Is Full-Frame the Coming Thing?

In the wake of the introductions of Nikon's first two "full-frame" (FF) (~35mm sensor size) cameras, to add to Canon's two, the internet is awash in prognostications that "full frame is the coming thing" (or "the whole market is moving that way" or one of several other common phrasings).

And that might be true...unless it isn't. It's equally possible that 35mm-size (~24x36mm) sensors will end up as an historically momentary diversion, an evolutionary dead-end; or a constant but smallish niche. In which case(s), full-frame won't have been the coming thing.

So which is it?

Impossible to tell. I mean literally impossible: the uncertainty is epistemologically absolute. No one knows what the future is going to hold in this regard.

That doesn't stop it from being a fascinating question.

The way I see it, there are two aspects of FF that bolster the case in favor of its future. One is its remarkably persistent popularity. I don't think it's true that what the market wants, the market gets. But it's probably true that what it really really wants, it gets.

And so far, the market (or at least a large enough subset to absorb the current FF offerings at an enthusiastic pace) appears to really really want FF.

I confess I personally misread this four and five years ago or so. I don't shoot with long lenses, but I've been working for photo enthusiasts for twenty years, and for that whole time I've observed the lust of the typical amateur enthusiast for long lenses. Everybody always seemed to want bigger and faster superteles. What I misread, I now think, is that I always assumed amateurs liked long lenses because of the pictures they could make with them—that is, they wanted narrower and narrower angles of view, for critter and sports photography and for that essentially voyeuristic kind of shooting that's equivalent to spying on the neighbors with a telescope. All of which seemed to be styles of photography that were perennially popular.

I now think that that was essentially wrong. What amateurs most wanted long lenses for was not for the narrow angles of view, principally, but for the prestige and pride of ownership of owning really big, impressive, expensive lenses. I thought the smaller 4/3rds and APS-C sensors would be seen as a huge advantage by all those tele lovers out there. 4/3rds, for instance, means that you can achieve the same angle of view with a 300mm lens as you used to get with a 600mm on 35mm. Since 300mm lenses are smaller, lighter, faster, and cheaper than 600mm lenses, that's all good, right? Well, it would be if what you were after was a narrower angle of view for the same focal length. But of course if the appeal of the 600mm lens is precisely that it's big and heavy and exclusive and confers great status on its lucky, rich owner, then the 300mm lens is not better.

I overestimated the utilitarian aspect of long-lens popularity and underestimated the status aspect. (Come to think of it, this might qualify as a persistent error on my part: I consistently assume that photography enthusiasts are more interested in pictures than they in fact are. But never mind.)

Depth-of-field issues were similar. For a hundred years, photographers bitched about limited d.-o.-f. Part of the rationale of the famous Group ƒ/64 was widely understood to include front-to-back sharpness as an aesthetic goal. Not having to suffer limited d.-o.-f. was always part of the allure of faster films, artificial light, and smaller formats.

...Until APS-C, that is. Then, photographers started a chorus bitching about the exact opposite! "I can't achieve selective focus unless I have full-frame," yadda yadda yadda. Some of these complaints came from people who actually knew what they were talking about and actually meant what they said, but in general such comments make me want to knock heads together. Set the zoom a little longer or move in a bit, fer Pete's sake. Aperture and format size aren't the only means of controlling d.o.f. (To say that typical photographers don't understand d.o.f.—either theoretically or in practical terms—is a massive understatement. And that's never going to change.)

Grump, Jr.
But enough of that. What's the other appeal of FF? It's that it matches well with existing camera technology—and existing camera technology also appears to have a very entrenched appeal. I've opined elsewhere that one of the great disappointments of the digital age thus far is that, after all the creative furor of its cradle period, digital has settled right back down to where camera design was in 1990—to a norm of Wunderplastik SLRs and dinky point-and-shoots, the exceptions being few and far between (some of the exceptions being the same exceptions that existed then, at least in terms of form—even down to a virtual replica M6, only digital this time). I was expecting, and certainly hoping, that digital would open up whole new vistas in camera design, and create whole new categories of cameras. Some of them I've even defined myself (the "DMD," I mean), to little effect. That very much appeared to be starting to happen back in the early part of this decade. But then it fizzled. The market might or might not be infinite in its wisdom, but it sure as hell is conservative!

When considering the "FF" question, for instance, the fit with 35mm camera body styles and existing 35mm lenses is the only thing that answers this question: why not bigger than 24x36mm? If bigger is better, why stop at 35mm size? Without a horizontally-traversing filmstrip to accommodate, the sensor can theoretically be any shape. Square makes the most sense of all the rectangles as far as lens coverage is concerned. And the square has two other big advantages as far as camera body design goes: it eliminates the need for "vertical grip" positioning, with all the redundant controls that add expense to battery packs and pro bodies now; and that, in turn, opens up viewfinder options.

You could argue that when using legacy lenses and flange distances, a square mirror wouldn't clear if it were too big. Right, but then what prevented the 24x24 format? Nobody tried that. Apart from the mirror issue (and I'll be waiting for the post-reflex-mirror era, like some pundits are awaiting the post-Bayer era), is it really that much harder to make, say, a 30x30mm sensor than it is a 36x24mm one?

I guess I'm fundamentally a grump* when it comes to this question. The Nikon D700 is a wonderful development from a digital-sensor viewpoint: its sensor, which is 70% of the point of the camera, is clearly a beaut. Tried 'n' tested, ably spec'd, very well established as fodder for enthusiast lust, and state-of-the-art in several performance parameters. Let me make myself perfectly clear: not complaining. (Really, okay?) BUT...(there's always that but)...from a camera design standpoint, almost nothing has changed in a decade; the D700 is a digital F100.

So why isn't Olympus making a whole range of point-and-shoot, and pocket, and rangefinder cameras around the 4/3rds sensor by now? Whither Canon's pellicle mirror as a digital solution? Where is the luminance-only sensor? Where is the larger-than-FF integral DSLR? Where are the square-sensor cameras? Where's the DMD? (The DP-1, nice though it is, meets the DMD spec in only a couple of ways, and misses the concept by a country mile in several others.) Where's the digital TLR? (The serious one—yes, I know about the mini-Rollei clone.) Where's the RD-2? The pro cameras with good viewing-screen only VF? Why doesn't the E-3 look just like an E-1 (the latter a brilliant design, I thought)? Why couldn't the D700 have been the size and shape and body material of an FM2 (or, heck an S3), and more clearly a single-shot camera, so as not to compete with the D3?

