« OT: New 'Pretties' | Main | Canons Take a Dunking and Keep on Clicking »

Tuesday, 14 July 2009


I've wondered how the Foveon sensor would fair if developed by one of the big manufacturers with the money, equipment and experience to move development along faster and further.

At this point, it's not possible to know outside the labs whether the relatively limited success of the technology is inherent or the result of lack of development and lack of competitive bodies.

BTW, Oly and Panny are deeply intertwined in 4/3 & m4/3 development and production. Most of the subsystems other than image processing and sensors are made by Oly for both brands and I believe Oly is using only Panny sensors for 4/3 these days.

So I don't know how well they are doing overall, but Oly is making some income for every 4/3 & m4/3 camera sold.


Slightly O.T.: I recently realized that Leica, instead of making all of us wait 'till the problems created by the M-mount's short flange to focus distance are sorted out by the sensor manufacturers(maybe never), could have made an M-like camera with a larger F to F distance, and a new set of lenses. After all, they've introduced at least 8 or 10 all-new lenses SINCE the M8 came out. Sure, it wouldn't be exactly an M, but the M8 isn't either!

I agree with the arguments, which makes it more baffling to me why Sigma tried to take on everything in one bite. It seems to me that the key to succes for a small company (or any limited budget environment) is to focus more on core aspects.
And then there is the issue of incremental improvements. Sigma are now on the nth Foveon chip camera and yet it still seems to have the same hang-ups and RAW processing support is as poor as ever. The handling, by report, seems unchanged from the DP1. Why introduce a whole range of accessory components?

This probably also applies to SLR manufacturers with lower volume (i.e. Pentax). I think I've read how they have been extremely selective over the functional requirements of their future cameras to focus their R&D efforts given their limited budget (same for their lens lineup).

On saying that, I would expect that there would always be a market for those requirements built into DP2. However, you've presented a very valid argument as that as a reviewer or a reviewing public without knowledge of these requirements or the budget limits -- we don't have to care. We just have to cough up our hard earned cash -- so in the end; we have the final say.

Yep. It's really too bad that Canikon's domination of the SLR business enforces such suffocating design conformity. The barriers to entry and cost of an unsuccessful venture are so extreme that no one dares try anything really different. I guess the micro 4/3rds system is the exception proving the rule.

Manufacturers could be a bit more daring back in the film era. I built a Pentax system, and was delighted with their quirky, compact cameras, even if their autofocus was never really competitive. The MZ-S in particular was a gem; it was much, much more compact than the Canikon equivalents, with a top plate slanted toward the user that made it far easier to read the LCD on a tripod. It had the most intuitive exposure controls imaginable, and accessories like the vertical battery grip were so sinuous it seemed like they were designed by H.R. Giger. Yet these clever design features never translated to market success. Canikon's market muscle created its own gravitational pull.

There's been a lot of noise about the DP2 review, and I was certainly disappointed by Mike's findings---not that I'd dispute them (the camera didn't work for him), but that I wish he had had the help of someone familiar with the camera to give him a hand (updating the firmware, demonstrating how to shoot with MF, etc.). My guess is that if he had had that benefit, he might have had an 'a ha' moment and gotten along better with the camera. Martin Doonan comments on Sigma trying to take on 'too much'---I would argue that they haven't. There's no face detection, IS, or many other P&S features in the DP2. They absolutely focused on the two core aspects of this camera---the lens is superb (it outresolves the sensor), and they've wrung out the best IQ out of this generation of the sensor so far (SD14 < DP1 < DP2). As Mike noted in his review, he can get images that are better than his DSLR.

I see a lot of 'mythology' regarding the camera. The handling is not unchanged from the DP1---the quick menu system is better and very useful (different than almost everyone else, but fast to figure out). The RAW processing speed is better---not fantastic 3 fps until your card is filled up, but shot to shot is really good (and better than the DP1---which is why I'll guess that there's a DP1mk2 in the future). A DMD camera defined by HCB would be one that allows you to set up and anticipate the shot and then execute. With the very nice manual focus wheel of the DP2, it's quite easy to do. In fact, you could set it up on a tripod, set the intervalometer, and go have some coffee and let the camera do the work...

