can only be done in Photoshop (e.g. retouching, red-eye correction, local contrast enhancement, sharpening or noise reduction via a plug-in). It just isn't possible to process a large batch of raw files without a trip to Photoshop, so why not use CS3 Bridge as your main tool?
• While it would be ideal to have an intermediate, uncropped version of each image, with full 16 bit depth and multiple saved layers and then have versions of each image for print and web applications, I have neither the time nor the disk space to do this. Instead, I'm going to standardize on a compromise. Each image will emerge uncropped from raw conversion as a 16 bit file in the sRGB workspace. Then, Photoshop will be used to crop for maximum composition appeal, apply noise reduction, sharpening, touch-ups, local contrast enhancement and fine-tuning of curves. I'll save as an 8 bit jpeg file as my last step.
• My rationale for converting to sRGB is based on the persuasive arguments of Ken Rockwell. If your target application is either a web post or a commercial photo finisher, then they are expecting sRGB files. There is no point in editing your photos in a workspace that supports a wider gamut if the image is going to be rendered from sRGB. You might as well live in sRGB and optimize that image in the workspace that will be used for viewing or printing. Try an experiment yourself. Take an image with lots of blown highlights. Bring up the Adobe raw converter with the blown highlight box ticked in Adobe RGB. Now, reclaim as many of the highlights as you can. Change the colour space from Adobe RGB to sRGB. Chances are that even more highlights are now marked as blown out. If you'd continued to edit the image in Adobe RGB, you might have taken the image all the way to the end of your workflow, converted to sRGB to send to your printer and then found out that highlights had seriously blown out. Even if you print your images on your own printer, do you want a result that closely matches what you see on your monitor (an sRGB device) or do you want unexpected colours on the printer resulting from colours that may have been out of gamut on your monitor, but show up on the printer?
• If I need to do another version of an image (e.g. to match it to a certain standard print size like 18x12), then I'll go back to the raw image and do it over again. Yes, this may mean that I won't be able to capture the identical image to my first version, but usually my editing operations follow the same path and I can come pretty close. This sort of thing happens so infrequently that it is preferable to keeping an intermediate image with all the layers intact.
• If I have a particularly good image, one that stands above the rest that has had a lot of painstaking care applied to it, then it will be stored in an intermediate form with all the editing layers.
• For printing, I'll continue to use my R800 for anything up to 8x10, but will continue to use Costco for 12x18's and Walmart for 16x20's and larger. The Costco 12x18's are $3.00. This price can't be beat if you factor in ink, paper and the cost of a larger Epson or Canon printer. I'd love to have the control that my own wide printer would give, but can't justify the expenditure when results from Costco are pretty darned good.

I'm sure Tim Grey will cringe at this workflow (and so will Rockwell, no doubt), but this is the compromise I've settled on for now. I have all the raw images to go back to if I change my mind, which I inevitably will.

If anyone has any suggestions or comments, please let them fly. What would you change?