Where's the creativity?

I hope it's happening, and just hidden in the dens and warrens of the camera company engineering departments. I'm sure those guys doodle at least as much as I do. Maybe someday a few newthink designs will see the light of day, without the "wise" market spanking 'em on the ass for not being the same ol' conservative same ol'.

Digital is obviously a lot more than just camera design. But, folks, camera design is stuck. Think rut, think sticky mud. Think hysterical screams of I will only buy exactly what I've bought before only better. Camera design is stuck. And only as long as it stays stuck will the concept of "full-frame" continue to make so much sense. 

Will that be forever? Okay, maybe. Will it be long enough for FF to dominate the whole DSLR market? Possibly, I guess. Will FF dominate over APS-C even if camera design stays in 1990 mode? Hmmph.

I'm sure you have your opinion. Of course, the answer is: nobody knows.

______________________

Mike

*David Vestal will always be the original grump. I'm just stealing his word.

Michael Reichmann's initial thoughts on the D700

Ken Rockwell on the future of Nikon FX and DX

Thom Hogan on Nikon's second FX body

ADDENDA: First of all, thanks for all the great comments (66 as I write this). I could have plucked any number of them out as "Featured Comments" but really, it's just fun to wade in and read them all. A few additional stray thoughts from me in no particular ranking of importance:

• I actually am not writing primarily about myself in the above post. Personally, I'd probably prefer FF, all other things being equal, simply because I've had problems with lenses in the digital era. Cameramakers in general are no longer updating standard medium-speed single-focal-length lenses, which are the type I prefer. I most naturally "see" in the 35mm-40mm-e range, and 28, 35, 40, 50, and 85mm-e is the extent of the range I use. There just aren't a lot of very good options in these types of lenses in reduced-sensor-size product lines that can really get the best out of the sensors. What do I want? Just a modern, optimized-for-digital and optimized-for-the-format 35mm ƒ/2 or 40mm ƒ/2 equivalent for either 4/3rds or APS-C, not too big. Easy enough to find in legacy versions for FF, virtually non-existent in digital versions for reduced-size sensors. (A few things come close.) What's more, the prospects of getting them in the future don't appear to be good at all, either. So I'd be better off in FF digital than I am in APS-C digital. (BTW, the combination of the D700 and the Zeiss 35mm ƒ/2 ZF lens is going to be killer, mark my words.)

• Related to the above: again, something I've written about in the past: as an analyst, my primary concern is not "what would be profitable for the cameramaking companies," but "what would be useful for photographers to do good work with." So sometimes when I hanker for specific products, I'm not actually suggesting that such products would make good business sense, necessarily. Just that they'd be useful to photographers. A different concept, I admit. Hey, I have my loyalties.

• It's true that there is something historically almost magical about 36x24mm. There have been numerous attempts to move away from it in both directions, especially smaller, and no matter how much "sense" they might make, 36x24mm just forges ever onward, continuing to exert its magic pull. Ironically, the tiny sensors used in digicams are probably the most robust departure from 35mm size in more than half a century. But maybe the resilience of "FF" is due in part to the fact that it really is a great compromise size, very versatile, a sort of convergence "golden mean" when many competing factors are all considered together.

• I also can see peoples' point about the form-factor of the 35mm SLR being a highly developed and well-optimized basic design. That doesn't really change my position on two important issues, both of which I've written about before more than once: 1) that there is currently a "fashion" that dictates that good cameras have to be big and small cameras have to be compromised (i.e., "entry-level," cheapened), and I'd love to have options that break this tyranny, even if the basic fashion continues unabated—specifically at least a few (more) choices of really high-quality, premium cameras that are small (the M8 does qualify). 2) I really do think that some more creative options would be wonderful to have. Specifically:

• The "DMD." I need to revisit this topic on TOP, but basically, a camera the size and shape of a Canonet with, say, a 4/3rds sensor and a fixed 20mm ƒ/2 lens (or an APS-C sensor and a 25mm ƒ/2 lens), made to be highly responsive for single-shot use. This has been, is, and will continue to be a no-brainer. Something like this is needed by many photographers. I'm frankly astonished it still doesn't exist. Cameramakers have so far made one attempt: the Sigma DP-1. Kudos to them, although the actual camera Sigma made was something quite different than I describe. The refusal of all other companies to make something like this doubtless reflects factors I'm not privy to, probably led by the belief that such a product would not be profitable. We still need it. I concede that it's possible that there are technical impediments, although I find that hard to believe. If you can believe this, I actually proposed to Apple that it make such a camera and market it as the "iCamera." It fits their pared-to-essentials aesthetic. No response, but it was interesting to see that several other people suggested that Apple should make a camera.
• The Coolpix 950 was a nifty design. Why is it gone? In general, the digital camera market hasn't even been true to the best designs of its own incunable period.
  
• Picture an intelligent melding of the Sony F-717 and R-1, with a square sensor and IS. Beautiful.
• Luminance only (i.e., B&W): Another topic I visit and revisit. We had it briefly with the Kodak DCS 760m, and reportedly there was going to be a B&W version of the Leica M8, which was killed by that executive they later fired. There really are significant technical advantages to having a luminance-recording sensor as opposed to converting color Bayer arrays back to B&W. Maybe this would not be profitable or even viable from a business standpoint, but it's a type of product that a subset of the photographic community would find useful and desirable.
• Alternative rangefinders to the M8. It's great that the M8 exists (even though I'm not a particular fan personally), so perhaps we should be grateful for what we have. Given its popularity, it's certainly a great deal better than nothing.
• There's one more camera I think we really could use that we don't have, but I need to write a whole post about it.
• ...And I'm sure other people have other ideas. But the point is that there ought to be room for a few things other than the ubiquitous Wunderplastik SLR. Even if said Wunderplastik SLR is "perfect" and doesn't need to be improved, itself.

• An idle final thought, something I touched upon in the main post. For several years, people have been saying (into their beer, mainly) that it would be nice if Nikon made a digital FM3a. Wouldn't that have made a great form-factor for the D3 sensor in a smaller body? (Even if it were polycarbonate and not metal.) Forget high frame-rate, forget the battery pack, forget all the AF wizardry and focus tracking and all that pro stuff. People who want that could just buy the D3. Just a manual-focusing, single shot little SLR with a great viewfinder and the FF, 12.1-MP sensor from the D3 for people who want the image quality. $2k instead of $3k. And it wouldn't have poached sales from the D3 one iota.

