Thursday, December 18, 2008

The case against DNG?

In a post today on his excellent blog "Serious Compacts," Amin describes how Adobe Camera Raw and Lightroom "bake in" lens corrections for Panasonic LX3 and G1 raw files when converting to DNG: Panasonic and Adobe - The Case Against DNG

Amin suggests that this is an argument against the viability of DNG as an archival format, but there's something sort of baby-and-bath-water about that argument to me. As I understand it, DNG is an attempt to create a standardized raw format. A standardized raw format would obviously be a boon to photographers, both for archival reasons, and because it would avoid the gap between the time when a new camera is released, and the time at which your favorite raw processor supports the camera's files. It would be a benefit to software developers as well, since they would not have to constantly scramble to reverse-engineer an unending tide of proprietary raw formats. It is, frankly, something we need.

No, this is not an argument against doing DNG. It's an argument against doing DNG wrong. Adobe and Panasonic, may I have your attention please: we shoot raw because we want to maintain as much control as possible. We use your software because we want our edits to be non-destructive. I know you know this, but it looks like you need reminding. Don't use vanity (oops, sorry our lens isn't so good after all) as a reason to undermine the very features that attract us to your products.

OK, that was the important part, but let me ramble on for a moment. Let's assume Panasonic went to Adobe and said, "Bake in lens corrections for LX3 raw files, or we won't let you put support for the files into your software." What should Adobe do? I'm not suggesting that Adobe should have said, "OK Panny, screw off then, no LX3 support in ACR." But I am suggesting that Adobe is a big grown up gorilla with plenty of weight to throw around, and I bet if they really tried, they could work something out. I'm unconvinced by Eric Chan's finger-pointing. Adobe has a responsibility to their customers, and if they want their DNG format to succeed it's even more important for them to make it appealing to customers than it is for them to make it appealing to manufactures.

But what if Panasonic didn't come around making demands? What if they came over all nice-like, and said, "Listen, we really paid a price for that tiny little 24mm lens, and we've got barrel distortion like you wouldn't believe. What can you do for us?" Well... what could Adobe do for them? Neither ACR nor Lightroom offer geometric lens corrections. What could Adobe do for you as an LX3 owner? Shrug their shoulders and make you shoot JPEG? Or use SilkyPix? Or live with crappy looking distortion? Or do your corrections in Photoshop (which then bakes in the changes of course)? Or make you wait six to 12 months for a full fledged lens correction feature, complete with all attendant UI changes, user guide changes, etc.?

I would suggest that there's a less sucky option. Adobe could special-case LX3 files and apply a non-user-accessible level of lens correction to the files on the way in (as indeed they have). Then, some time down the road, when they finally get around to actually putting lens correction features into ACR and Lightroom, those changes to your LX3 files would suddenly be accessible. You'd see the "barrel distortion" slider already set to 127 (or whatever) for all your LX3 files, and you could then back it off if you so wished. How about it Adobe? Any chance that these behind-the-scenes adjustments will become user-accessible in a future revision? Is it too late for that?

(P.S. to Panasonic: I like a wide angle lens as much as the next guy, but if you can't make a wide angle lens that's genuinely good, maybe software trickery isn't the answer? What's the problem here? Is physics getting in your way?)

Sunday, December 14, 2008

Limits breed creativity...

...or at least they force you to think. I recently purchased a Sigma DP1 compact digital camera. At last count there were approximately 137,924 reviews of the DP1 online, and I don't intend to write another. But I am going to say a few words about why I wanted this camera: it has lovely, lovely picture quality, and everything else about it sucks.

Granted, "sucking" might not immediately make sense as a reason for purchase. And I am perhaps being somewhat glib. First of all, not everything else about it sucks. It looks great (and I'm not ashamed to admit that's important to me), it's built like a tank, and it allows full manual control. Everything else sucks.

How could sucking possibly be good? Well, it isn't. Frankly, I'd be happier if Sigma addressed the myriad flaws in a future revision. But here's how I was able to turn this into a reason for purchase. The camera I was using prior to the Sigma (and which, frankly, I still use in addition to the Sigma) was the Contax U4R. This was essentially the last version of the Kyocera SL300R (see my previous post). It's really small, it literally fits in my shirt pocket. And it's faaaaaaaast. I pull it out, hit the shutter release, and it fires off pictures at machine gun speed. What's not to like?

