2.18.2010

How would I design a perfect camera for me?


When I was a kid fast cars were king in the U.S. and we spent a lot of time with notebook paper and Bic pens drawing fantasy cars.  What would our ultimate car be like?  Of course there would be hood scoops and spoilers but also big fat tires and engines that virtually bulged thru the hoods like steroidial biceps.

Then we got all grown up and discovered "trade offs".  We'd trade acceleration for fuel efficiency.  We'd trade hauling capacity for aerodynamics. And we'd trade cool for value or reliability.  Now we drive Hondas and Toyotas and the errant Ford or BMW.  No Super Hemi Charger SS stud cars.

In early days of digital we dreamed about our ultimate cameras. They would have "at least" 6 megapixels!!!!  Would shoot at least 2 frames per second!!!!  They would have very clean files right up to ISO 400!!!!!  And they would write at least 8 big raw files to the memory card before you "hit" the buffer and hit the wall of waiting.

I was putting together my current stuff for an event shoot this weekend and I started thinking about what I want out of a camera today.  Here's the basics:  I want 12 megapixels and I want them at 5 shots per second in raw.  We've already got that.  I'm set there.  I want more that 18 raw files in a row without a slow down or hiccup.  We've already got that.  If you need more you're shooting something so frenetic and weird that the rest of us don't even know about it.  I want a great finder and a great rear LCD that I can swivel around and look at from many angles.  Okay.  My Olympus e3 and e30 both do pretty much all of this without breaking a sweat.  For nearly all of my work I need a good, solid ISO 200-800 and I've got that from just about any DSLR I pick up.

So what do I want that isn't out there?  This will sound strange but I'm ready for electronic viewfinders.  I was an early adopter with the Sony R1 but the performance was nowhere near convincing.  It was the Olympus EP2 that changed my mind.  The overwhelming feature?  Being able to set a specific aspect ratio and see exactly that ratio in the finder, edged by black.  I'm partial to the square but occasionally I'll set a 16:9 ratio for images destined for hi-def monitors.  Wonderful. And it's also wonderful that the finder matches the output to such an exacting degree.  It's also great to see a histogram or a quick, hi-mag view live on the screen.  I'll go out on a limb and say that this is the future.  In five years all camera manufacturers will have abandoned prism finders and implemented very good EVF's.  It will go hand in hand with the inexorable drop in DSLR prices.  And we'll love em.

I want a bigger EVF in a body like the current Olympus e3.  Maybe we could make that body a little smaller and lighter as well.

If we take out the pentaprism couldn't we also do away with the moving mirror?  That would reduce the number of moving parts in a DSLR by a huge amount.  All that would be left is the aperture stop down mechanism and the shutter.  So cameras would be lighter and more reliable.  Not bad.  And not having to charge a mirror would also save on battery power to offset the increased use by the EVF.

The next step (and it already exists) is purely electronic shutters.  No moving curtains.  The advantages are twofold:  Fewer moving parts and faster flash syncs.  At this point the only moving part of the camera body would be the in body image stabilization.  Just about nothing to go wrong mechanically.

So now the cameras would be smaller, lighter, less expensive and more reliable.  Not bad.  Not macho but not bad.

Starting to sound like a tool that's becoming transparent.  Almost invisible.
At that point we can turn the prowess of engineering to creating lenses that are smaller and lighter but have as good or better performance.  Really,  if you could have all the performance of a Nikon D3 in body that weighed three quarters less but had all the peformance parameters you needed and was 20% of the price wouldn't that be good.  Let's face it, the barriers are gone anyway.  Why continue to carry around the heavy and bulky legacy of the mustache wax days?

On another note:  I think I figured out how the medium format camera manufacturers screwed up.  We saw these cameras as a replacement for our studio (read:  moveable standard 4x5 and 8x10 cameras) but we only looked at the resolution.  In the ten years before MF digital came into existence we were well past the need to use big cameras just for resolution.  The real advantage of the view cameras that the MF's replaced (and which is rarely duplicated) is the ability to use movements.  To shift, tilt, raise and lower, and swing each standard, independently.  By doing so we had total control over product/subject geometry and also distribution of focus.  Everything else about large format film cameras was a red herring.