...And it would probably have sold 1/10th as much as the D700 will sell, which would be bad except for the fact that the people who bought it would probably have liked it 10 times as much.

You can see why I'm not a camera company executive. —MJ   

Comments

[Bill Mitchell]

The 35mm camera was so successful because it fit perfectly to the size of human users.
Much smaller (Pentax 110) and it becomes too petite to handle comfortably -- the tiny little lenses are dinky.
Much bigger (Pentax 67) and the interchangable lenses are too large and heavy for most of us to carry around and handle.
Like the three bears' porrage, chair, and bed, 35mm size is "just right."
(Has no one noticed that the reduced-size DSLRs are just a big and heavy as standard 35mm cameras?)

[John Skillman]

The coming trend should be towards 'computational photography'. Multiple, small, combination sensors/lenses. The results of combining multiple views of the same scene can offer stunning dynamic range and even 3-D. I agree with your assessment of 'tradition' in the camera market and would also argue that engineers want to make bigger sensors and lenses because it's a bigger challenge than making one superb lens/sensor combo. and then differentiating cameras by the number of sensors included.
I have come to the conclusion that the next 'innovative' digital camera is going to come from a non-traditional player like Adobe, Apple, Sharp, Nokia, AMD; or a camera company with nothing to lose like Kodak.
I was hoping Pentax might do the DMD, but their corporate overlords Hoya probably want to sell more optical glass in over-sized lenses and Samsung probably wants to make large sensors for the reasons above, and to sell to camera manufacturers who don't have their own sensors fabs. I expect we will continue to see bigger and more over-spec'd traditional designs for the forseeable future. I don't blame them because the costs of marketing innovative are massive and extremely risky. The entire existing infrastructure would have to change significantly. It's too bad...

[Eric Ford]

This must be the third time I've read your telephoto theory, and it always cracks me up. The longest focal length I can tolerate using regularly is about 105mm, but 85 is plenty. This is one of the reasons I want full-frame: my 35, 50, 85 and 105 are none of the above. Unless I buy a $1700 uber-zoom, wider than 35 is Compromise Land.

Re: depth of field, I freely admit that I can't fully grasp it, and when I see a website claiming to explain it once and for all, the numbers, charts, and physics jargon make my head spin and I give up at about the moment I see the words "airy disk." What I do know is that the DX format satisfies my thin DoF cravings.

I agree with your feelings about the development of digital. Why don't we have interchangeable sensors for DSLRs, for example? Then we could choose to use a monochrome sensor or a color one, a high sensitivity one, a high-res sensor, and when new ones are designed our camera bodes wouldn't need to be replaced... just talking about it makes me kind of bummed about the missed possibilities.

[Archer]

Urg. 135 format is, IMSIO (In My Semi-Informed Opinion) a bit of a dead-end. The argument against it coming down in cost is straight-up Moore's Law. Chips drop in price as the size drops. 135 sized sensors, by definition, will always be 24x36, and hence will only benefit from modest reductions in price. Furthermore, those same reductions will apply by a factor of x^2 to 4/3's, and x^1.5 for APS-C. Why do you think digicams use 1/2.5" as opposed to 1/1.8" sensors? 135 will always be at a significant cost disadvantage.

As for boring design, Amen Brother! SLR's are pretty silly from a mechanical perspective, but they do have the advantage of allowing very sophisticated TTL auto {focus,exposure}. I see two interesting ways to go. One is the video-still convergence, another being really clever use of rapid reading of subframes from the sensor to do clever/fast auto-stuff using the main imager. Oh, and if a sensel can be made small enough, I can see oversampled imagers that have different "full resolution" images depending on the f/ratio selected, eg, resolution selected to oversample the spot size of the lens.

Lastly, as to why cameras are so damned boring? That has an easy answer: Maitani retired.

[Joe]

Whew! You earlier asked for the best of TOP, and this is a best example for me, and many past bests like this as well. Thank you.

On topic, 90+% of consumers will pay up for "good enough", and digital has given them low-expense results as good or better than 35mm film, and easier to use. The self-contained tiny sensor point and shoot camera fits that bill.

Every larger sensor camera (APS-C variants and FF)have been designed for the higher priced, higher margined accessory, starting with the lens choices. FF may be on everyones lust list but I believe all current sensor sizes will continue to be made for many more years even as production costs fall.

As for new designs, traditional shapes imply value. 30 years of the digital personal computer are basically unchanged, and modeled on 30 more years of computer terminal interaction. A quick look around will find hundreds of like examples in everything we use.

[Dan Khong]

While it may be impossible at this point in time to tell whether full frame is the coming thing, the fact of the matter is that there is a vast accumulation of film based lenses out there waiting to mate with full frame cameras.

I suspect that the digital sensor technology may have reached a stage in which affordable full frame cameras is beginning to dawn.

Dan K.

[Tom V]

"I overestimated the utilitarian aspect of long-lens popularity and underestimated the status aspect."

Chicks dig guys with big Canons.

Or that's what we like to think....

[Dwight]

I'm the guy who has never voted for a winning presidential candidate. My camera choices seem to do about as well as my political choices.

It seems that there is a need for good cameras with SMALLER sensors rather than bigger. Ricoh's GRD gives me pictures with great DOF. Such pictures aren't possible with bigger sensors. Why doesn't someone make an interchangeable-lens rangefinder with a small sensor?

I've never understood the big gap between small sensors and mid-size. It seems that the DMD should have a sensor that's a just a little smaller than four thirds. They could make a square out of 4/3 by cutting off the long edge. The new format would be called "Three Quarters of Four Thirds"

[Bron Janulis]

Like you Mike, I have certain demands on my resources, that preclude extravagance. I've recently altered my approach; instead of a D200 or D300, I got a G9. Instead of upgrading to CS3, I bought Elements 6. FF is meaningless to me at this point; and the image quality out of my antique D100 is just fine, as I never print larger than what I can produce on a letter size printer. I even print less, and make DVD's that are shown on a 19" TV.

In general, FF has no meaning as I tend to shoot in a mid range, 35 -55 sort of stuff, so I'll save my money to be "wasted" on kids. As to the future, it will be here soon enough, and as historians, we can dissect the errors.