Ted makes a great case for small companies taking risks. I've admired Sigma for doing this---they saw something unique with Foveon and struck a deal with them (eventually buying the company), and are in this for the long haul. As a family run private business, they can execute on a much longer timeframe than publicly traded 'what have you done for me last quarter' companies. The fact that they spent THEIR cash to acquire Foveon and continue to work on new chips is a great sign, as was their efforts to partner with Fujitsu to make a processing chip that would speed the X3 processing (one of those technical nits that makes their life more difficult than the rest of the camera makers).

When it comes down to it, the DP2 is a camera that beats all others this size when it comes to image quality. I'm a huge fan of the Foveon sensor, because I believe it makes a better print than most other CFA camera---and my friends at work who shoot Nikon and Canon always wonder why my photos on the walls (12 x 18"s) have that 'quality' that they can't seem to get. The DP2 is a very usable camera---not for sports. Not for indoor color work in crappy incandescent light at ISO 3200 (although this generation is much better at higher ISO color). It does marvelous B/W work (funny coming from a 3-layer color sensor). It's good to know that there's a company committed to making this type of product (and they're committed to making other aspects work as well---but they do hit technological hurdles)---just like it's good to know that there are Corvette's out there that have wacky levels of performance and power. These products aren't for everyone, but for those who want or need it, they're available.

This reminds me of the folding MF camera Voigtländer recently produced. People complained that it was terribly expensive for what it was, on the grounds that (say) Nikon could produce something which obviously was a lot more complicated and featurefull for less money.

While that is true it completely fails to account for Nikon having both much, much larger expected sales, and also being able to absorb the risk of a product not being profitable in a way a smaller company can't.

Not so much a comment on the specifics of the post (but I think Ted had hit the nail on the head), but to simply say this is the type of well thought, insightful writing that keeps me coming back to TOP. Can you imagine finding this on another site?

To make the 'vette analogy more comparable, Sigma should be able to take a Canon/Nikon and replace the CMOS sensor with their own Foveon - and I guess this is more in line with what they do with their lenses.

In fact, what they are doing with the DP2 (and other cameras) is akin to a small company producing a complete car in competition to the 'vette. The only way this can succeed is by making it much better, and hence, enormously more expensive.



Pak: I think you are right about Pentax's piecemeal strategy. The K10d gave them a sealed fullfeature body with grip option, the K20d concentrated on the sensor-related issues, the K7 concentrates on refining the mirrorbox and body features. There is some (online) complaint that the sensor improvement seen from K10d to K20d has not been seen again going to K7; that is to miss the point in my opinion. The sensor was the focus of effort with the K20d. Pentax cannot dramatically improve all aspects of their camera together, they don't have the resources it seems, or the price band to sell it into. But if some aspect of your camera, however isolated, is at a particular moment class-leading - that keeps you in the game.

This all makes sense, but still, what I don't understand, is why no big name produces something with in the spirit of Sigma dp2 or Olympus pen. Do they fear competition with their own DSLRs? I don't think that's realistic, at least for a dp2-like camera. If Canon G10 had a larger sensor and maybe a more "pro" lens, it would probably beat any competition.

In discussions of this and similar topics, everyone seems to forget the Epson R-D1. Perhaps the very fact of that omission is instructive concerning the marketing and "mind-share" challenges of a small-production, high-quality niche-appeal digital product.

The R-D1 is broadly similar to the M8 (manual focus M-mount rangefinder with AE) - albeit using sensor technology from 5 years ago - at a price 50% lower than the M8 (or less in the used market). But I find it interesting to note that Epson have seen fit to issue two updates to the R-D1 since its inception in 2004: the R-D1s in 2007 and the R-D1x in 2009. Although both offered only incremental changes, with the original 6MP sensor remaining in place, they seem to have found a market sufficient (apparently) to justify the effort, and I've rarely heard regrets from someone who bought one.

