Reading Tea Leaves
PMA 2008 and The Future
I didn't get to the PMA show in Las Vegas this year. A winter snow storm cancelled my flights, and rather than struggle my way there I decided to spend the weekend at home, watching the Superbowl, thinking about the industry, and eventually writing this article.
Since there are any number of sites that report on the new products being shown at PMA for the first time (most were announced during the week leading up to the show, rather than at the show itself) you'll have no trouble finding pictures and specs of all the I-gotta-have-it items as well as the yawners and me-toos.
But I thought that it might be fun to ponder the significance of both what was announced at PMA, what wasn't, recent industry trends, and where they might be leading us. Read the tea leaves, in other words. So what follows is what is known in the biz as a think piece. Few facts, mostly opinion. Agree with me? Disagree? Fine. Care to make your opinion public? Join this sites discussion forum and let everyone know what you think.
The Megapixel Race
Things are still silly in the digicam field with shirt pocket cameras now up to about 12MP. This means 2.8 micron pixels (or maybe even less) which if this trend continues will begin to impinge on the size of the upper wave lengths of light. Stuffing photons into these little holes is going to start challenging the laws of physics pretty soon.
Update: Ironically, to both put the lie to and validate my comments here, Kodak has just announced a 5MP sensor for use in cell phones that has 1.4 micron pixels and speeds up to ISO 3200.
In the DSLR world sanity seems to be settling in, with pixel counts in the 12 – 14 MP range becoming the norm. The high end of the pro DSLR market seems to be at the 21 – 24 MP range, and while that leaves room for the lower end of the market to still move upward, the ceiling isn't going to get much higher once pixel count gets above 25MP and photosite sizes below 5 microns, because noise will become too big an issue at anything other than moderate level ISOs. Photographers now want image quality above pixel count, or at least I do.
It also needs to be asked – which photographers need these large files? Typical commercial uses (magazines, newspapers, etc) are easily satisfied with files in the low to mid teens, and while a larger file, such as from a Canon 1Ds MKIII, is great for making a 24X36" print, how many people actually need this? Of course a larger files means a greater ability to crop and still get a usable image size, but this then starts to stress lens performance; a bit of a vicious cycle.
So, as far as I'm concerned, anything north of a high quality 12 Megapixels is fine for most applications, and 20+ MP files (whether from a DSLR or a medium format back) are only needed in the work that I do for my most critical landscape work and some commercial projects. (For example, I have a commission to document a major urban renewal project, and in addition to an eventual coffee table book have been told that wall-sized blow-ups for a presentation center will be needed. So, I'll be shooting much of that with a 39MP medium format back.)
Horses for courses, as the British say. In other words, look realistically at the work that you do and choose a camera that is appropriate for the task. If you're a travel photographer you'll have different needs than if you shoot wildlife, or fashion, or products, or life styles. One size does not fit all.
For some six years Canon has been essentially alone in offering full frame 35mm DSLRs. Contax and Pentax took a shot at it some years ago, with Pentax dropping out before hitting the market, and Contax withdrawing after only producing a very small number of cameras. Kodak (remember Kodak?) also did so with their 14n, 14c, SLR-N and SLR-C, which were noble but highly flawed attempts.
The Canon "1Ds" series, at 11MP, then 16MP and now 21MP was otherwise the only mainstream game in town, and gained a lot of mileage in the pro community because of it. Conversely, Nikon lost a lot of ground in the pro market because of its lack of a full frame camera and thus very high pixel counts. Many stock agencies started demanding higher res files and Nikon shoots simply couldn't provide them. The LBCAST sensor fiasco didn't do much to bolster Nikon's reputation when it came to sensor quality either. The faithful stuck with Nikon but there were a lot of defections over the years.
But beginning during the second half of 2007, and in strength at the start of 2008, Nikon has turned the tables. In August '07 Nikon announced the full frame D3, and since it started shipping in November it has received nothing but praise.