Viva APS-C :)

Bron

One minor point, aside from my own unimportance as to marketing, is that the best, smartest or most efficient technology does not always win; Apple vs Windows, Beta vs VHS, etc. etc.

[Bron Janulis]

P.S.

As to the big lens thing, well, duh! SUV's, too.

[Janne]

I do think larger-than-ff is coming, once sensor manufacturing gets cheaper still (which is a slow process). You already have it, in Leaf, Mamiya and Hasselblad medium format cameras, just not at a price a normal hobbyist can afford.

When the main components - the sensor - is expensive, the whole camera will be expensive. And to buy an expensive camera, buyers reasonably expect high-quality, high-performance components overall, not just from the sensor. So 35mm, and even more medium format is all geared towards the top-of-the-line, no cost spared-style equipment, which of course increases cost even more. But as we are sort of gradually seeing with 35mm sensors, the price is coming down, and veeerrryyy gradually do the rest of the package also come creeping down in ability, quality and price. At this rate we'd still be _at_least_ five years away from 35mm overtaking APS, and medium format overtaking 35mm in market segments. More like 8-10, if ever.

But this creeping downward in price is happening for APS as well, only faster. And while people expect a "real", "professional"-looking black DSLR at 70k yen (which, face it, is a rather large chunk of money for a camera, historically, and a fair amount of money, period), they'd be much more receptive to variations in design and reduction in capability if the price was a third of that.

If we see a shift where the high-end becomes 35mm - or medium-format drops enough in price to overtake 35mm altogether as the highest-end amateur gear - APS equipment will have become cheap enough to compare to film camera prices. And at those prices - think 20k yen for an APS body - there is plenty of room for experimentation.

[Ctein]

Dear Mike,

I think you're right about design being stuck and staying stuck for a good while.There are sound design and market reasons for that to happen. So, given that, I think the question of FF ascendency resolves to a different question.

If a pro FF kit ends up being not substantially more expensive nor bigger than, say, a quarter-scale (QS) sensor kit, then FF will be the dominant format. If a QS kit is substantially cheaper/smaller (let's say "substantial= factor-of-two" for argument's sake), then FF will eventually become a niche product and QS will dominate.

This is much the history of film. The dominant pro format moved from 8x10 to 4x5 to 120 rollfilm to 35mm, relegating the bigger formats to niches. It stopped at 35mm because there really was no profound gain in going smaller. Consequently, half-frame and the assorted subminiature formats also became niche items.

I picked QS somewhat arbitrarily. Maybe it'll be 1/3-scale. I dunno. I do know, for a fact, that QS sensors will be entirely capable of meeting most professional needs. That's not to say bigger isn't better, just that smaller can be good enough. Again, much as with film. Does anyone dispute that sheet film gives better quality than 120 roll and 120 roll than 35mm? Of course not. But cost/convenience trumps quality when the quality's good enough.

I do not have an opinion, though, on which way this will play out. I can happily argue either side of whether or not there will be a substantial difference in cost/convenience between FF and QS over the long run.

pax / Ctein

[Steve White]

Two reasons that I may seriously consider a D700 (or a D3X) - the 14-24 f/2.8, and the 24 PC-E. This is the first glass I've really, really wanted in quite awhile.

[Nicholas Von Staden]

I hope somebody is listen to you in the industry...your readers are shaking their heads.. YES...Try a survey on the blog on 10 things what we want in a camera and vote one each? It would be very popular I'am sure...let me see...........a shutter you can feel not hear....50 zone auto focus with manual override...Lens,a 28mm 1.4 all in the size of a canon G.. among other things.....like you said they're in a rut... with design by committee

[Bromley]

One point i disagree with (it's only natural i suppose!), but apart from that i agree with pretty much everything you wrote - you are certainly correct; camera design really is stuck in a rut. What a shame. You never know though, perhaps one day someone will make me that digital Mamiya 7, with a 6x7 sensor... Just for me.

(Just in case you're interested in which point i disagreed with: Yes, i know that you can move closer or zoom in more, in order to control depth of field - but that then means the composition and/or perspective has to change, and you are no longer taking the same photograph. For example, when making photographs with medium format cameras (eg. my mamiya 7) it's possible to subtly blur the background at modest distances from the subject with normal/wide lenses, which i feel really makes subjects stand out from their respective backgrounds, without losing any of the environmental information that i wanted to include. It is a characteristic of MF i really appreciate.)

[Adam Richardson]

The conservativeness is disappointing. If you looked at the last gen of 35mm film point-and-shoots and SLRs, and skipped over the period that brought us the articulating Nikons and the Sony F707 and similar (dead-end experiments) and looked at today's digital offerings, you'd think nothing had changed with the internals.

Remember how early cars looked like horse and buggies because the manufacturers hadn't grok'd what the new internals meant for the external form-factor and possibilities of arrangements it allowed? It's almost the opposite with cameras where there was the period of experimentation and now we've gone back to buggies.

Heck, 25 years ago Luigi Colani collaborated with Canon to conceptualize film-based cameras that are way-more radical than anything we are seeing today. (http://www.canon.com/camera-museum/design/kikaku/t90/03_5systems.html)

This exploration resulted in the Canon T90, which for better or for worse set the paradigm for all SLRs that have followed - Canon's current offerings look virtually the same (except less elegant IMO). On TOP you've referenced the post I did on this on my own blog quite a while back (http://www.richardsona.com/main/2006/9/29/design-classic-canon-t90-slr.html)

[Ken Tanaka]

"Is Full-Frame the Coming Thing?"

Two answers: No and Yes.

No, because full (35mm) frame sensors have been around for quite a while. I've been using full-frame cameras since the Canon 1Ds Mark II. So it's not a "coming thing"; it's a "now thing".

Yes, because it will be the marketing bogey that all of the secondary manufacturers will try to hit in coming years.

I don't think that camera design "creativity" will come from form sensor factor. The manufacturing barriers are too enormous, Nor will we see a big resurgence in mid-20th century camera designs-turned-digital, such as 35mm rangefinders. All who wanted ashore on that island are probably already ashore.

Rather, I believe that we'll see incremental innovations in sensor technology. Yes, it looks like the Foveon, everyone's former great white hope, is a flop. But there are other possibilities on the horizon.

For example, Sony recently announced a "breakthrough" in potential sensor dynamic range and clarity. They have reportedly developed a chip that captures an image on its back-side rather its lens-side. They claim that this new design offers a significant leap forward for an affordable price.

