Entry 8, 20 Feb: Death of a multi-millenium paradigm

I had intended to stick to adding to this blog at least once a week but got side-tracked by work on a tutorial. One day I may even "publish" it. But today, back to the ol' blog slog:

An essay on a major photography site recently referenced the 10,000 hour rule, which in turn set off a squall of forum posts as to the nature vs. nurture of talent. Few seem to realize that there is a 10,000 hour rule for audiences as well as for artists and performers.

I had occasion to show a selection of my best prints to a person who has been doing photography at the casual level for decades. I put him in a chair in a location with strong 5000K lighting, then did an informal test for near-vision acuity by asking him to read some fine print. No problem. I then let him work his way through a box of some two dozen 17x22 inch prints with no leading on my part. Both his comments and his time spent per picture made it clear which ones particularly grabbed. What was also clear was that in spite his ability to see how images were rendered on paper he consistently paid zero attention to that aspect of the pictures. In fact, he had a bit of an epiphany that he could tell which pictures were going to grab him in the first fraction of a second of seeing them.

In each case he looked through the print to the subject matter. If I had printed any of them at 100 PPI he might have noticed. Two of the prints were multi-shot stitchings with exceptional, wrong-end-of-telescope clarity. I may as well have down-sampled them to 150 PPI before printing. There were prints with meta-textural rendering that I would feel comfortable showing next to a Rembrandt. Unseen and unnoticed.

This is in no way a criticism of this particular person (who can identify grapes and vintages with alarming accuracy). I'm not even talking about a general lack of visual perceptual refinement in the population. My point is that it takes a large investment in time and attention to achieve competence in art appreciation as to achieve competence at art production. In the (perhaps vast) majority of cases those who achieve that level of competence are themselves artists. It would not be much of a stretch to modify the cliché from "art for art's sake" to "art for other artists' sake".

This might suggest that anyone who has achieved excellence in photography can expect anyone else who has achieved excellence in photography to keenly appreciate what he is doing. But even among that small subset of the population, there are photographers who do documentary work and photographers who do art. And among those who do art photography, there are realists, pictorialists, expressionists, surrealists, abstractionists, and conceptualists – each a niche with its own microcosm of challenges and concerns. My guess is that there are at least 1000 people on the planet who have invested the necessary time and attention in just the right visual specializations to "get" what I'm doing. And of those 1000, maybe 10 or 20 happen to have the necessary human characteristics to also feel passionate about the same subject matter that I feel passionate about.

Another comment the particular viewer I mentioned above made while looking at the prints was that it was a shame that to date I hadn't had some sort of public showing of my work. But in fact I have a permanent one-person exhibition on display in every city, town, and hamlet in the world. It's called a web site. The whole point of my little print-showing experiment is that the difference in rendering between a large, finely crafted print on cotton rag and the travesty that is a web-sized JPEG image is irrelevant to 99.99% of all possible viewers. Not only can anyone in the world gain admittance to my permanent exhibit for free, they can bookmark it for repeat viewings or even right-click to download (and therefore take possession of) any picture that appeals. For free.

Probably 990 of each 1000 visitors to my site arrive here upon searching for a particular equipment review or tutorial on a particular photography-related topic. Of those 1000, maybe 15 or 20 ever go on to take more than a glance at the photo galleries. The vast majority of viewers get to this site because they are interested in picture-taking, yet they are mostly uninterested in picture viewing. Superficially, this makes no sense. People who do photography already have a large body of pictures perfectly suited to their personal preferences to look at – namely the pictures they take/make themselves. If a photographer is going to bother to look at anyone else's pictures, it's going to be someone who is at least interested in the same photographic specialty. Birders will look at pictures by other birders. Macro photographers will look at pictures by other macro photographers.

Let's take a seeming detour. I've lately been reading the blog of a certain pro photographer. I've read forum posts by other pro photographers. If there is one theme that runs more consistently through what they write than whether to use medium format or dSLRs, it's how much faster they have to tap dance each year in order to stave off extinction. Fewer weddings are shot by paid professionals. Few magazine spreads ditto. The world was already saturated with amateur film cameras before 2000 AD, but certain practicalities of post-production and distribution permitted a pro-based economic niche to persist. Those practicalities disappeared when digital photography appeared. All that's left keeping many a professional employed is inertia, and there's not much even of that.

