So, here's the overview of what I go through preparing a photo for display at my top quality: Scanning or raw conversion: I try to get the overall color balance close here. Otherwise I work to get maximum information, rather than the best-looking image (the information gives me more options in the later stages). So I don't go to set final black and white points here, for example, I try to get all the information out of the original into Photoshop. For that and other reasons I produce 16-bit outputs at this stage. Noise Ninja: At ISOs above 400, and some other times, I run Noise Ninja on the background layer (or a duplicate copy of it). Sometimes I then back out of that and go back to put it in selectively with the history brush. Perspective adjustment: I consider, at least, whether I want to straighten anything, or adjust perspective. I've found this useful for everything from architecture to a picture of my and my wife sitting on our sofa (I'd shot from too close, and the hand on her knee was too big, and change the perspective adjusted that). Maybe 25% of photos I work over actually get a perspective adjustment. Spotting: Especially for scans, but also sometimes in digital if I've been shooting at wide apertures for months and suddenly go outside in the sun and stop down to f/16 and all the dust-bunnies on the sensor say hi! If I have a clear idea of cropping in mind, I may skip spotting work outside the area to be finally exhibited (and may have to come back and fix that later). This goes on a separate layer, or the same layer as the simple retouching. Retouching: Especially if the image contains a human face as a main subject. This is on one or more separate layers, and will use rubber-stamping (and spot healing tool), painting, and sometimes layers with special blending modes and hand-painting on them. Sometimes I'll have to extend a background or something, too. I've had to resort to inserting noise into burned out areas in an attempt to tone them down without making them look too horrible. All sorts of things can happen here. Curves: A curves adjustment layer to get the overall look about right. Also white balance, if it needs refinement past what I did in RAW conversion. Most commonly I'll click something for gray balance, and adjust two or three points on the composite curve. Sometimes I hand-adjust the individual color curves as well if the lighting is complex or weirdly-colored. If there's nothing for gray, I'll manually compute a numerical average of high, middle, and low samples and adjust the middle based on that, that takes points on the three color curves. Dodging and burning: For me, this usually takes the form of additional curves adjustment layers with layer masks. It's rare for me to get away with fewer than two curves layers in a serious exhibition print. Local contrast enhancement: I find it's worth considering trying an unsharp mask with about 64 pixel radius and 5-25 effect to see what it does, especially in landscape photos. Well over half the time in landscape photos, this does very interesting things with the local contrast and produces a much more intense image. (This is applied directly to the background or background copy; I may then have to go back and fine-tune the curves adjustment layers). Other weird adjustments: I might use a color-balance adjustment layer, in particular, if things were weird enough (I combine it with the curves adjustment normally, if it's simple). Misc: As needed. Also any borders or text overlays or whatever, beyond the simple photo (mostly I don't do text overlays or borders on photos I'm doing the exhibition-quality treatment of). I'll use Focus Magic selectively sometimes if it looks necessary. (And this version is the version I save and keep in a PSD file). Cropping: I try to decide this late, but it interacts with the retouching steps and the perspective correction (and sometimes I have to extend the background in some part of the image to make the rectangle I want to crop to after perspective correction). Resize: for web display; I always print from a full-res version. Sharpening: Generally using smart sharpen. Always at final resolution, and done specifically for the size I have in mind. Color profile conversion: For web versions. Also reduce bit depth to 8. Um, just to clarify, I'm not claiming any great chops as a printer (by saying "exhibition-quality" over and over). I'm using that to indicate photos I'm doing *my* best on, as opposed to snapshots and such where "good enough" arrives much earlier. It's very, very rare for every one of those things to be done to the same photo, but I really do consider them all along the way. -- David Dyer-Bennet, dd-b@xxxxxxxx; http://dd-b.net/ Snapshots: http://dd-b.net/dd-b/SnapshotAlbum/data/ Photos: http://dd-b.net/photography/gallery/ Dragaera: http://dragaera.info
