Add a comment or send Thom feedback on this article.
5/1/07: Original
7/2/07: Updates from comments

  Thom's Compact Camera Challenge

If the camera companies can't design it, let me step in and show them how.

   


Despite almost two decades of compact digital camera production, it seems that the camera manufacturers still haven’t realized that there is a substantial-sized market not being served by current compact digital cameras. Meanwhile, the trend towards more and more me-too designs continues, with little to distinguish compact digital Brand A Model A from compact digital Brand B Model B. And all the while, the camera manufacturers consistently complain about lower selling prices, lower gross margins, and thus lower profits. Some even complain that the market is too tough to distinguish their product in. Give me a break. There are so many untapped camera markets out there that if I had my own camera company I’d be cleaning up.

(It seems I hit a nerve here. Everyone who's written seems to agree with the above premise, including some folk who work at camera companies. The key words are "not being served." Seems that an awful lot of folk agree with that sentiment. By "awful lot" I mean that I can point to nearly a thousand such responses in my email alone.)

My analysis of the situation is simple: the camera makers simply don’t have any idea of what many of their customers really want. Instead, they make what they know how to make and what they think has the largest possible volume, and they keep trying to do so at lower costs in order to survive the margin hits they keep encountering. When all is said and done, today’s Coolpix P5000, for example, isn’t a lot different than the previous Coolpix P4, so why would I want one? And the problem doesn’t just live in Nikon’s consumer digital camera production; you see the same problem at Canon, Sony, Fujifilm, Pentax, and a host of other producers. Heck, Canon has averaged eight new compact camera models a year for a decade, yet the latest ones aren't really any different than the first ones (other than a few more noisier megapixels).

True, sometimes we do get something a little different, such as the “all weather and underwater” versions of the very same compact digital camera designs that come in non-protected cases. Or perhaps we get a new feature, like sensor-based vibration reduction or the much-marketed-but-not-so-useful face detection autofocus. But for the most part what we get is the same design iterated over and over, just with cheaper sensors with more megapixels and more integrated ICs to cut costs. Silly thing is, what's being implemented over and over is not even a good design. If I see another compact camera where the down-key on a Direction pad brings up a Focus menu so that you can get to the Close-up focus mode, I think I’ll puke. Buttons are overburdened in bad and inconsistent ways, “modes” exist all over the place, and the ability to control things gets buried deep into a multitude of poorly designed menu systems, mode dials, and multi-purpose buttons. Human interface design appears to be a subject that the Japanese engineers skipped in their education.

Personally, I’m sick and tired of all this mediocre me-too camera design. Not a single compact digital camera meets my requirements for such a product, and I think that most serious and professional photographers would agree with me on that. Rather than just sit and rant about the onslaught of me-too compact digital cameras, though, I decided to do something about it.

To help the designers along, at the end of this text you’ll find the basic specifications for a consumer and pro version of a camera I’d judge as adequate. Simply put: larger sensor, high-quality lens, and user control. Virtually every specification I list basically falls into those three categories, which tells you something about just how miserably the current crop of more than 60 million cameras being sold a year fails. The one thing that isn’t in those categories is a dedicated autofocus system (rather than double-purposing the imaging sensor as almost all current designs do), and this requirement basically points to the other failing of all current compact digital cameras: they aren’t responsive enough.

If the camera makers can’t quite figure out what such a camera looks like, I’ve even done some mock-up images of the camera. And if that weren’t enough, if any manufacturer is interested, they can have their product management team directly contact me and I’ll provide them with the user interface document for the full set of controls on my “adequate compact.” (Serious inquiries only, please. I’ll also publish a list of any company that requests this UI document, so the public will know if you looked but didn’t listen...)

(It seems that some take my mock-ups too literally. Let me point out that I used the Nikon Coolpix P5000 as a base simply to chide Nikon; this is a Nikon-related site, after all. The P5000 succeeds at everything except sensor, lens, AF speed, and direct UI, which is to say it is essentially a consumer point-and-shoot with manual controls that aren't always directly accessible. The lens in my mock-up is in a retracted position, and after doing some follow-up math, probably is a little too small in the mock-up. But my mock-up does exactly reflect my UI document--in other words, the controls are all there.)

