I recently read Jeremy Clarkson’s I Know You Got Soul. For those who don’t know, Jeremy’s one of the hosts on Top Gear, which can basically be thought of Mythbusters if the myth was “I heard it won’t go that fast”. An interesting read, all about machinery that was the technical embodiment of hopes and dreams, quixotic or not. One of the things Jeremy mourned was the loss of Concorde (never, ever “the Concorde”. Concorde.). Never economical, the only supersonic passenger jet capable of managing three hour hops between Heathrow and JFK was grounded when one of them went up in flames a few years back. Jeremy wrote, rather wistfully, that technology was only supposed to improve. But a few years ago, you could make that trip in three hours, and now, quite simply, you could not.
Of course, not everyone could afford Concorde. The entire travel industry has realigned towards bulk transfer of people at the lowest cost possible. Airbus may very well be the most presciently named company in history. Concorde was many things, but it could never be “just a bus”.
But this is not to be a post about travel — having finally made it home after five weeks on the road (a shattered soul with a smile on his face, as predicted), I could do without planes for the moment. It was a fun trip, yielding a veritable bumper crop of pictures from my Fuji F30.
I have never met Nathan. Having read this, I am thoroughly convinced: This is a man who misses Concorde.
Nathan is writing about a perceived end to the revolution for digicams. He’s actually arguing, no wait, the revolution isn’t over…one can only conclude he’s thinking about the high end for SLRs, where things have somewhat settled. The war is over, digital has won. Except for the most extreme large format work, and probably motion pictures, CCD’s have won.
But it’s the midrange where everything interesting is happening. This is where the vast bulk of photography occurs — how many megapixels is that photo you didn’t take, because your camera was too big and, well, gaudy to carry around?
Put simply, there’s a revolution coming to candid photos: No more blur, no more flash.
It can’t come fast enough.
We’re already seeing the beginnings of this realignment, with Sony, Panasonic, and a number of other vendors picking up on optical stabilization in either the lens or the chip assembly. By moving at precisely the opposite rate as your jittering hand, they can cancel out quite a bit of drift presently hitting the sensor. Optical stabilization would be nice, but I prefer Fuji’s solution to the problem: Make a much, much more sensitive CCD. Fuji’s F series is able to take excellent, low noise photos at ISO 800, and can do a reasonable job at ISO 1600. It’s not yet good enough for the nightlife, but for even moderately well-lit indoor environments, flash becomes unnecessary.
Very nice. But we can do better.
Before I got into security, I played around quite a bit with computer graphics — even wrote a photoshop filter! So I attend SIGGRAPH whenever I can, and always track the latest papers coming out of there. Without a doubt, the most exciting hack I saw out of them this year went by the name of “Coded Exposure Photography: Motion Deblurring using Fluttered Shutter” Put simply, suppose you find yourself taking a picture without enough light. Should your camera:
1) Take the photo with the shutter open longer, letting things get blurry, or
2) Take the photo with the shutter open not long enough, making everything dark and noisy?
Well, this hack adds option #3:
3) Keep the shutter open longer, but have a second shutter in front of it, flickering pseudorandomly at very high speed. This flicker will convert what was once a smooth blur into a series of overlapped images…which is much easier to deblur in software.
It’s beautiful. Between this, and some really cool work extracting blur patterns from blurry photos(see Removing Camera Shake from a Single Photograph), the war is on against blur, and by extension, flashes.
Watch. Sometime in the next eighteen months, the megapixel war will end (16? 24? 500? Who cares?) and we’ll actually start seeing stats on things that a) vary significantly and b) actually affect photo quality: ISO ratings, noise levels, etc.
Honestly, I’m waiting for the digicam industry to go this route, just to show off how bad the cell phone cameras really are (see, SLR guys? Point and Shoots have even stolen your superiority complex!).
Now, will dynamic range take off? This I think is a possibility. People are really getting into HDR (High Dynamic Range) photography; tools like HDRShop make it much easier to composite images taken at different exposure levels and merge them together into one, tone-mapped, hyperreal mix. What I suspect might happen is that some camera manufacturer will cobble together a photo mode that takes photos with and without a Flash, then saves the combination for offline processing (perhaps by MSR’s FlashNoFlash code?)
Of course, in a dream world, we could non-destructively read a CCD, meaning we could sample the sensor after a thirtieth of a second, a tenth of a second, a half a second…
HDR images don’t actually look particularly real — it’s an effect of the tonemapping. But reality is overrated 🙂 Some rather cool stuff coming. Just…500mpix? There are more interesting things to build.
(And yes, I’ll be writing quite a bit more now.)