In January Sony announced its 25 Megapixel full-frame sensor and shown prototypes of its forthcoming flagship camera where it will find a home – due later this year. Of course no one doubts that a version of this sensor will find its way into a Nikon D3 variant, likely sooner rather than later.
Competition and Complacency
When Nikon announced the D3 and D300 last August, along with a slew of high-end pro lenses, it was clear than the company was finally serious about getting back in the game. These new cameras have proven this, and the new stabilized super-telephotos, 24-70mm, 14-24mm, Nanocoating lens technology, and now three brand-new tilt/shift lenses, are, if nothing else, a clear signal to Canon and the pro community that Nikon once again intends to compete in every segment of the pro and prosumer market.
For its part Canon is to be applauded for everything they've done to move sensor size and image quality forward during the past 6 – 7 years. Its stabilized super-telephotos and tilt/shift lenses of the late '90s were breakthroughs, and were influential in moving a lot of photographers into the Canon camp (me included). Certainly the Canon D30 camera of 2000 broke the hold of prohibitively high prices for DSLRs, and established CMOS technology as viable for high-end camera sensors for the first time.
But this is now 2008, and there is a definite feeling out there that Canon has not been as innovative recently as they have been in the past. Many of its smaller competitors have introduced advances such as in-body stabilization, providing shake reduction to all lenses. Sensor shake for dust reduction was introduced by others first, as was Live View. Articulated LCDs are finding their way into more and more DSLRs, making Live View more usable for some, but this has not yet appeared on any Canon DSLR.
But, lets not quibble over features. The point is that Canon is the 800 pound gorilla of the industry, and though they've done a lot of innovation over the years, they have not, in my view, taken sufficient advantage of their lead recently. This has left openings for other companies to exploit, and they are now doing so.
An example of this is wide-angle lenses. It is no secret to anyone that Canon's lens line-up, though very extensive, is not as strong as it needs to be in the wide angle range, especially WA zooms. Full frame high resolution sensors such as found in the 1Ds MKIII push lenses very hard. Wide angles are harder to design than medium focal length and long lenses, especially when used on DSLRs.
For this reason many demanding Canon users buy wide angle lenses from Leica and Zeiss and use them with adaptors. This means no autofocus or even autodiaphram, but for many critical users the superior image quality of high-end lenses from these two companies makes the compromise worthwhile.
Canon appears to be aware of this weak spot and has recently been upgrading some of its WA lenses. But apparently not well enough, as shown in this comparison of the Canon 16-35mm f2.8 L, Nikon 14-24mm f2.8 G and Contax Vario Sonnar 17-35mm f2.8. Incidentally, I've been shooting with the just-released Nikon 14-24mm and am astonished at its image quality. This may well be the finest ultra-wide angle zoom ever made; one that rivals most primes in its focal range.
Nikon is a traditional rival to Canon for the high-end market as well as the consumer segment. But Sony is the new kid on the block. With its acquisition of the Konica / Minolta camera line and technology a couple of years ago Sony showed its intention to become competitive in the prosumer if not pro market itself. It is also the only other major camera company with its own sensor fabrication capability other than Canon, and their new 25MP chip announced at PMA shows that they fully intended to exploit that capability. (Pentax through its hook-up with Samsung now is also somewhat in this category, but that combo has so-far failed to live up to its potential).
Olympus continues to march to the beat of its own drummer. I was dubious about the 4/3 format when it came out, and have been every since. Nothing that I've yet seen has convinced me otherwise. Their Pro bodies using this format are no smaller than top-of-the line reduced frame DSLRs, and while Olympus lenses are generally just about as good as it gets, they are expensive and not all that much smaller either. I don't see the just-released E-3 as a convincing follow-up to the E1, even after a 4 year hiatus.
Olympus' consumer grade DSLRs have a lot of appeal, but their smaller sensor will continue to be its Achilles Heel and will prevent them from being able to be competitive in the high end, as full frame sensors continue to drop in price and increase in resolution. It's like the difference between medium format and 35mm. Size matters, and in digital as with film size has clear advantages.