In my original version of this article, I was going to offer US$10,000 to the first company that could both match my specs and produce a workable compact camera. After circulating a preview to a few industry folk, they all said the same thing: no amount of monetary incentive would change design decisions, so it would just look like I was grandstanding with a fake check (really, I would have gladly written the check). So, here's the new tactic: just write your favorite camera company and point them to this article along with the comment "I'd buy that for US$xxx" (you fill in the xxx with what you'd pay). So Nikon, I'd buy that for US$899. Today. When was the last time you sold a Coolpix for US$899? Get the hint?

Thom’s Compact Camera Challenge Specs
thom hogan's compact design, Copyright 2007 Thom Hogan

Camera type: non-interchangeable zoom lens compact camera (amateur and pro versions)

Sensor

  • Sensor type: APS CCD
  • Sensor size: 23.6 x 15.8mm
  • Total pixels: 10.75 million Bayer pattern
  • Effective pixels: 10.2 million
  • Image sizes: 3872x2592 (Large), 2896x1944 (Medium), 1936x1296 (Small)
  • Image settings: Tone compensation, Sharpening, Color Space, Saturation, Hue, White Balance, B&W
  • VR: sensor-based vibration reduction

It was pointed out to me (correctly) that my thinking is a bit Nikon-centric, thus the APS sensor. 4/3 would be perfectly acceptable, I think. The upcoming Sigma (right column) is 1.7x and looks like it is reasonably close to my specs, if not my design goals. Thus, I'll say that we need a DSLR-type of sensor (4/3, 1.7x, APS, etc.). The design goal behind that is to achieve real image quality, indeed, DSLR-like image quality in a smaller, more carryable camera.

Lens

  • Focal length: 18-45mm (~28-70mm equivalent)
  • Maximum aperture: f/3.5-5.6
  • Aperture range: f/3.5 to f/22
  • Focus: 1.5’ (0.5m) to infinity
  • Other: removable lens cap, filter rings
  • Optional: screw in wide-angle (0.75) and telephoto options (2x)

Lens (for Pro version)

  • Focal length(s): 18, 24, 35mm (fixed settings; tri-focal rather than zoom); Alternatively: 18 or 20mm fixed if tri-focal too ambitious for the optical team (it shouldn't be)
  • Maximum aperture: f/2.8
  • Aperture range: f/2.8 to f/22
  • Focus: 1.5’ (0.5m) to infinity
  • Other: removable lens cap, filter rings
  • Optional: screw in full frame fisheye accessory, optical viewfinder (aligned with lens)

There's been considerable discussion about my choice of sensor/lens combination. Several folk have written saying that the illustrated lens wouldn't work with the sensor size, as there would be edge telecentricity issues. First, the illustrations are for collapsed view and were more to suggest control placement than size--the lens would obviously have to extend during shooting. Second, I assume that the over-sensor items would have to be adjusted (microlens displacement towards the edges, for example). Finally, as I noted elsewhere, the camera very well may need to be jacket-pocketable instead of shirt-pocketable. But I'll also point out my style of design is different than the Japanese tend to use: first, imagine what it should be like; next, drop pre-conceived notions of "it can't be done." In my experience, if the goal is clear, the design is possible. Indeed, because the design may need some out-of-box thinking, this is where clear, unassailable patents come from (as opposed to those incremental we-thouight-of-a-slight-tweak-and-are-going-to-patent-it-anyway designs).

Another comment about the lens specs: the important aspect of this camera is image quality. There are three critical elements to that: sensor, imaging ASIC, and lens. Most of the Japanese companies have great imaging ASICs now. The sensor is obviously sub-optimal on all current compacts. But so is the lens on many. One only need look at the quality of the Leica lens on the Panasonic high-end compact to see what I mean. That means no plastic elements and use exotic materials (flourite elements) to get the quality up. Even in the low-end DSLRs we have companies that don't quite get this. Nikon's 18-55mm kit lens so outperforms any other one that I've seen and used, it's clear that the other companies are cost cutting at the expense of quality. The overall design goal of the product in question here is simple: top-end DSLR image quality in a compact camera design. Sometimes design goals come in conflict with one another. In this case, the lens quality conflicts with size of camera. "Compact" is a vague range, not a specific cubic area you must hit. I'd not give up the weight specification, but I might give up a bit of the depth specification, in order to resolve goal conflicts. This gets me back to a common theme in my criticism of current camera designs: serious photographers know intuitively which way a goal conflict should be resolved, but these folk aren't being integrated into the design process to help get it right the first time.

