Monthly Archives: September 2009

On the BART with Walker

Photographs

The modern Leica at work.


On the Bart. G1, 30mm, f/5.3, 1/20, ISO 800

The Bay Area Rapid Transit (San Francisco subway/tube/metro) was rocking and rolling mightily when this was snapped, suggesting that Panasonic is onto something with its OIS anti-shake technology – check the shutter speed, above. Don’t be expecting this feature in a Leica until the M14 comes out three decades hence, at which time the economy will have recovered and some dumb ass banker will lend you $100k to buy the thing.

By the way, the auto white balance in the G1 is every bit as good as that in the Canon 5D, which is to say it’s awful. I had to mess with the color sliders in Lightroom 2 to get a semblance of accurate skin tone in this snap.

Doubtless Walker Evans would have killed for this sort of technology, not to mention color, when he snapped this one back in 1938 as part of his subway series:

Evans used a camera concealed in his coat triggered with a remote control. No such subterfuge was called for in my picture.

Experts and pictures

Cameras

Trust your eyes, not charts.

Frequently one sees performance of digital cameras compared using JPG images. This could not be more useless. Different sharpening algorithms in the cameras being compared render such comparisons meaningless, compounded by the poor dynamic range of the JPG file format. JPG is the fastest way of making a fifty dollar camera out of a thousand dollar one. Add to this technical ‘analysis’ the fact that different lenses at different price points are being used and what you end up is so much noise.

Some try to counter these issues by comparing unprocessed RAW image output. Even worse.

All digital images require anti-aliasing – the process of removing ‘jaggies’ from image details. For example, when the Leica M8 was first ‘tested’ reviewers gushed over the native definition of the RAW images produced by the camera, comparing these favorably with Canon, Pentax and Nikon DSLRs’ output. Chalk and cheese. The Japanese cameras use a strong anti-aliasing filter, placed in front of the sensor. The Leica M8 (and M9) uses none, preferring in-camera software to do the same thing to images rendered with its previous generation Kodak sensor. To get the best out of the Japanese cameras’ RAW images substantial sharpening must be applied in Lightroom (or whatever you use) to get the best from the original file and make a proper comparison.

Reviewers would have it that this makes the Leica image superior as no sharpening is needed, conveniently disregarding the fact that sharpening has been done in camera. They will then compound this misinformation by comparing a fixed focus lens on the Leica (because that’s what they were given to test and are too lazy/ignorant/conflicted to do it right) to a lower quality zoom on the Japanese competitor.

As I have pointed out many times here, the 5D I use – depending on the lens used – mostly needs a sharpening setting of 40-60 in LR. This works fine for the 24-105, 85/1.8, 100 macro and 400/5.6. The 200 f/2.8 ‘L’ is so sharp that it needs a setting of 20 at most. The 20.8 and 50/1.4 I owned (both sold) needed 70 or more at wide apertures. The Panasonic G1 uses so little in-camera RAW processing that it needs even more sharpening – my default import setting for the G1 with the kit lens in LR is 90.

So what’s the right answer to the question : “Is this camera capable of delivering the resolution/speed/handling/whatever that I need?” I’m afraid that, with the host of variables in the digital process, and the broad range of skill sets and needs among users, the only right answer is “Try it and see”.

After all, would you trust a surgeon who never operated before to work on you? Because that’s what these reviewers are – hardly a one of them can take a picture to save his life and what works for you will differ from what works for the guy working for ‘click-through’ advertising dollars using free gear from the manufacturer. He’s not about to bite the hand that feeds him.

So for the reader who asked here the other day “Should I sell my LX-3 and get a G1?” there’s no right answer to that question. How can the needs of, say, a street snapper, compare with those of a macro or sports enthusiast? Depends what you want to use the camera for – there are no Swiss Army Knife solutions.

By all means read the technical reviews to get a sense of the gear but don’t waste your time on comparisons with other makers’ equipment by those unqualified to pontificate. Try it yourself.


Expert camera reviewers. G1, 14mm, f/4, 1/2000, ISO 100

An online camera review reader having just finished yet another mind numbing twenty page special:

Public schooling

Photographs

Urban art at its worst.

Walk south along that chicest of San Francisco’s streets, Fillmore, and turn west on McAllister and you are suddenly in an area which feels like a concentration camp. Lined with housing estates for the poor, each comes with a healthy ration of locked gates, barbed wire and surveillance cameras.

Then, just when you though things could get no worse as you approach City Hall, something like this confronts you:


Lefty. G1, 35mm, f/5/5, 1/400, ISO 100

An appropriate tribute to one of the greatest government crimes of our times. Public schooling. This mural suggests that this chap finally cracked and set about himself with a heavy object, doing the students justice. Can’t say I blame him. Just imagine turning up to school only to be confronted with this daily ….

The Panasonic G1 and GH1 sensors

Micro Four-Thirds

Sensor design is still in its early days.

A friend of the blog, has forwarded a fascinating technical piece which addresses changes in digital sensor design and suggests the oligopoly held by Canon and Nikon, who account for 78% of the DSLR market, is likely to weaken.

Most significantly, the article provides evidence that the Panasonic sensor in the GH1 (the G1 with video added) is anything but the same as the one in the G1. I would hate to have to compete with a behemoth like Panasonic which can roll out the first workable Electronic View Finder in the G1 only to completely redesign its sensor for the GH1 released shortly thereafter.

I quote: “Upon receiving the DMC-GH1, we fully expected to see a fabrication process similar to the DMC-G1, but partnered with a new design to add HD video functionality. While still early in our analysis, we have been pleasantly surprised to see Panasonic switch fabs and radically re-design their fabrication process and pixel architecture. ”

By the way, the GH1’s sensor is oversize compared to the one in the G1 so that when the user changes aspect ratios (16:9, 3:2 or 4:3) the total pixels used stay much the same.

The article also addresses innovations by Samsung in the sensors used in the fine Samsung/Pentax DSLR cameras and suggests that Canon is sitting on its laurels for now. In fairness, I have to add that’s no bad thing given how wonderful the FF sensor in my 5D is, but I have never known complacency to be a winning strategy.

Click the chart for more.

For those who thought sensor designs had peaked this piece will reinforce the fact that we are still in the early stages of innovation which, coupled with the new breed of Electronic View Finders, will make the next decade a Golden Era for new and increasingly responsive camera designs. In fact I expect that a decade hence, the pentaprism and flapping mirror will have disappeared from all but the most basic cameras (and the Leica DSLR, of course), confined to Rube Goldberg’s (Heath Robinson’s for UK readers) garbage bin where they belong.


On the BART. The G1’s antiquated sensor will do fine for now!
G1, 31mm, f/5.4, 1/80, ISO 800