Review: Adobe Photoshop CS5 – In-Depth

nvkhkhr

Prime VIP
Adobe’s announcement of Photoshop CS5 was accompanied by some of the most talked-about new Photoshop features to date–Content Aware Fill, for example, represented a paradigm shift towards simplicity in image editing. But behind all of the hype with a big release like this, there’s always the possibility of a slightly less shiny reality. In this review I’ll be focusing on Photoshop CS5 in-depth, and exploring the practical side of its new features. Previously Geek.com posted early thoughts on the CS5 Master Collection.
As a graphic designer and photographer, I probably spend more time in Photoshop than I do in my living room. (Sad, but true.) It’s like a home base for anybody working in digital content creation; it’s used by designers, photographers, illustrators, fine artists, video and film creators, scientists… you name it. It’s a common tongue spoken amongst creatives. And as such, Photoshop has become the hub of the Adobe Creative Suite, now in its fifth generation. Even its name has almost become an understatement; it’s certainly not just for processing photography at this point. It’s the software universe’s equivalent of “Kleenex” (or “Hoover,” for readers in the British English world)—one does not retouch; one Photoshops.
With Photoshop CS5, Adobe may have created its biggest headline yet with “Content Aware Fill,” a feature that has generated more buzz (and humor) than any other feature to date. Content Aware Fill bears some resemblance to features that were available elsewhere before CS5’s debut, but its level of integration to the application’s workflow is so tight (simply pressing the “Delete” key can bring up the Content Aware Fill option) that it’s sure to become much more of a retoucher’s staple than a niche tool as CS5 floods the market.
Content Aware Fill (Before)

Content Aware Fill (After)

Having used it quite regularly for a while now, my impression of Content Aware Fill is, for the most part, very positive. It is killer for quick fixes, especially those in backgrounds, areas of low complexity, or away from the main focal point of the image being retouched.
Josh's "Mess-O-Rama" before content aware fill

Josh's "Mess-O-Rama," after content aware fill. Even with this extreme example, the results are quite good. In this example, the algorithm has more than enough data to create a convincing interpolation.

It can save a ton of time when compared with clone tools, especially in situations where more complex selections or masks would be necessary; Content Aware Fill is usually able to blend the resulting interpolated image data into its surroundings, making perfect selections less than necessary. But with larger and more complex areas, Content Aware Fill may not always provide an instant solution. It draws image data from its surrounding context, and so it works best in highly predictable scenarios. In my buoy-in-the-water example, I was able to make a crude selection around the buoy, and one Content Aware Fill later, the buoy was completely gone. And at first glance, it seems like a perfect solution—save for the big vertical seam up the center of the interpolated area. There simply wasn’t enough relevant image data for it to fill such a big void without artifacting.
Photograph before content aware fill.

Photograph after content aware fill. I tried to delete the "T." It seems that letters cannot easily be deleted with Content Aware Fill if they're near other letters.

Even stranger, I’ve found that whenever I use Content Aware Fill to delete one letter from a word in a flattened graphic design, it almost always replaces the letter with another letter from the same word.
But I think it’s useful to bear in mind that Content Aware Fill does not read minds. It is still an algorithm, and we can only expect so much of it on its own. It does a great job in so many circumstances that you’d think it could read minds sometimes. The most promising aspect of the tool is that, in one step, it brings the user very close to the final result. Of course it would not be difficult to get rid of that vertical seam in the buoy shot with a bit of manual cloning. And that’s the beauty of Content Aware Fill; it does some of the heavy lifting automatically, but it doesn’t put us Photoshop experts out of a job. (Read further for ways that Adobe is putting people out of jobs.) But I think it would be nice, in the CS(X) future, to see Adobe upgrade the Content Aware Fill algorithm with some additional options, perhaps more like the Alien Skin Image Doctor Smart Fill interface. That would lessen the randomness of the results, but, then again, perhaps it would ruin the fun.
Personally, I think another less-heralded Photoshop CS5 feature deserves top billing: the new “Refine Edge Selection” functionality redefines how complex masking is done. Again, Adobe may have taken some cues from existing third party software, but there’s a lot to be said for implementation and ease-of-use. CS4 had some similar features for refining selection edges, but they were in their infancy by comparison. It’s now possible to use quick-and-dirty tools—like the “Quick Selection Tool,” which seems a heck of a lot better than the Magic Wand in CS5—to create professional-quality masks and selections, even around out-of-focus edges, or hair. It feels like cheating, frankly! I’m sure I share the sentiment of many other Photoshop pros who have spent years perfecting combinations of clipping paths, feathered edges, and hand-painted layer masks to achieve satisfactory results. Now a whole day’s work can be accomplished in a few minutes. (Apparently the circumstances are even more dire for the proud Guild of Rotoscope Artists, whose entire profession may be reduced to a single “Rotobrush” tool in After Effects CS5.)
notschlock Ma-noir-ah, original photo.

Ma-noir-ah, extracted with "Refine Edge." Note lack of blue fringing that would usually be present. CS5 is able to remove edge color casts automatically.

Ma-noir-ah, bent with Puppet Warp (because I'm not that strong.)