I should also point out that not too long after I first published my article, Nikon begin surveying users about a serious compact camera. The options in their survey were a 24mm f/2.8, a 28mm f/2.8, a 35mm f/2.8, a 28-85mm f/2.8-4 and a 24-50mm f/2.8-4 (these are 35mm equivalents). The last two are awfully close to my suggestions. Still think such lenses aren't possible?

Viewfinder

  • Type: Live view on minimum 2.5”, 230k pixel LCD (highly consider swivel)
  • Optional: optical viewfinder fits in hot shoe (aligned with lens)

Another point of contention by many. I'd say about half of the supportive emails also said "but the optical viewfinder needs to be standard." Personally, I'd rather have a swiveling LCD before an optical viewfinder, as it provides a more tangible benefit (and potentially solves the "arms out" shooting position that some people want to avoid, especially for candid street shooting). Doing an optical viewfinder that's useful isn't trivial. You have difficult parallax and lens angle of view adjustments you need to get right for it to be more than a crude approximation. The usual Japanese optical viewfinder design these days is "80-85% view so that you're pretty sure that what you see was captured, even if other things end up in the photo." Getting beyond that requires a great deal of engineering and complexity. I'd rather compromise on this aspect than have the product delayed because of a critical path in the viewfinder mechanism. Before I started to use the Leica D-Lux 3, I might have said otherwise, but I've found its LCD the first that I can use in almost any environment to frame with, and now regard the LCD as something I can live with.

Storage

  • Media: Secure Digital, supporting SDHC
  • Logical: Compliant with DCF 2.x and DPOF (PictBridge optional)
  • Image formats: Compressed RAW (minimum 12-bit), JPEG
  • Buffer: Minimum 3 shots RAW, 10 shots JPEG
  • Additional features: Image comment in EXIF

Focus

  • Type: Independent multi-segment phase detection AF
  • Range: -1 to 19EV
  • AF types: Single Area, Dynamic Area, Closest Subject Priority or Group setting
  • Other: Focus lock
  • Manual focus: locks on AF point when switched, value then changed via Direction pad

Yes, there are probably other ways to do fast focus other than phase detection. I'm willing to compromise on the manner in which focus is done as long as it isn't done via running the standard imaging sensor long enough to get a focus derivation (the technique used in most compacts today). There needs to be virtually no focus lag even when the shutter hasn't been previously pressed halfway. That's the design goal. (I should point out that I built specs from a set of goals, which I didn't publish; camera companies will have to pay to play if they want a clear set of design goals from someone who knows how to formulate them ;~). One camera company designer who's ideas were forwarded to me in response to this article suggested one way of achieving my goal that isn't phase detection. That's what I mean by my above comment "if the goal is clear, the design is possible." He obviously guessed my design goal here and did exactly what I suggested: come up with another way of doing something that currently isn't done.

Exposure

  • Exposure modes: Program, Program Flash, Aperture, Shutter, and Manual
  • Meter: Sensor-based matrix, center-weighted, and spot (central)
  • Metering range: 0-20EV
  • Exposure compensation: -3 to +3EV in 1/3 stop increments
  • Flash exposure compensation: +1 to -3EV in 1/3 stop increments
  • Bracketing: 3 or 5 shots in 0.3, 0.7, or 1 stop increments
  • White Balance: Auto, Incandescent, Fluorescent, Daylight, Flash, Cloudy, Shade, Measure, and set any K directly from 2200 to 9000.
  • ISO: 100 to 1600, Auto ISO with maximum setting and trigger control
  • Other: exposure lock

Shutter

  • Type: Electronic
  • Speeds: BULB, 30 seconds to 1/2000 or higher in third stop settings
  • Frame rate: 1.5 frames per second burst speed

Flash

  • GN: minimum of 27’ (9m) at ISO 100
  • Control: Standard TTL, Galen’s Fill, Manual
  • Other: hot shoe for external flash

Battery

  • Type: Rechargeable, removable lithium ion
  • Capacity: Minimum 500 JPEG shots under normal shooting conditions