Somewhere roughly between two and three thousand years ago a rather peculiar idea originated in both western and eastern civilizations – that there were certain people who specialized in the production of visual recordings of recognizable scenes. As these recordings became more skillful and more realistic the institution of the visual artisan began to gel. In order to acquire the necessary skill an artisan needed to devote large swaths of time to the practice of his craft. This was not possible when hand-to-mouth survival dominated life. It because barely possible when empires allowed the existence of a parasitical leisure class to act as employers.

Due to industrialization the wealth of whole populations has increased to the point that both patrons and artisans came to exist in very large numbers. But now the pattern in the kaleidoscope has shifted and re-formed once again. All this hinged on the ability by a relative few to produce visual recordings of recognizable scenes. Now everyone and his grandmother can do just that with a $200 digital camera, not to mention using a mere smart phone.

If anyone cared about the actual quality of these visual recordings, then pro and art photography might have a chance to survive or even thrive. In the vast majority of instances what people care about is whether subject matter is easily recognizable or not; and even the most humble digital camera can pass that test in the large majority of situations. Image stabilization, high ISOs, auto focus, auto exposure. These are enough to get the job done. Not well, not connoisseur thrillingly. But adequately.

So we are today living in a world in which everyone is a photographer and anyone who cares to play in that sandbox is an "art" photographer Those who want visual recordings can make visual recordings themselves and do so at the level of quality/competence they have need for. Those few who want to own art photographs are happily engaged in producing their own art photographs.

I can spend my time and energy attempting to tap-dance even faster than the world's best tap-dancers in search of that ever more microscopic piece of the appreciative audience pie. Or I can resign myself to maximally pleasing a guaranteed knowledgeable and sympathetic audience containing exactly one member. I'm too old to tap-dance, so I've resigned myself to the second alternative. I have a box containing several dozen exquisitely seen and rendered photographic prints. I know they're exquisitely seen and rendered because my audience unanimously and vociferously agrees to that proposition.

Glad I'm not trying to make a living as a tap-dancer. At my age my heart wouldn't be able to take the pace.

---

Feedback: Please e-mail any feedback to webmaster@daystarvisions.com. I will try to post all but the most scurrilous comments and reply as appropriate on this page. Please indicate whatever alias you wish your comments to appear under in your message; actual names will never be used unless you expressly request that I do so.

---

Return to top of page

 

Entry 7, 11 Feb: Trying to Comprehend the Nikon D800

People have been saying for several years now that megapixels don't matter and that X megapixels are sufficient, where X = 6, 10, 20 or whatever. In a previous blog entry I addressed that point in theory. Now that a relatively affordable 36 MP is a reality I have a more visceral reaction. I personally don't need 36 MP, or even 24 MP. But ever since I started digital photography in 2000 I've felt like there weren't enough MP. Now I feel like I've been stooped over for the past 13 years living under too low a ceiling and finally I can stand up straight. The fact that – for me – 36 MP is an excessive number is beautiful. I can now shop for any MP count that appeals to me and do so without having to re-mortgage my home.

Perhaps more amazingly, this will be a CMOS implementation with an approximately 5 micron pixel size which means excellent DR and at least good high ISO. For me, it will never again be about when will that affordable, sufficiently high IQ camera come to exist, but when will it appear in a carry-all-day package?

But there is another question sparked by the D800 that in certain circles is raging like a wild fire: Does the D800 make those super-expensive digital medium format cameras and backs obsolete? If not, does it force them to a more affordable price point?

The most common argument that dMF cameras in the 40 and under MP range will necessarily have magically better IQ because bigger pixels always trump smaller ones. The pixel size of the Phase One IQ180 – surely the most-coveted MF sensor going – is 5.17 microns. the most likely pixel size for the D800 is 4.9 microns. Anyone want to argue that the 1/4 micron difference is going to make a meaningful difference in favour of dMF? The real question is whether a 5 micron pixel from the older CCD technology can come any where close to the performance of a state-of-the-art CMOS pixel? If it were as simple as bigger pixel = better pixel, then image files from the $2500 Nikon D700 with its 8.4 pixel pitch would necessarily trump image files from the $40,000 IQ180, not to mention the 5.93 micron pixels of the $10,000, 40 MP Pentax 645D.