The real "new thing" on the immediate horizon may very well be HD video cameras designed specifically for dual-purposing. Canon already has a robust line on their shelves and also has immense video lens technology and experience. They may just decide that this is the technology weapon to renew their distance from those slow-footed-but-pesky Nikon people next year. (Nikon has zero video assets.)

So, Mike, I'd advise you to just party like it's 1999 with a new full-frame Nikon or Canon. What you see is fundamentally what you'll see in the photo market.

[Erez]

Mike,

I think it's at least the 3rd time I read an article you wrote about lens size. Yes, most amateurs like to show off with their photo equipment. Most also concentrate on what they can't do with the equipment they have. I have a feeling it's part of the culture we live in more than anything else.

Regarding the crop-sensor angle of view, I must say, I can't see the point. The fact is that the magnification stays the same, all you do is crop the photo. So basically if you had the same lens and take the same photo using crop sensor and FF sensor (using the same zoom value) you will get two identical photos, one is a subset of the other. So in essence a crop sensor doesn't enlarge the subject somehow, it just captures less of the scene, and as such not very compelling.

Erez.

[Aubrey]

Hi Mike,

I really enjoy your views on photography and this post is no exception.

I think that people want the best equipment when it comes to things that they are interested in whether it is cars, hifi or photography. Thats why people want Full Frame because it is perceived by most people to be better which is a carry over from the film days where Larger format equals better quality.

I agree with you that the camera companies have not been very creative when it comes to innovation and new formats in the digital world. Square format would be nice as I liked it when I had a Bronica 6x6 many years ago. I don't blame companies for not experimenting though as that might put their profits at risk if the change is not accepted by the customer.

I'd like to see HDR built into the camera as well as a non bayer sensor like they had in the Kodak 14n. There is still a lot of innovation that would improve the quality of the image if there was a company out there willing to take some risk.

Regards .......... Aubrey

[Jay]

Dear Mike,

There is one totally new direction for digital design, the camera phone. It does not interest me now but who knows about the future.

I have been using Leica M's for 43 years, so I'm set in my ways with an M8. But I do want a really good compact digital camera with very good high ISO. Say like the Hexar AF with its great 35mm lens. The lack of an optical VF is a deal breaker. A more compact one like the Contax T3 would also be desirable I tried the G9 and it does not meet my standards. Hell, even my old Canon QL17 GIII beats out the current compacts.

[R. W. Bloomer]

Maybe the 35mm SLR IS some kind of perfection: maybe the meaningful advances in camera design have long since taken place. Maybe what EVERYONE has secretly been wanting all along is a Nikon DSLR that lets Nikkor manual focus prime lenses do what they were built to do.

[Mike J.]

"The fact is that the magnification stays the same, all you do is crop the photo. So basically if you had the same lens and take the same photo using crop sensor and FF sensor (using the same zoom value) you will get two identical photos, one is a subset of the other. So in essence a crop sensor doesn't enlarge the subject somehow, it just captures less of the scene..."

I don't think this is a very good way to conceptualize what's going on. Lens focal length doesn't change magnification regardless of the format size, and perspective is the same, but lenses have to be designed for image circle, too. A smaller format only "crops" a larger format with the same lens if the lens can COVER the larger format. And then, what's the opposite of the concept of "cropping"? Because the fact is that the 300mm on 4/3rds has the same angle of view as a 600mm on 35mm, and a 300mm on 35mm "crops" less, in your terms. But then consider that a 300mm lens on 6x7cm is the same angle of view as a 150mm on 35mm (more or less) and the FF format "crops" it, and a 300mm on 8x10 is a NORMAL LENS.

Most people would not say that the 300mm lenses on an 8x10 and a 35mm are the same, except that the 35mm format crops the image. Rather, they would look at the more readily apparent property, angle of view, and say that the 300mm on 8x10 more closely resembles a 50mm lens on 35mm.

It's also not good conceptualizing because the 35mm lens can be designed for a much smaller image circle and it can be much brighter and higher in resolution on the smaller format. With great differences in format, very different designs can be used. A typical 300mm designed for 35mm would cast an image on 8x10 film about the size of, what, a tennis ball?

Conversely, you wouldn't say that, say, a 7.1mm lens from a digital point and shoot is the "same" as an ultra-wide lens for 35mm, except that it's "cropped," would you? No, because the digital p&s lens can't cover 35mm or even come close. So the "crop" concept when understanding format is really only is true when you are considering magnification ONLY, in isolation, ignoring several other very important aspects of lenses, most prominently angle of view on any particular format as well as coverage.

The "cropping" figure of speech isn't technically wrong, in certain ways and under certain conditions (although as a vocabulary word it should be reserved for cutting off the edges of captured images), but it's just not a very good way to think about all that's going on in the interaction between focal length and format.

Mike J.

[John Hicks]

Grump Jr. you're making my head hurt in the best tradition of the original Grump.

"digital has settled right back down to where camera design was in 1990—to a norm of Wunderplastik SLRs and dinky point-and-shoots, the exceptions being few and far between"

Absolutely correct; the megapixel war is comparable to the lenspeed war, the motorspeed war, the hexaphotocybernetics war etc.

We have a Technoplastik Wunderblaster known as a Canon Digital XT, a couple of near-misses, Canon G9s, plus the assorted old dinosaurs including the 8x10.

The XT pics are nice and clean with their odd (but perhaps correct) digital look while the G9 pics are a little gritty; otoh have we all forgotten Konica 3200 film? I guess so. Yes, the G9 viewfinder is the pits; you'd think Canon would do far better. But that's beside the point.

Full-frame. What's full-frame? 8x10? 2x2? Of course once "full-frame" cameras were regarded as mere toys. Maybe in fact they still are.

"camera design is stuck. Think rut, think sticky mud. Think hysterical screams of I will only buy exactly what I've bought before only better. Camera design is stuck."

Absolutely. It's the same-old same-old, just different imaging (gads I hate that word) mechanisms. Where's my square-format (digital) OM-4t? Where's my digital Noblex?

The mass market, of course, prevails.

"I now think that that was essentially wrong. What amateurs most wanted long lenses for was not for the narrow angles of view, principally, but for the prestige and pride of ownership of owning really big, impressive, expensive lenses."

Don't tell me you're just now learning this? In my camera-dealer days I could probably count on a finger or two the number of people who bought those big honkers who talked about usage. Not owning, but using them for a specific purpose. It's the gold-Rolex effect; essentially a way of saying "mine's bigger than yours."