Puppet Warp is another interesting animal. On the surface, Puppet Warp’s mesh-based pixel warping tools may seem only useful to circus poster designers, and other digital pachyderm manipulators. But the auto-generated Puppet Warp mesh allows for an unprecedented level of control over pre-clipped objects, using an intuitive interface which feels akin to playing World of Goo. I’ve found it to be quite useful for mapping two-dimensional designs onto complex 3D surfaces; the kind of operation that wouldn’t be quick or easy using Free Transform. For me, it’s another excuse to be quicker, and less precise, without sacrificing quality results. Features like this that make CS5, in the long run, a very good investment for certain users, depending on the type of work being done.
In Photoshop CS5, “Merge to HDR” has been upgraded to “Pro,” and while there are some nifty new features, such as ghost removal, I would not have been so hasty in granting it a “Pro” suffix. When compared with more professional third-party tools, Adobe’s Merge to HDR Pro plays second fiddle. Of course, Adobe’s interface is simple and easy to understand, which is a bonus compared with some of the more esoteric applications out there. But it’s also slow, kind of clunky, and the results are nowhere near what I’m accustomed to from packages like FDRTools. My take is that the feature is geared more towards “HDR-sploitation” than accurate, professional use.
My proof? Image > Adjustments > HDR Toning. This newbie is essentially an amalgam of adjustments and filters (Shadow/Highlight, Vibrance, Exposure, etc.) blended together into an unholy soup, designed to mimic the look of exaggerated high-dynamic-range to low-dynamic-range (HDR to LDR) tone mapping. I may be a curmudgeon, but this worries me. Could faux-HDR and ridiculously overdriven local contrast be the “lens flare” of the decade?
Fisheye photo, before Photoshop CS5 lens correction

Fisheye photo, made rectilinear via Photoshop CS5 lens correction

Speaking of lenses, CS5’s Lens Correction functionality is pretty good; comparable in many ways to the professional’s tool of choice, PTLens. Compared with PTLens, though, Adobe’s Lens Correction lacks as much out-of-the-box compatibility. It ships with only a fraction of the calibrated lens/camera combination presets that are necessary for quick corrections. But I’m sure this will change over time, and Adobe has created an automatic calibration tool to help users quickly build the database. It’s now possible to perform lens correction and calibration directly inside of Camera RAW, for non-destructive lens corrections within the RAW workflow. Camera Raw 6 also adds film-like grain effects on a nondestructive basis, along with improved noise reduction and more control over post-crop vignetting.
The new Mixing Brush tool and overall improvements to the brush painting engine in Photoshop CS5 should prove to be an asset for digital painters. Brushes can now use physically-simulated bristles. This breathes far more life into the tablet+Photoshop combination, so for Wacom users, it will be a great new toy.
Photoshop CS5 Extended’s new Repoussé 3D (oh là là, Adobé) seemed to me on the surface like “Adobe WordArt,” but it’s actually capable of some impressive 3D manipulation. It takes a lot of adjustment initially, but I find myself using it for quick 3D text extrusions. It’s certainly a lot less time consuming than creating an entirely different file in dedicated 3D software like 3ds max, rendering, and then outputting to Photoshop. Edits can be made natively within Photoshop CS5. Repoussé doesn’t do everything, though, so I’d still recommend dedicated 3D software for more complex tasks.

The supporting cast of less in-your-face new features and tweaks are straight up evidence that Adobe listens to customer feedback:

  • It’s now possible to automatically delete empty layers (using a script), to keep your .PSDs nice ’n’ tidy.
  • Using the ruler tool, one can now straighten an image automatically; no more multi-step time wasting.
  • Remember how annoying it was to open a boatload of images at once, and then be asked, one by one, “Save changes?” before exiting? Now it’s possible to deliberately “Close all without saving.” So now you can leave work at the end of the day much faster.
  • A preference is now available that allows “Save As” to default to the file’s original folder, or to the user’s last used folder. Another time saver.
  • When dragging a file into Photoshop from an OS window, the file will now be placed into an existing open document as a Smart Object (unless the file is dragged away from the document’s canvas and towards somewhere else in Photoshop, in which case it’s opened normally). This can make the File > Place operation quite a bit quicker, but it can also get on your nerves when you simply want to open the new file separately.
  • Scrubby Zoom is one of those “huh?” features that might confuse the heck out of new CS5 users. By dragging right/left, the zoom tool now defaults to “scrubbing,” rather than marquee-dragging. It’s zippy and efficient when used effectively, but it takes a bit of getting used to at first. I wanted to leave it off in preferences, but since I’ve been using it, I like its level of precision.
  • Color selection using the eyedropper is now easier to see, thanks to the heads-up “Sampling Ring.”
  • Last but not least, performance: the first time I ran Photoshop CS5, even with all of my plugins, presets, and what-have-you loaded up, I was shocked at how much faster it loaded than its predecessor. The speed is noticeable throughout the application, as many functions (and the overall feel of the program) just seem more spry, and more stable.
Overall, Adobe Photoshop CS5’s improvements over the last version will, for some, make this a must-have release. Even though some of the big-ticket improvements might seem like novelties for those who might be unaffected by them, there are enough tweaks, timesavers, and most notably, speed and performance enhancements to make CS5 worthwhile. At around $650 for Photoshop CS5 (and $1,000 for Extended with the 3D features), it’s actually a very worthwhile investment for any professional. The base version might provide a little more value for money, depending on how often you will use 3D within Photoshop. But upgrades are $200 (cheap, frankly); CS3-and-earlier users especially have no reason not to upgrade, and CS4 users will see major differences in performance and workflow.
 
Top