The Singularity is Near
I've mentioned it on these pages before, but I'll do so again. If you want to get a feel for what exponential technology change in the semiconductor field is going to do to our society (and ourselves) read The Singularity is Near by Ray Kurzweil.
My point is that the camera business is now very much a semiconductor dominated business, and thus is subject to Moore's law, which depending on how you interpret it tell us that chip density doubles and prices are cut in half every 18 months. For the most part this has been true for some 30 years, and knowledable observers agree, continues to look like it will remain so for the foreseeable future.
That's why your $1,000 desk-top computer now has more computing power than a multi-million dollar room-filling Cray supercomputer used to, and why a DVD player that used to cost $900 now literally sells at Best Buy for $25. (Did you ever ponder how much that DVD player costs to manufacture so that it can sell for $25? There's parts, packaging, documentation, allowance for warranty, electrical and safety approvals, shipping (often half-way round the world) and of course the profit margin for the manufacturer, importer, wholesaler ,and retailer). That's mass manufacturing for a global marketplace and Moore's law in action.
Not to beg the point, but as I am writing this I got an email promotion from a major US photographic retailer offering a Kensington 4GB CF card for free. Yes, free. Pay for it and receive a rebate from the manufacturer for the full retail price. (Let me know when they start actually paying us to buy things).
Because the camera business is now very much the semiconductor business we will continue to see exponential growth. This means higher resolution sensors, larger sensors, and lower cost sensors. I have little doubt that within 1 – 2 years we will have cameras with full frame sensors of about 16MP for under $1,000. (Drop me a line on Dec 31, 2009 and let me know if I was right or not. My guess is that this is, if anything, a conservative forecast.)
Cameras though are, of course, still very much mechanical devices. There are expensive mirror assemblies, and prisms and chassis, just as there always have been. But the real cost of digital cameras is in their semiconductors, and these are subject to Moore's law. There's no reason to think that it will become otherwise. And do keep in mind the $25 DVD player that used to cost almost $1,000 just a few years ago.
Medium format is, in my view, at a critical juncture. With 21 – 24 Megapixel DSLRs either now available, or soon to be, the 16 – 21 MP medium format back's raison-d'etre is no longer. We can quibble as to whether a 21MP back has higher image quality than a 21MP DSLR, but given the advantages of cost, size, weight, convenience, long lenses, wide lenses, high shutter speeds, and high ISO quality, the result of this battle is a foregone conclusion.
That means that it's 30MP and especially 39MP backs that are the new sweet spot for MF, and I wouldn't be surprised to see anything smaller than 30MP essentially disappear from back maker's lines over the next 6 months.
There has been industry buzz for at least two years now that higher resolution chips are on their way, and so I wouldn't be surprised to see backs of around 60 MP announced, if not shipping, by Photokina in September. This is really medium format's only salvation, keeping its size edge over full frame DSLRs.
This begs the question of who needs 60 MP? The answer is, the same people that today need 30 – 39 MP. Photographers, both pros and others, whose commercial or personal work demands the absolutely highest image quality, and cost be damned. But, and it's a very big but, image quality can not be compromised just for the sake of chasing more megapixels. This is the slippery slope that medium format back makers will find themselves on – needing to create backs with more megapixels to stay ahead of the top DSLRs, but doing so without compromising the highest possible IQ.
As for the players, there are currently just three companies making medium format SLRs (Hasselblad, Franke & Heidecke, and Mamiya) and five back makers, Hasselblad, Mamiya, Phase One, Sinar and Leaf. (Leaf and Sinar's Hy6 camera are OEM'ed from F&H).
My sense is that the market is too small for this many participants, and that a shakeout is inevitable. Some of these companies have large corporate parents, some have outside investors, and others are management-owned. Regardless of their corporate structure, all companies need to make profits, and shareholders eventually lose patience when they don't. My guess is that 2008 will be a year of further shakeout in medium format, and only the strongest will survive.
Who will those survivors be? I have my guesses, but will keep them to myself for the time being since they are simply conjecture rather than being based on any hard evidence. But it's safe to assume that the medium format industry will likely look quite different in February 2009 than it does now.