Obviously we need another Maitani. Is he (or she) out there? And would a camera maker take the risk today?

I guess we'll see.

-jbh-

[JR]

I was very disappointed when Nikon did not produce a camera like Pentax, to simply accept all the Nikor lenses. Even better, would be a back with a sensor that could be attached to any camera, old or new, within reason- say since the 1970's.

I find that for most photo enthusiasts, making good photographs is not about fast lens, fast focusing, etc. So interchangeable backs would be a good way to go for everyone. I don't feel like handing out thousands of dollars each time there is a change in technology, however small. All I hear from these discussions is that people want pro digital cameras to be as versatile as point and shoot cameras, much the same way as the old discussions went about the relationship between the different formats in film cameras.

JR

[Michael Steinbach]

The "DMD" always makes me think what if they dropped an APS-C sensor into a Nikon 28TI body....

I was shooting a wedding two weeks ago and a guest was the local Nikon Rep, Mike. I asked why not? Mike's response: "I suggested it many times, but you just never know what is on there drawing board." With reference to what Michael Reichmann just wrote about Nikon's Board (see link)
http://www.luminous-landscape.com/reviews/cameras/700-thoughts.shtml Nikon could have many more surprises ahead.

[Chico Fernandez]

Dear Mike,

I've spent 40 years writing and photographing fly fishing. Years with a Leica M4-P and M3, and later Nikon F3s and FM-2s. And I never, ever, composed better than when I had my Mamiya 6. Something about the square format and the RF's bright line finder.

Anyway, I fully agree with your article. There is nothing new. The D700, exciting as it is, is just an F100. And when indeed are we going to see p&s cameras with larger sensors?

Presently I use 5D and L lenses for work, and G9 for play, and soon also for some work.

Anyway, well said, as always.

Chico F.

[John Camp]

I think the main driver for FF was legacy lenses, and the fact that most professionals had a pile of them -- and now the 35mm size that was adopted because of legacy lenses is driving production of a new generation of "digital performance" lenses that are sized to FF, so we are creating a new legacy at the 35mm size...

There is some talk of an upcoming Leica R10 that will have a larger-than-35 sensor; most people who believe this will happen (and the possibility has been mentioned by Leica officials) expect a sensor size that will have the largest real estate that will be decently covered by Leica R lenses -- not a 2:3 aspect ratio, but something closer to 3:4 or 4:5. There have also been suggestions that the R10 will be a medium format camera with a FF crop which will use R lenses, and a new batch of Leica lenses that will cover the full MF. That seems unlikely to me, because I'm not sure Leica could afford to do it, but Leica desperately needs something that will distinguish their unbelievably expensive cameras and lenses from the Canons and the Nikons. So that could be interesting. And most Leica fans expect to hear something by Photokina.

My son jokes about it ("I can't possibly come to dinner now -- somebody's made a MISTAKE on the Internet") but honest to God, reading a Digital Photography Review thread on DOF or ISO can unhinge a normal person. Especially weird are what car guys call "comparos" when people try to figure out if a Canon 5d (say) is better than a Nikon D2x, and they analyze tiny fractions of jpg photos taken by two different guys at two different times with slightly different focal-length and quality lenses, while in the background is the usual Greek chorus chanting, "This is useless. This means nothing."

The chorus is right, of course, but really, who cares?

JC

[Geoff Wittig]

The 35 mm form factor for cameras and sensors is here to stay, simply because the immense 'installed base' provides far too much inertia to overcome. The same principle controls the computer world. The crypt-keeper legacy of Intel's X86 architecture and associated archaic code keep desktop and laptop computers nailed to the floor, despite hardware advances that could provide so much more. (And you Mac lovers, stop smirking. Your machines are now also part of the Borg...I mean Intel...collective.) Sigh.

On the other hand, a reasonable argument can be made that the 35 mm form factor is a sort of optimal nodal point for camera design, and this 'gravitational pull' is difficult to escape. Going to a smaller sensor clearly imposes some limits on noise control and resolution, so if you insist on a really excellent 16x20" or larger print as final output, or want room for cropping, then 'full frame' 35 mm sized D-SLR's will be very hard to match. On the other hand, any larger sensor requires a bigger lens to cover it. Practical realities in precision optics manufacturing quickly escalate costs out of sight...unless you're willing to accept some optical compromises (i.e. higher levels of aberration) which then defeat the purpose of the larger sensor. The 35 mm size may indeed be Goldilocks' "just right".

Certainly more rational lens coverage could be achieved with a format closer to square than the 24x36mm standard. I sort of like the idea of pushing out the full frame sensor to a 36 mm square. This would maximally exploit the image circle of existing "35 mm" lenses without requiring a whole new hardware standard. Won't happen, though. Too much inertia.

[Paul De Zan]

"Maybe what EVERYONE has secretly been wanting all along is a Nikon DSLR that lets Nikkor manual focus prime lenses do what they were built to do."

I think that's it right there. To paraphrase Jacob Bronowski, there is nothing new in camera design because there is nothing new in photographer's thought. What I can't understand is why these "optimal" DSLRs are all so damn huge. Many think the Nikon D80 is an ideal size; I think it's a boat anchor. D40/D60-sized DSLRs ought to be the standard, not the exception. The physical bulk of these things is the one carryover from film cameras that I neither like nor understand; there's certainly no overwhelming technical reason for it.

[Fazal Majid]

It's amusing how fans of systems that don't have full-frame belittle the importance of it. Nikonians certainly did until the D3 came out.

Full-frame sensors offer the following advantages over APS-C or lower sensors:

1) They offer stellar high ISO performance, far exceeding that of film. When you see the dismal ISO 1600 performance of an Olympus E-420, you can understand why the Four-Thirds format is getting no traction.

2) Their viewfinders are big and bright, nothing like the nasty tunnels even a D300 feature. Sure, there is no reason why an APS camera couldn't have a big viewfinder, if its optical system were designed with a full-sized mirror. That would introduce flare, however, and bulk up the camera.

These two factors in themselves won't automatically translate into market dominance, but if the price premium falls to a reasonable level, most people will go for the more capable camera, except in the ultra low-end Rebel XS/D40 segment.