While Nikon and Canon are on the right track with their tilt and shift lens I'm hoping that someone comes out with a modular camera which is small and affordable (anything over 8 megs.....) and can do actual tilts and swings and shifts.  The size of the sensor is immaterial.  In fact it should be easier to make longer TS lenses for smaller formats.  Once they have that licked they can start working on tilting and shifting zoom lenses.  With 12 to 20 megs and full movements we will have achieved what we already had twenty years ago.  It's all about control.

16 comments:

Neal Thorley said...

Some really good points there.
It's surprising really that hasselblad or mamiya haven't come out with a "LF style" bellows and movements body, that takes the standard film and digital backs from their medium format range. It would have to be a lot more portable and downsized a little from your standard view camera but wouldn't be a bad idea.

just a thought.

Enjoying the posts, keep em coming!

iern said...

..I am still wondering what is stopping DSLR manufacturers from implementing tilt/shift sensor in their cameras...imagine a camera which makes all the legacy glass on it behave similar to large format camera with back standard movements..
on that note I am also kinda surprised that we don't have more tilt/shift mechanical adapters for legacy 35mm glass use on APS-C and smaller sensor cameras. In particular, m4/3 seems like the ideal format for using both tilt-shift sensor and tilt-shift mechanical adapters for legacy glass.
As always, a fun read from Kirk, that got my brain moving sideways :D

Herman said...

Neal: Fuji has one. For 6x8 format.

Tex Andrews said...

I agree entirely concerning the matter of "transparency" of the tool. I think one of the reasons traditional visual arts media survive is that their inherent simplicity makes them more transparent. Now, DSLR's are hardly simple devices, but they have started to get to a point that user customization can make the device nearly transparent in the hands in practice. For my perfect camera, I someday hope to see a DSLR or non-mirror rangefinder-esque camera that is simpler---not dumbed down, though, but high quality-- and affordable. And yes, like you I look forward to a day when movements could be more easily brought into play with digital at an affordable level.

Pierre-Yves St-Onge said...

Coming from a video background I totally understand the want for EVF. Current implementation of liveview in my Canon body is great to see when my highligth blows, to fine focus without the need to recompose and to visualize how the scene render in B&W but it is unusable outside most of the time. EVF would solve this and, as an added bonus, we could use those advantages handeld with goood stability.

Mirrorless camera makes sense if EVF are adopted and is a better alternative for hybrid (photo+video) cameras. It should be easy to redesign a new lens mount with backward compatibility à la 4/3 Micro 4/3. Future gonna be exciting!

obakesan said...

I'm so glad some other photographer seems to see that LF was about movements. Focal plane control, its not just about increasing depth of field, its about controlling it. I feel that my 4x5 gives about the same image quality as a 5D, but I still use my 4x5 (despite the hassle) because it does things nothing else does.

depth of field - dare to be shallow.

Eric said...

I doubt that manufacturers will have totally abandoned the prism in 5 years. I do think small mirrorless cameras are going to do away with the entry level DSLR's, maybe even if 5 years. However, I bet the same people that buy D300's, 7D's, D700's, 5D's and K7's will still be buying that same kind of camera complete with prisms 5 years from now. The only problem is high end DSLR's like that only comprise a small percentage of total DSLR sales, so I wonder if companies will still care about the DSLR market if it's been reduced down to less than 5% while EVIL's account for 20+ percent like Panasonic predicts. I think they'll always be a niche for the optical view finder just like there is still a niche for rangefinders, but the days of the optical view finder as the top dog are coming to an end. EVF's just have way too much potential since is becoming apparent that still and video cameras are merging into one sector.

Marshall said...

Interesting points as usual.

For iern: I think that there are a number of limitations that make it challenging to look at a tilt-shift sensor, not least of which is that the cameras are pretty stuffed with components already. In addition, there's a limitation on the distance to the back of the lens, as well as a likely concern about the precision of return to "straight". It's easier to consider this for aps-c cameras, though, since there are already a lot of lenses (i.e., full-frame 35 lenses) which have coverage beyond the edges of the sensor, but otherwise lens coverage is an issue.

iern said...