Maybe Ted's point about recouping the R&D costs associated with high-performance digital component production explains why the R-D1 sensor hasn't been upgraded. Or maybe Epson is moving slowly to avoid a big front-loaded expense: The R-D1s updated the firmware with some fixes/features and enlarged the RAW buffer. The R-D1x upgraded the image processing engine and enlarged the LCD. Maybe the next version with bring in a sensor with a more competitive pixel count?

As a user of many Foveon products, including each of the Sigma versions, I would take issue on incremental improvements. In my experience, Sigma has made many improvements in the cameras and software. One can see from comments at the dpreview Sigma forum that users are happy with these improvements, particularly the changes to the raw processing software - I know that I am. Incremental improvements is in fact an area where Sigma shines, as shown in their many firmware updates which have not simply addressed bugs but have also enhanced operation, quality, etc. The big knock, with which I agree, is that they have been slow to address the bigger issues, perhaps for reasons described in this excellent article. Mike

I like the article. Well thought out and easily understood. Low volume "edge" market competitors are always at some disadvantage. Their engineering design and parts procurement teams try to do as much as they can on so few resources (talent, money).

One thing surprises me, and perhaps its my lack of understanding, but Sigma makes some very fine optics. Does the sales of lenses not adequately fund further product development?

I remember when Vivitar was a force in the industry and offered some interesting, unique, and quite decent photographic tools. Their California based engineering design team could specify product that could be built by any number of manufacturing facilities.

I'm shocked at Leica and their M8. Whatever happened to "superior" German engineering?

"what I don't understand, is why no big name produces something with in the spirit of Sigma dp2 or Olympus pen."

Well, they don't yet. There have been some rumbles that Canon is doing some serious market research in the category.

But why don't they provide body-integral IS either? Probably because they've tied their flags to lens-based IS and make more money that way, and don't feel sufficiently threatened by the alternative. Of course they could be developing that too, I don't know.


You call this small company taking a risk?

I call this small company NOT taking a risk. What they essentially did was "Let's build a camera, but let's not spend too much on development. Just in case it won't fly."

And guess what? It does not fly exactly for that reason.

The ideal small camera -- teeny, compact, fast interchangeable lenses; at least APS-C sensor size; shoots raw only at not less than 5 fps; converters to support at least one line of DSLR lenses with all the contacts supported; through the lens optical viewfinder ('cause that's the only way you're going to *get* an optical viewfinder for arbitrary lenses) is both a stonking engineering challenge and going to be sold into the most nitpicky, cranky, talks-on-blogs, wants-the-moon-for-sixpence market segment going.

So success is difficult; if you don't achieve it, you take a loss; if you do achieve it, you've cannibalized your own DSLR business, and probably *also* take a loss.

It's quite possible that the three? four? five? companies that could do this in engineering terms have looked at the business case and gone "we don't need the pain".

The Corvette comparison is interesting since I was just talking to my wife this morning about the car business models used by Lotus and Fisker (as I read in this month's Car and Driver). They are trying to cut down development times by staying small and outsourcing most parts, engines and, in Fisker's case, assembly. I think it's a great model compared to the dinosaur car manufacturers. The dinosaurs are so big that new car development takes years and millions of dollars. Lotus developed the Evora in 27 months (according to evo.co.uk) and it already is sounding like a great car albeit with some issues like unreadable gauges in certain light conditions. (Sounds like a camera viewfinder problem too).

Plus, the big dinosaurs won't be able to make quick changes to product lines when market demands change. Small companies are hurt is in distribution, supply chains and costs. Lower volume means higher prices, they have to find dealerships willing to sell and service their cars and parts makers may not be willing to help with a low volume sales. There are benefits too as new manufacturers may not be wedded to steel producers and can quickly introduce new technologies like carbon fiber. But just like Leica, the only products we will probably get from these makers will be expensive niche products like the Lotus Evora or the Artega GT.