[Bill Bresler]

Re: The appeal of hellishly long and heavy lenses...
Years ago I was shooting an annual horse racing event with the longest lens the paper issued us snappers, a 300mm f4. I ran into a friend, shooting for one of the big dailies in town. He was shooting the race with a 600mm f4. It was a monster. I asked him how it felt to drag that heavy glass around. He said it didn't bother him, because whenever he shot with it, lots of women hit on him. I guess I developed a serious case of lens envy.

[Mike J.]

"And I never, ever, composed better than when I had my Mamiya 6. Something about the square format and the RF's bright line finder."

Chico,
I had a Mamiya 6 and remember it fondly. This is something you might remember...the Mamiya 6 was originally targeted at wedding photographers...that's who Mamiya figured would buy and use it. All the early ads were directed at wedding use. Lo and behold, it never caught on with wedding shooters, but editorial and magazine photographers loved it and it became quite a fad among them. So sometimes even the camera companies don't know who they're making things for...sometimes products find their own markets.

Mike J.

[McD]

Discovering the practicality and effectiveness of small sensor all-in-one world has radically changed my relationship with photography.

Why choose a DSLR cost and weight multiple of ten or more? Some professionals may need the features. I've learned that I don't.

The DSLR world is driven by insane profits at the moment, but I expect it to be eroded as the alternatives become more attractive and we become more practical. Think of the demise of medium format. Change can come quickly.

[David Dyer-Bennet]

Observing my own use of my 70-200 zoom across the transition from film to 1.5x digital, I find that I'm out at the 200mm end just about as often as before. That is, the lens previously wasn't actually long enough, it was just the best I had (of any convenience; the 300/2.8 weighed a ton and the 500/8 was kinda slow, plus they were both manual focus).

Looking at lens releases, we see the Sigma 50-150/2.8 (roughly a 70-200 equivalent for 1.5x), but we don't seem to see it flying off the shelves, or many other people copying it. In fact, we see *more* 70-300 consumer lenses than we used to. I see here more evidence that people wanted longer lenses than they mostly had on 35mm film.

Similarly, 50mm macro lenses were the norm, then the Nikon 60mm. Longer ones were also available in some lines (and Tamron had that lovely 90mm since forever). And then 105mm started becoming more popular. And then 1.5x DSLRs came in -- and 105mm kept right on taking over, and you started hearing about 180 or 200mm macro lenses.

I'd *love* to see some of the experimental bodies you mention. I definitely want the non-reflex interchangeable-lens large-sensor digital, and would be very interested in the monochrome one (that should be good for a couple of stops improvement in sensitivity, plus of course an increase in real resolving power).

[John Gillooly]

i am a cross platform person, but it seems like we need the engineers at Apple to attempt a camera. They seem to have the ability to think about what could be, as opposed to trying to tweak a current product to make it a bit better. it would be very interesting to see what would happen if very talented people went at it with no preconceived notions of what it should be.......

john

[Aaron]

As mentioned above, it seems like the next big thing will be something really new - not a digitization of an older design.

For example, twin lens reflex cameras have already lost a round to 35mm cameras - I'd think manufacturers would be rightfully leary of introducing a digital TLR. They were overtaken by 35mm film cameras - what would allow them to enlarge their market share in digital? That's not a rhetorical question; I'm genuinely curious if someone has a good answer.

[Jaakko]

Mike: "4/3rds, for instance, means that you can achieve the same angle of view with a 300mm lens as you used to get with a 600mm on 35mm. "

Same angle of view but not the same image quality. Very long end of focal lengths are all about reach which unfortunately is not achieved by cropping -> I would any day take 35mm equivalent with same pixel pitch to replace my 1DMKIII for long tele work if price and buffer speed would be the same ;)

Another thing is legacy of 35mm film SLR background. Once I have been earlier using SLR with 35mm and 50mm lenses, same time enjoying big viewfinder and nice oof, the 1.6x crop was alway feeling 'not right'.

My prediction is that prosumer/pro will eventually go towards full frame while there are still lot of use for crop cameras. For example Olympus E-420 + 25mm pancake is a lot more practical carry around than equivalent full frame dslr.

[erlik]

"digital TLR. They were overtaken by 35mm film cameras - what would allow them to enlarge their market share in digital?"

Hmm. How about now that LiveView broke the psychological barrier and isn't a "solution looking for a problem" anymore?

You could have perfect LiveView A/B with a TLR.

Only, in this age of instant gratification, I don't see many people learning to cope with the parallax error.

[Dave Brown]

First, wouldn't a major advantage of a full frame sensor be the availability of shorter focal lengths? I've heard (though not from authoritative sources) that there are physical limitations that make it difficult to make really wide lenses for small sensors. If this is true, then I would have an incentive to buy a full frame camera, but not at current prices. In a world where cameras become obsolete in ever shortening time spans, I just can't justify $3000.00 for a full frame camera body. Back in "the good old days" I could upgrade my camera myself simply by changing the film; now when new sensor technology is available, I have to buy a whole new camera to upgrade.

Second, as long as we're fantasizing, what I would really like to see is a sensor with random pixel distribution (as opposed to a grid pattern). Not only would this do away with Moir'e issues, but it would help eliminate some of that "home video" quality we often see in unprocessed digital images.

In the final analysis, cameras are only tools. It would be ideal if we could adapt them to the way each of us, as individuals, work. Since digital cameras are mass market items, that simply isn't going to happen. So I suppose that as always, we'll learn to live with, and occasionally exploit, the weaknesses of our equipment.

[Antonio Dias]

If one wants to use 35mm lenses from wide-angle to telephoto using their designed for image circle, a full-size 24x36mm sensor is required. The so-called "partial-frame sensor" telephoto advantage is no advantage at all.

[Peter F]

Praise be to you for questioning full frame dogma. And really, "dogma" is the right word here, because a lot of the chatter you hear around this subject sounds suspiciously religious (i.e. faith based). For a while I assumed everyone wanted FF because it will always have a noise advantage over smaller sensors. But when you look at the numbers... is that really enough?

What I'd like to see is a facts-based discussion about how much smaller and lighter bodies and lenses can be when designed around APS-C sensors (ahem, TOP?). I'll be disappointed if the industry abandons APS-C (or leaves it to the entry-level models), and I could totally see it happening because with the way silicon manufacturing scales, FF systems won't have to cost as much as they do today for very long, and when prices come down... well, like Mike says, you can't predict the future. But offer a guy a medium Coke for $1.00 and a super size Coke for $1.10, and you know which one he's going to pick.