Precisely because of the reason of coverage, I believe that 4/3 is better suited to tilt/shift sensors and lens adapters. You get a lot more of leeway with smaller sensor.
Cameras full of components...yes, but we are not talking about 30 degree off axis movements of the sensor, more like 5-8 degrees, which should not be too hard to achieve. Getting the sensor back to straight reliably and quickly is just an engineering problem, not really anything unsolvable ;)
It all becomes even easier if you lose the mirrorbox...

Rick Moore said...

Regarding the EVF: I have learned to embrace LiveView in my studio. I realize that is not an EVF but still carries some of the same idea.

Where I do not see EVF gaining any traction is with active sport shooters. As an example, I can not hold long glass stable enough for shutter drag panning without wrapping myself up in the camera and moving from the waist.

I am guessing that we are still going to need a prism and viewfinder as there are applications where the EVF does not work to satisfaction.

kirk tuck said...

But Rick, with an EVF like the one in the Olympus EP-2 you can cinch up right into the camera just as you would with a regular finder and shoot that way. The fix will be making the response time of the finder monitor system smooth enough and fast enough so that it is a virtual optical finder.

Side note. I'm sitting at the car dealership waiting for my car to be serviced. Why is there always one really loud, anti-empathic person sitting in the midst of a room full of people writing and reading quietly, having a loud, long conversation of her/his cellphone? It's so annoying. We've really used some technology to destroy our culture.....

Rick Moore said...

Kirk, I stand corrected. That EVF on the EP-2 with the viewfinder is very cool. Obviously I was speaking without knowledge in my post.

Can the EVF be programed to lighten only the view in the finder in a dark setting? I am thinking not to change the exposure but to change the brightness in the finder.

Now, the more I think about it, the more I think EVF is darn sweet.

Marshall said...

Just one? At this point, I consider myself lucky if it's not three, all trying to talk over each other and complaining that they can't hear the phone for some reason...

David Bateman said...

There are tilt and shift adapters for M43! Although not together. So for $150 from Fotodiox you can buy a shift adapter from Contax, Canon FD or Minolta MD/MC to M43.
Also ball tilt adapters work for nearly all lenses from an Italian company for 145 Euro, more will come I bet from other adapter companies.
I have e-mailed fotodiox asking for a Nikon to M43 shift, as for me with some Nikon glass this would be more useful. I encourage everyone here to do the same, please!

Mark V said...

I thought Hasselblad did come out with a bellows type body that emulated LF, the Flexbody. You can hang a digital back such as the Phase One P21 off that thing and have decent movements, no?

ginsbu said...

It's interesting that you talk of an EVF camera as becoming "transparent". As much as I welcome the advent of mirrorless systems and see the very real advantages of EVFs, my one reservation is that they don't feel at all "transparent" to me: I'm very much aware of the EVF interceding between me and my subject; it feels like I'm watching TV compared to the immediacy of a good OVF, especially in less that great light. That's why I'm drawn to Fuji's hybrid viewfinder approach, allowing you to choose between OVF and EVF benefits on the fly.

I am very much looking forward to the introduction of a compact and affordable camera will movements. I hope to see an APS-C body that puts a mechanism for movements in place of the mirror box, with FF lenses offering a sufficiently large image circle. The main problems are providing sufficiently fast AF with those lenses on a mirrorless camera and the need for new ultra-WA lenses. One could come close with mirrorless cameras and T/S adapters, but an integrated solution designed from the ground up for movements would presumably provide more comfortable and efficient handling. At the very least, I hope we'll see adapters offering both tilt and shift capability soon.

One cheap-ish option for WA shift capability that might be worth trying is to mount the Sigma 8-16 on a m4/3 body using a shift adapter. 8mm is still quite wide on 4/3, and DOF and resolution should still be quite generous even with the lens aperture stuck wide open. I'm not aware of anyone who's tried it though.