"Do you...think about composition like in terms of rule of thirds when the camera is capable of shooting 360 degrees in one shot?" --dpreview post
I think you now know why I'd like to strangle the person who started that whole "rule of thirds" compositional shortcut ;~). Everyone's looking for a shortcut, I get it, but rule of thirds is actually a dead end most of the time, not the short route to something useful.
What people want in their composition is balance, but balance doesn't take one form. You wouldn't, for instance, place a subject on a one-third point if the goal was to show isolation, for instance. In such a case, the composition needs to be highly offset, with something well outside the one-third lines balanced against the rest of the frame. Or maybe a small subject in the direct center of the frame.
The key to composition is knowing what you want to say, not knowing where to put a subject in the frame. I really need to make a Sesame Street style video to illustrate this ("behind", "next to", "near", "far", etc.).
One thing I always laugh about with rule of thirds illustrations is that they'll draw the lines, and then the actual subject focal point isn't actually on an intersection, it's just nearby. Some rule. The rule seems to be "put a subject about here, or about here, or about here, or about there." The resulting photo may or may not be "about right."
"Shop your closet first." --subject line of an email received
A reader after my job ;~). Love his wording, so I'm going to steal that line in the future.
But the point is extremely well taken. Too many people go to the B&H site when planning for an upcoming trip rather than pull things off the shelf of their gear closet. Yes, we have more megapixels now, and some recent lenses are miraculously sharp, but you bought and fell in love with that 58mm f/1.4 for a reason. Rediscover that reason before trying to replace the lens.
Not a strange thing said, but a strange thing shown: press release photos of new lenses with chromatic aberration, improper color rendering, poor focus choice, and so on.
I'm going to pick on Tokina for a moment, as they're just the latest that have done this. Here's their press release photo for their latest lens:
Notice how blue it is? There's a magenta/blue tint throughout the image. Terrible JPEG compression artifacts live in several places in the image, as well. The shadows are crushed. White isn't white, but "paper". Now let me clean it up a bit for them:
See the difference? I thought so.
Now here's the thing: if a camera or lens company can't render their press images cleanly, accurately, and consistently, would you trust their engineers to have designed the lens cleanly, accurately, and consistently?
Not that the camera and lens companies are the only ones that have this issue—it's so easy to fire most of the marketing team since they don't "create revenue" and contract out work to a bunch of interns—but you'd think that if you want folk to believe that your product accurately captures what's in front of it, that you could do that in your marketing, too.
"It's very hard...not technically impossible...I may do my best." --Fujifilm engineer describing whether 160mp in-camera pixel-shift images might be a future feature on a firmware update for the X-H2, in an interview available on YouTube.
The actual answer is far longer than my abbreviation, above. My reading of the entire response in the original language is essentially the way the Japanese say "no." The Japanese will often answer a question in a way that suggests that "yes" might some day be a possibility, but what they are actually saying is "no, we won't do that." It's a cultural thing, where direct confrontation with the questioner is avoided by giving some respect to the question.
But here's the thing: none of us care if something is hard to do for the product designers. What we care about are features that allow us to get our job done more productively. Pretty much every pixel-shift implementation so far requires you as a user to go out of your way to set the camera to collect a set of images, then requires you to use some terrible piece of computer software that was coded in haste to create the pixel-shifted file. Some require you to go further and process the file once stitched together. From a user/customer standpoint this is not what we want. Let me easily set the camera to pixel shift, let the camera create a combined file (and perhaps keep the originals, as Nikon now allows with many of their similar in-camera features), and let that combined file be one we would process with our normal tools, in a normal fashion. Update: someone reminded me that the Leica SL2 produces a 187mp DNG file directly in the camera. Which is just a further indictment of the Japanese reluctance to do the same.
When you stop caring about what the customer really wants and instead base feature creation on how easy it is to implement for you, you've just created a cardinal sin in my product management book. You want my money? Give me the right implementation. I don't have extra time to waste on extra steps in my workflow that don't produce optimal results.