[Paul Mc Cann]

Cameras with long lenses as photographic jewelery is a definite phenomenon but then consider how many of those little point and shoots have really extended telephoto zooms with only a label to proclaim the fact. One has to assume people want these long zooms. Can it all be the bikini-on-the-beach effect?
Saw a guy in a Spanish cathedral one day with a truly enormous long lens on his camera. But, hey, it was white.

[Grega]

I for one don't know which format I want. My longest and dearest lens now is 105mm and it is perfect for my needs on my D80. I don't need anything longer and I like this particular lens so much that I wouldn't want to kill it by using it on FF. On the other hand, I miss a wide prime on D80. So for me, I guess it is both formats or at least FF with enough resolution that makes a 1.5 crop reasonable (and at a reasonable d80-ish cost). I hope my current camera keeps working for the next 4 years or so.

I don't think larger than FF DSLRs (or at least considerably larger) are gonna happen. The lens mounts limit the size of the sensor and I think I speak for most of us if I say that we don't won't a change in the lens mount.

What I'd like to see is ability to change sensors in the camera. You could have a high res one, a low res one with less noise and a B&W one. When there is a considerably better sensor made, you buy the sensor only, stick it in the camera, do some firmware upgrading and done. But that's the old hey-new-film-is-out thinking.

[hugo solo]

Can people take seriously when you said I´m a photographer and you carry an Ixus in the photographer´s mind they need big cameras,from the end of the 80´nothing change much in design, the first flat screen tv was invented in the 50´s.

[fotoral]

It's the viewfinder.

[Hugh Crawford]

I think a big part of the demand for "full frame" 24x36mm is from photographers who used to use medium format and photographers who really like wide angle lenses.

Say you were formerly a medium format camera user with a two roll a day habit, and decided to switch to digital. A full frame DSLR seems cheap, like for free cheap, but a medium format back costs as much as a nice used car, and depreciates much faster.

There is/was an ocean of cheap medium to large format stuff out there that was affordable to a lot of photographers that can't afford the digital equivalent. Pre digital, my camera buying habits were along the lines ofvarious Autocords, Hasselblads, homemade lumps with a 47mm Super Angulon on one end and a graphic roll holder on the other, old press cameras, a 70mm combat graphic, a 4x5 RB graflex. And you could always sell them for about what you paid for them, although I still have one of all the cameras I mentioned other than the 47mm Super Angulon beasts.

The jump from a APS-C camera to a full frame with more than twice the image area is not that big a jump in price, but the jump to the next size larger is pretty painful.

My feeling is that "full frame" 24x36mm is the sweet spot for serious photographers that would have never considered spending serious money for a 35mm film camera.

BTW, my recipe for the ideal small digital camera would be a camera with a square sensor, say 30x30mm with an electronic shutter and a live view screen and a Leica M mount. No mirror, no rangefinder. The beauty of an M mount is that you can adapt just about any 35mm lens ever made to it. It would make a great SWC replacement.

Also, even though I prefer square formats, why hasn't anyone done a revolving back digital camera? It would be so much easier than revolving back film cameras were.

[Tom Dibble]

"Regarding the crop-sensor angle of view, I must say, I can't see the point. The fact is that the magnification stays the same, all you do is crop the photo. So basically if you had the same lens and take the same photo using crop sensor and FF sensor (using the same zoom value) you will get two identical photos, one is a subset of the other. So in essence a crop sensor doesn't enlarge the subject somehow, it just captures less of the scene, and as such not very compelling."

This is absolutely correct, but only meaningful in a limited sense. IF you are using the same lens on you APS-C as works on your FF, AND you have the same pixel density on your FF as you have on your APS-C, AND you have a significantly faster storage mechanism with a significantly larger capacity to accomodate those extra pixels, THEN the APS-C sensor is nothing but a pre-cropped FF sensor.

On the other hand, the APS-C allows you to break each of those rules.

1. The lenses can be smaller and lighter for equivalent edge-to-edge quality (especially note the vigneting and chromatic aberrations around the edged of average consumer FF lenses).

2. Sensor pitch is generally more dense on APS-C, specifically because they then do not need to tax the image processing systems.

3. At the same time, they end up with fewer pixels and thus still tax the image system less, allowing for significantly faster continuous speeds for significantly longer bursts for significantly less cost than is possible (today) in FF sensor cameras. This advantage will diminish in importance over time, of course.

IMHO, the main point against a square sensor is the same one against standard-aspect-ratio television sets (just try to buy one today ... I dare you!): people don't like feeling like some visible bit of technology they purchased is not being used. Oh, they'll buy a camera with a hundred features they do not understand and will never even attempt to use, but if to get a visually pleasing "golden ratio" print they need to cut out half of the square image, and they need to do this on every single picture they take ... I don't think they're ready for that.

On the other hand, I don't mind the "black bars" on my television sets (my last CRT-based set had a "widescreen" mode where it actually condensed the horizontal lines of the image so that the actual displayed resolution was the same with or without black bars), and already crop just about every single picture I take in post-processing. I also find it really agonizing cropping a landscape to a portrait in post simply because it is so rare I captured enough vertically to make it work, and a square captured aspect ratio would help with that tremendously. IMHO, the "ideal" camera would capture a square image, but automatically "crop" that to the central "golden rectangle" portion in either portrait or landscape, allowing the photographer to reconsider orientation and "shift" the cropped window in post, yet provide simple printouts to common photo print ratios.

[Andrew]

Mike,

You stated "I overestimated the utilitarian aspect of long-lens popularity and underestimated the status aspect." This is clearly true and one of the main reasons why the D700 is so big and the D3 and 1D family are enormous. You could just as easily have written "I overestimated the utilitarian aspect of small camera popularity and underestimated the status aspect [of having a brick to hand around the owner's neck]."

[beuler]

In the last decade Olympus got burned thrice trying to innovate in camera design. The E-10 and E-20 used a pelical mirror and got bashed by the review sites for robbing x amount of stops from the sensor. The E-1 got bashed for (wait for it...) not having enough mega pixels and too much noise at high ISO. The E-330 is the strangest case - that camera received poor reviews because it had Live View(!) and didn´t look like a 1990´s slr.

No sir, the camera market is not ready for form-factor innovation, it´s all about being able to photograph in the dark.

[Andrew]

Has anyone tried the 3CCD approach (or 3 CMOS) for still cameras? Might be expensive but it sure helped a whole load in early DV camcorders and for years in pro video.