What's not to like is that I wasn't thinking about my photography any more. This is the sort of picture I was taking:

It's a picture of a funny sign! Everyone loves funny signs! What's wrong with it? Well, nothing's wrong with it, but it's not the sort of thing I'd be likely to print out and hang on the wall. My photography was becoming almost exclusively an offhanded document of random things throughout my day. That has a certain appeal, but I also found myself wanting to get back into making "fine art photography," for lack of a less pretentious term. I wanted to make pictures that would compel me to spend printer ink.

My friend Eric (an engineer on the Adobe Lightroom team) once said, "Most cameras are better than most photographers." I couldn't agree more, and I would never blame my tools. But I do think that the nature of our tools guides us towards a particular sort of output, and I wanted a camera that would force me to think, because I obviously wasn't going to think on my own if I didn't have to. I saw that McCain sign, pulled out the camera, turned it on, zoomed in, and shot off 14 exposures: elapsed time, literally about six seconds. Just turning on the Sigma takes practically that long.

Slowness is not a virtue in a camera. But for the sort of photography I'm hoping to do more of, thinking is a virtue, and everything about the Simga forces me to think and to make conscious choices. The fixed focal length lens imposes limits on composition which force me to think before I push the button. The fact that once I've taken a picture it's going to be a while before I can take another forces me to think before I push the button. The fact that the camera produces better results on full manual than it does on automatic forces me to think before I push the button. Being forced to think means that now I'm taking pictures like this:

No one's offered me a Pulitzer, but I'll tell ya this: it looks good on my wall.

Wednesday, December 10, 2008

What have the camera companies been doing with their time?

I'm going to imagine a camera for a moment, will you picture it with me? I like to be able to carry a nice camera with me wherever I go, so this camera we're picturing has to be small, really truly pocket sized. Just big enough to hold in the hand, and not much more than half an inch thick. The picture quality has to be good, otherwise why bother? I need to be able make decent looking prints at least up to 8x10. I need at least a standard 3x zoom. And of course I want it to be fast, very fast. Startup time of under a second. Shot to shot time of under a second. Shutter lag, including focusing, of under a second. And I want a continuous shooting mode that allows me to shoot at least 3 full res pictures per second until the memory card is full. I like to shoot video occasionally, at least 640x480 at 30 fps, and I want to be able to zoom while shooting. Oh, and even though it's a point and shoot, I like to have some manual control. Manual ISO and white balance are pretty standard of course, but I also want aperture priority mode, and a shutter priority mode that allows long, multi-second exposures for cool night shots. Oh, and manual focusing.

Who am I trying to kid? A half inch thick point and shoot that fast? With decent picture quality? And seriously, manual focus? That's just being unreasonable, right? What am I trying to do, describe a fantasy camera from some magical utopia full of ponies and rainbows? No. I'm describing the Kyocera SL300R, and it was released in September of 2003.

Which I think begs the question: what the heck have the camera companies been doing with their time for the past five years? Digital Photography Review just posted a comparison review of nine current ultra compacts, you should read it. Go on, I'll wait. Guess what? Only one of the nine cameras is as small as that Kyocera. None of them are as fast, not even close. They all allow shooting video at 640x480 (one of them at 720p), but they don't let you zoom. Manual controls? Forget about it. Picture quality? No better, often worse. Oh, sure, most of these cameras have 10 megapixels instead of the Kyocera's 3.3, but to what end? The chips are the same physical size, and the increased pixel density means that noise, and noise reduction artifacts, eliminate any potential resolution benefit. That, combined with generally crappy folded optics prone to smearing and fringing, means that these cameras, just like the five year old Kyocera, can produce decent prints only up to about 8x10. How's the modern crop of cameras doing so far? Size: worse. Speed: worse. Video: worse. Manual controls: worse. Zoom range: no better (in most cases). Battery life: no better. Picture quality: no better.

So again I ask you, what have they all been doing with their time? Sure, they've been pixel-packing, I've no doubt that's a difficult engineering task. But since its detriments outweigh its benefits, it seems like a dubious enterprise at best. They've invented an awful lot of clever stuff like face tracking and smile detection. But why, exactly? Not even my mom wants smile detection, and no one wants to avoid having to think about her camera more than my mom. In a careful reading of that review of nine state-of-the-art ultra compacts, I can find only one single genuine improvement over the Kyocera, and that's optical image stabilization (which only five of the nine cameras have).

Or do they know something I don't? After all, those companies are all still making cameras, and Kyocera isn't. Are people that impressed by a big megapixel number? Are gimmicky features really that much more important than size, speed, and picture quality? Answers on a 3x5 postcard please. Or in the comments.

(P.S. I didn't even mention the coolest feature of that Kyocera: a rotating lens. No one makes rotating lens cameras any more, not even Nikon, and I demand to know why. But that's a subject for another post.)