Physical

  • Size: “jacket pocketable”
  • Weight: no more than 14 ounces without battery and card.
  • Other: durable material (e.g. magnesium alloy construction), minimal weather protection (e.g. some gasketing)

Other

  • Standard tripod socket
  • Arca-style bottom plate built-in on Pro model
  • Nodal point and focal plane markers

Bonus points

  • Optical focal length coupled viewfinder built-in instead of optional viewfinder in hot shoe
  • Monochrome Pro version made available (no Bayer sensor)
  • IR Pro version made available (no UV/IR blockage)
  • Would be nice to have the tripod mount the nodal rotation point.

Camera doesn’t really need, but optional at manufacturer’s discretion

  • USB connection
  • Noise reduction controls
  • Image rotation
  • In camera cropping
  • Face priority AF
  • Red eye fix

Camera absolutely doesn’t need

  • Scene modes
  • Audio recording
  • Movie recording
  • Video out

A lot of folk said they wanted everything I specified but that video or audio should be included. There are reasons why I didn't include them, actually. The primary one is that I want to specify nothing in the design that takes any extra engineering cycles in addition to the primary goal of the product: high quality stills. Such features can be added to a second generation of the product, if really desired by a majority of users. The trick with design goals for a new product class is to get the 80% "necessary" part dead on right and ignore the 20% "nice" part in initial design. Since you're working on a short, fixed time frame, you want to spend 100% of your design time on the thing that makes the product different, not the "nice to have" features.

Can this camera be produced today? Sure it can. It could have been produced a couple of years ago (the Epson RD1, now discontinued, actually comes closest to the concept, though it was designed in the slot above where I'm designing [both in quality and price]). Most of the specifications listed above are easily do-able (and many exist on some compacts today, though often through less-than-satisfactory steps for setting them). The big switch is using an APS sensor instead of the much smaller sensors currently being used. This, in turn, impacts the lens design, which would be the biggest challenge to design (mostly to keep it small--there's nothing particularly challenging about the focal length or aperture specs). Indeed, the lens and AF would be the big ticket engineering items for any company attempting this product.

But you'll note that a lot of the design specs are simple things: filter rings, focus lock, settable white balance, the return of raw captures, etc. A few others (Galen's Fill, for example), are things that the designers just don't "get", probably because they're simply not exposed to enough serious photographer input worldwide. Yes, I've intentionally left such specs vague in this document--you'd think by now that at least one ambitious camera company would have picked up on the hints by me and others that they aren't designing to our needs. Rather than give them all the straight-out idea--though it shouldn't be that difficult to figure out--I continue to try to push them to start a design dialog with me and others, as there are many, many more things that we all wish for in our cameras that are simple yet overlooked features. The first company to start that dialog is going to get a head start on the others, and thus have a marketing advantage.

I've also left the UI specifications out of this article. Again, this shouldn't be rocket science, but as an upcoming comparison of a handful of compact digital cameras will reveal, not a single company seems to have managed to come up with the simple, intuitive design we photographers all need. Yes, we need simple, intuitive, and direct control. When you're shooting you're making hundreds of decisions, and changes to one decision can trickle down to another and yet another and still others. By throwing modes and control interactions into the mix, you essentially create UI designs that slow the decision/control process to a snail's pace, and thus produce a risk of the photographer missing the shot. The photographer can't afford that, so they start to set the "automatic" options, improving their chance of catching the "moment" but lowering the likely quality of the capture. A serious camera like I've outlined here shouldn't have that compromise. Heck, the P5000, F30, G7, and D-Lux 3 shouldn't have that compromise, but they do.

In short, I'm mad as hell and won't take it any more. You should be, too. I'll bet that every one of us would pay US$300 more for a P5000-like Coolpix that met my specs, above. Meanwhile, every compact camera creator is bemoaning the status quo. Kodak has pulled out of the lowest priced compact camera market. Nikon is losing market share faster than I can keep count of. Canon is iterating so many "same idea, new packaging" models that I'm beginning to think they'll run out of model numbers (80 models in 10 years). Sony seems to wander around in random directions, trying lots of game-like UIs on top of slightly off-center designs, as if getting the joystick interaction right will fix all their woes. Samsung seems to think outside the box on UI and design, but still doesn't get the camera right. Panasonic seems to come the closest to a photographer-centric design, but their sensor just doesn't live up to the quality of the camera design.