It will be interesting to see if some of these companies can slowly build up sales and models to encroach upon the big boys. I'm still hoping a 2000 pound hatchback with a modern 4 cylinder engine and modern safety standards is possible one day. Even my Honda Accord feels heavy which just seems not right for a Honda.

pb here again.
a semi-retraction of previous post. I, obviously, at first, did not get the point.
But... those companies mentioned aren't newbies, some are older than you and I(65) and are the innovators, (not Callaway).

WE do need to encourage them at all times and at all cost, or this whole thing (photography) will come to be Lemurish.

thanks for listening. pb

There is a good analogy with the automobile business -- Mazda.
The Miata -- conventional technology, nothing ground breaking, just a different execution, and carved a niche market other makes had to hasten to imitate. This is Micro four-third -- conventional technology, just different execution. But it will not be main-stream, just a niche, considering all the limitations again the benefits. The Conika will follow with Miata look-alikes? Yes, if they realize there IS such a niche market.
Then there is the RX series sports car with the rotary engine inside. Mazada almost got pull down but their investment in the rotary engine technology. If not for their conventional line of car models, Mazada is no more. THe 1st few iterations of the RXs were problematic. It is almost there now with the RX8, but still, it is another niche. The Feveon will be perfected after a few iterations, and if not for the successful lens business, Sigma and Foveon would have been history already. Let's sit on the fence and watch history unfold (or repeat?).

Does anyone actually know how many DP1 or DP2 have been sold? and how does it compare to their targets?

Foveon and the rotary engine are a good comparison. Both interesting (though over-hyped) technologies.

But the difference bewteen Mazda and Sigma is that Mazda were putting their revolutionary product into a technology with which they already had much experience. Sigma were putting a revolutionary product into an entirely new (for Sigma) product range.

What might have been interesting is if Sigma had of partnered up with, say, Pentax to deliver a Foveon based camera range.



How great to see almost back-to-back two such concise articles as "The Price of Being Small and Taking Risks" and "Reality Is Not Arithmetic"!

Both are simple and direct, nicely reasoned, clearly laid out and accurate.


You have to start somewhere. Sigma has never built a small digital camera before the DP-1, and that was the first iteration.

The DP-2 improves on the DP-1 in many ways - as another posted noted handling is much better (quicker to access a much wider range of abilities in one or two clicks, like auto-bracketing) with much better shot to shot times. I personally have found that when using the quick focus mode on the DP-2 (people/mountain) that focus in good light is 2x-3x faster than the DP-1. SPP now does a better job processing RAW images than it used to, even for the older cameras, and they'd managed to keep Lightroom/Photoshop integration up to date to provide a strong mainstream alternative to processing RAW images.

To me it makes no sense for Sigma to build the most expensive camera possible, until they get the fundamentals strong - the basic camera usability. They are pretty much there with the DP-2, now they can play with adding more expensive components in future cameras to improve operations. But we should not forget that $650 is already "pricing to value" above a really cheap digital camera, but at a level that a lot more people are willing to buy into.

Sigma has shown with the SD-X series of DSLR's that they are willing to grow organically and improve systems over time incrementally. I see no reason they will [not] continue to do so, after all after each camera Sigma has put out there have been people claiming it was the "last camera Sigma would make" and each and every time they will have been wrong. Meanwhile the other camera makers seem content to let Sigma expand the niche they are addressing with the DP-1 and DP-2 - not even the EP-1 fills the same need (since that is really a small replacement for a DSLR system and will live or die on those merits).

This is a really interesting post. I kind of agree with what you are saying, on the whole.

The business reasons for launching a product or not can vary a lot and not always a real advance. Shooting with a DP2 and a Canon G10, I did not see impressive differences. For a moment, comes to my mind the history of APS (film), remember? It was “the future” for 35 mm film format. Just a couple of years of pure excitement…. and burned money. Before invest in Foveon what if to wait for a time test? Sigma is probably making the bride beauty to sell the patent for Nikon/Canon domain.

The comments to this entry are closed.