Sigma continues to provide a fascinating sideshow to the actions of the major players. At PMA it announced that its long-delayed DP1 digicam, using a 4.6 MP Foveon X3 sensor of 20.7mm x 13.8mm size ( 7.8μm.) It should begin shipping soon.
Foveon sensors are controversial. They are the only sensor on the general market that does not use a Bayer matrix. Rather, each photosite is three layers deep, with separate Red, Green, and Blue sensitivity layers, just like film. Consequently Sigma and sensor designer Foveon call a sensor like this 14 Megapixel, even though the image created is only 2652 × 1768 pixels, which my calculator tells me equals 4.6 million.
Experience tells us though that a Bayer system loses about 1/3 effective luminance resolution because of the Bayer array, compared to a none-matrix sensor array such as Foveon's, so things aren't as wide apart as they first appear. Add the fact that a Foveon sensor doesn't need an anti-aliasing filter and again it gains a resolution advantage over what the numbers alone might indicate. But there's no getting away from the fact that Foveon chips simply don't have the resolution-producing pixel density that more traditional sensors do.
On the other hand, they do produce a very clean image, though not as magical a one as some devotes would wish, and which they fervently that claim it has. The soft core of the Foveon chip is its poor blue channel. This can lead to noise, especially when shooting under blue-deficient light sources such as tungsten illumination. But, the photosites on the DP1 are huge by comparison to anything else seen thus far in a digicam, and so I'm very keen to find out what files from this camera will look like.
Sigma also is flexing its muscle when it comes to lenses. They are one of the big-three independent lens makers, and though some of their consumer grade lenses are merely ho-hum, their better (read more expensive) glass can be quite good indeed.
At PMA Sigma showed several new high-end lenses, including a preposterous 200-500 f/2.8 weighing in at some 16kg, and looking like nothing less than a bazooka. On a more down-to-earth level their new 120-400mm and 150-500mm stabilized long zooms both look like they will offer photographers an interesting alternative to the inevitably more expensive offerings from the major camera makers.
Printers, Papers and Inks
It's almost a cliche to say it, but there have been more advances in colour printing in the past ten years than in the previous 100. Today's pigment-ink based inkjet printers are capable of producing sharper, wider-gamut, and longer-lasting colour prints than any prior process or materials. And within the past couple of years advances in inks and papers as well as printer's dithering algorithms have allowed these printers to produce neutral and lovely B&W prints that rival silver gelatin in almost every aspect.
New papers, such as those utilizing Baryta, which use Photo black ink and which are fiber based, are beginning to once again revolutionize fine art printing. No longer do we need to choose between the esthetic advantages of fiber-based papers and the use of photo black ink with its denser blacks.
I may be going out on a limb here, but I believe that over the next year or two, as even more appealing fiber-based papers which use Photo Black ink come to market, the fine art community will transition away from the use of matte black inks. Somewhat ironic, as this likely will start to happen coincident with Epson finally producing large format printers than take simultaneous blacks without swapping, like their Stylus Pro 11880 recently reviewed here. Still welcome, of course, but maybe a bit behind the marketplace's previous needs.
The Bottom of The Tea Cup
So what do the tea leaves tell us? There's no question in my mind that this is the golden age of photographic tools. As photographers rather than hobbyist we currently have equipment that enables us to produce images of higher quality than ever before.
Whereas in the past some scientists and engineers in Rochester, New York or Tokyo decided on the palette of our colour images and the tonal value distribution of our B&W photographs, today with silicon sensors instead of film, and the powerful raw software tools available, we are now the ones that determine the fundamental characteristics and look of our images. This is revolutionary, and still not well understood or appreciated by many who still are stuck with the notion of absolutes. Colour exists in our heads, it is not an absolute, and it therefore demands interpretation by our minds and hearts. This is what lies at the core of being a photographer today.
But whether these new tools and their associated freedoms will enable us to produce better photographs is a question to ponder and for a future discussion.