This isn't rocket science folks. Photography is a mature and well-known craft. Layering gimmicks upon the product perhaps makes for interesting marketing messages, but not necessarily better pictures. Those of us who care about photography want to take better pictures. That requires better cameras.

So, camera companies worldwide: are you ready to make a better camera?

Super Bonus Camera
One email I recieved made me consider an alternative design approach. So I'll give a brief outline of it here, especially since it's based upon an existing Nikon design. A couple people suggested just using two jacket pockets with a D40 and I'd have my "pocketable" high quality camera. Okay, I'll go further than that.

Take a D40x and get rid of the mirror system. Design a new lens to fit into the now open space within the camera. A properly designed 30mm f/2.8 wouldn't even stick out as far as the righthand grip (we know the existing 45mm f/2.8 wouldn't ;~); I suspect an 18-30mm f/2.8 could also be that compact, though I haven't tried to work the optics on that. Telephoto reach could be achieved easily via a screw-in option, ala the old Coolpix models.

You have two things you need to fix: (1) you need an LCD behind the viewfinder eyepiece, and (2) you need a new AF system. The first is trivial (already done with the Coolpix 8800) while the second isn't (at least judging by Nikon's Coolpix AF designs, which have been pathetic).

The result is actually about the same size as the Coolpix 8800, only with less lens protrusion. All the engineering to create it basically boils down to the lens and AF system, as everything else about the D40x would basically stay the same (okay, the viewfinder changed, but that's a straight swapout with another existing design). Heck, even from a documentation standpoint this is a no-brainer, as the current D40x manual wouldn't even need much change.

The question is whether such a camera would cannibalize D40x sales. Yes, it probably would to some degree. Faced with a choice of a US$799 D40x with 18-55mm interchangable lens and a US$799 Coolpix 40x with 18-30mm fixed lens, a good portion of folk would pick the Coolpix version. But so would a lot of pros looking for a walk-around quality compact. But again, I need to talk about design goals. If your design goal is to open up a new type of camera market, then you'd actually strip some features off the D40x (it doesn't need Scene exposure modes, it might not need a built-in pop-up flash, you just use the Classic UI and remove the hand-holding UI bits). With a bit of judicious design goal differences, you do create distinctly different products that appeal to different markets. A little pricing adjustment (D40x at US$699 and the CP40x at US$799) and you've forced the issue.

Anyway, it just goes to show that the camera companies aren't being fully creative and flexible with the products they aleady have.

And Finally...
After this article appeared, I was once again contacted by someone with the right credentials and contacts wondering about raising capital to produce something that I would design (yes, it's happened before). I'm not against such a possibility, but there's a really high hurdle here. Let's assume for a moment that I can do what I say I can do: lead a design effort to produce the best quality compact possible, and do so in minimal time. One problem is that there are quite a few intellectual property hurdles to get around. That is do-able, but potentially takes effort away from the primary design needs, potentially increasing the time to market. The bigger problem is that there is almost no hurdle for the existing camera companies to do the same product as I've defined. What you don't want to do is prove the market exists and then have the existing companies simply clobber you.

Apple's recent iPhone introduction is a good illustration here. There's no reason why Nokia, Motorola, Palm, etc., couldn't have produced the iPhone design (though Apple had an advantage in using OS-X as the base platform). They didn't, so Apple has a window of opportunity for the moment. Moreover, the other companies don't have an iPod and iTunes to put into their devices, so those other companies have some engineering, partnering, or acquisitions to do. Still, it's unrealistic for Apple to believe that the iPhone will be unique two years from now. Apple's best hope is that the momentum they gained from the initial entry (and hopefully, good solid updates) is enough to carve out a permanent swath of the cellular phone market. That's one reason why the intro hype was so important to Apple. All in all, Apple got to market with momentum, a disruptive product, and a time gap before competition appears that's desirable. That may be enough for them to succeed.

But the device we're talking about here is not a do-all disruptive device, but a singular-purpose evolutionary device. It's a little more difficult to defend that turf from the big boys. As I noted above, Nikon is already surveying customers about APS compact cameras with specs similar to what I suggest. Exactly how well could New Camera Company with one product hold off Nikon if Nikon decided to play in the same park? Not long, I think. This increases the problem for New Camera Company: it needs to be more than a one-product pony, and it needs a hook that can't easily be stolen by the existing companies. That's the primary reason why I continue to lobby existing companies for the right products rather than make it myself. Not that I wouldn't entertain leading another tech venture leap-frog effort. But it probably wouldn't be the camera I define here, as it isn't a big enough step forward to make the iPhone-like splash that's necessary to disrupt the market.

What about the Sigma?
Nine months ago Sigma announced the development of the DP1, a compact camera that uses a large, high-quality sensor (1.7x Foveon). It was clear from the original announcement that they didn't have all the details of the camera right. To their credit, they seem to have done a bit of rethink of some of the specifications, which may be why it's taking so long to make it to market. Personally, I'm hoping they get the eventual product right and I can just simplify this article to: "other camera companies, see what Sigma did."

In the Meantime...
I've been using a number of different compact digital cameras trying to find the one that comes closest to my needs. That "closest" is still a long way from being right. At present, I can only recommend four candidates, and even then only half-heartedly. In order of preference:

1. Leica D-Lux 3 (also available as Panasonic LX2). From a control standpoint, LeicaPano gets most things right. For running manual controls while shooting this camera shows it was designed by a photographer, whereas the rest appear to have been designed by Play Station junkies. The menus and deeper UI leave a lot to be desired though (try figuring out how format a card--at no point does LeicaPano use the word "format"!). The LCD is good. The speed is good. The shooting controls are excellent. The lens is outstanding, trumping everything else by a long shot. Unfortunately, the pixels are mush. Even at the base ISO looking at a 100% view you'll have a hard time seeing edges. Noise and NR conspire to ruin the low-level integrity of the data, making this really a camera that can't go beyond a modest print size. Still, my favorite of the bunch, by a fair margin. (It's been pointed out to me that the predecessor model had better pixel integrity. This brings up another sad truth: camera companies aren't so much interested in image quality as they are in marketing features ["more megapixels!"])

2. Canon G7. All around decent, but seemingly excels at nothing. I shoot slightly more slowly with this than the Leica. The lens is not as good as the Leica's, but I'm much happier with the underlying pixel data, at least at the base ISO. The lens lets this camera down (too much CA and distortion, and it's a little slow in aperture). As others have written, it seems like the G7 is a step backward from the previous G series models--less professional and more amateur-oriented seems to be the design goal that was followed. I'll bet that didn't increase sales one bit.

3. Coolpix P5000. Overall, a disappointment. The biggest weakness is focus, which is slow, not manually controllable like the Leica, and often hunts uncontrollably in close up mode. The lens also is a bit underwhelming and slow in aperture at pretty much all but the wide end. On the flip side, it's small and has an optical finder, and the wide angle lens accessory is nice (though a bit dramatic on the front of the miniscule camera). As you might note from the front page picture as I post this, I've found I like the P5000 as a B&W camera. The one thing Nikon did that resonates is allow the full customization of B&W options, including virtual color filtration. Why they followed through on this but not on other functions is simply beyond my ability to understand; the design feels incomplete.

4. Fujifilm F30. The one-trick pony. If it's high ISO you need (indoor shooting without flash), this is the best of breed, without question. Fast startup and focus help, too. But the camera is let down by its proclivity to produce objectionable highlight detail and its limited manual control. You also won't like the use of xD cards. And, again, the lens is a wimp.

A number of people pointed me to the Ricoh GX100 or GR-D. These cameras have a decent user interface (though not quite as good as some suggest--the reason why the command dial should be at the rear of the camera and flat to the top versus on top of the camera and perpendicular has to do with the speed at which you can go from setting a manual function to shooting; any design that takes your finger off the shutter release is sub-optimal, IMHO). There are a few eyebrow raising choices on the camera, including the ability to set apertures that absolutely will have visible diffraction in them, and again the sensor is the low point in the design. The design target seems about right for this camera, but unfortunately the image quality doesn't come close to clearing the hurdle I want cleared.



bythom.com | Nikon | Gadgets | Writing | imho | Travel | Privacy statement | 2007 Thom Hogan. All rights reserved.