If you click on this DxOMark link, then click on the Measurements tab, then click on the SNR 18% tab, you'll see a significantly better performance from the two medium format sensors than the Nikon D7000. That matters because the Nikon D800 will almost certainly perform very similarly to the D7000: it has more pixels, but unlikely dramatically better pixels (somewhat better, since Nikon will use its best sensor support components). In terms that I can understand, the SNR graph says that recent dMF cameras have measurably cleaner image files than the D800 is most likely to have ... and the difference does not simply scale with pixel size.

But that's not the end of the story. Click on the DR (dynamic range) tab to see that the D7000 bests both dMFs in DR, which is just as important an IQ factor as cleanness. Click on the Color Sensitivity tab to see that dMF retains an edge in bit depth. Again: these are not D800 numbers and DxOMark is not uncontested as an arbiter in these matters. What I feel I can assert with a reasonable expectation of accuracy is that no one will be able to demonstrate the superiority of 40 MP and under dMF prints at any size over prints from the D800 at the same size. If the most veteran MF shooters can even distinguish dMF prints from D800 prints in a double-blind test situation, I'm pretty confident it will be due to other factors than image quality, such as DOF.

I'm sure there are many other factors than IQ that will keep dMF photogs in the fold. Shallow DOF and low base ISO are two that spring immediately to mind. Similarly, the D800 and its inevitable competition from Canon and Sony will have an inherent functional flexibility that will make them preferable to dMFs for many tasks even without the price differential.

I also asked whether the D800 will force dMFs to a more affordable price point. This is a crystal ball question; unfortunately mine is experiencing a strong bout of digital noise right now ... can't quite make out the answer...

Return to top of page

 

Entry 6, 09 Feb: Yet more on the OM-D/E-M5

I've read quite a bit of commentary on the E-M5 AKA OM-D since its official announcement. I'm rather baffled by the lack of commentary on two features. If the new 5-axis in-body IS lets me handhold at even one stop slower a shutter speed than before, that's a big deal to me. If "the world's fastest AF" results in the sort of decisive moment responsiveness of an all-manual 35mm camera from 30 or 40 years ago, that's another big deal to me.

Update, 13 Feb: Yet another big deal is not new or unique to the E-M5 but shared by it with all other mirrorless cameras. Thanks to the mirrorless design these cameras have real-time exposure information both as a live histogram and as an accurate presentation of the scene on LCD or EVF. Formerly, we might have counted this as compensation for the CD-AF downside of slow AF speed and the EVF downside of refresh lag and between-shot black-outs. The high speed AF and twin image processing chips in the E-M5 tantalize that we may now be getting real-time exposure presentation without paying the expected penalty. That's the sort of thing that puts some meat on the bones that mirrorless is destined to put an end to the 60 year SLR reign.

In my last post I was pretty hard on the probable IQ of the E-M5, if only on purely theoretical grounds. Yes; I too would love to see my dSLR's IQ shoe-horned into the E-M5's smaller form factor. The official specs are now out and they list the sensor as being Live MOS, Live MOS means Panasonic. Of course, we have to wait for test results and owner feedback, but if you want to believe that Panasonic gave Olympus a better sensor than any they themselves have used, I'm sure I can still get you a great deal on the Brooklyn Bridge if you act quickly.

The sensor is not the only critical IQ component in digital image creation by any means. We can count on Olympus to squeeze every possible drop of goodness out of whatever sensor they've got to work with. But I wouldn't expect that to translate into more than a 1/3 stop DR, for example. If Panasonic's most recent cameras are still plagued with pattern noise, you can confidently expect Olympus to do better. But I don't expect the beautiful, 13.9 DXO ISO numbers or anything close that we see in Sony's latest 16 MP APS-C sensor.

Now let's flip that on its head. I recently bought a Canon S100 pocket P&S, in theory as sidekick to my dSLR. The S100 has a pea-size, 7.6x5.7mm sensor inside – that's just 1/5 the light gathering area of even the 4/3rds sensor inside the E-M5. Yet if you could have bought a roll of 35mm slide film before the digital era with the IQ of the S100's image files you would have thought you'd died and gone to photography heaven. The S100 is a 12 MP camera with a superbly tuned lens in front, so resolution and acutance are likely every bit as good as the best B&W neg film ever made. Yet this is a colour "material", not B&W, and the colour is just as good as you care to make it. Slide film rarely exceeded 5 stops DR, but the most conservative estimate for the S100 would be 7 or 8 stops of practical DR at base ISO. Unlike slide film but much like neg film, the S100 can give you usable high ISO output up in the 800 and even 1600 range. If we could say that the inherent limitations of 35mm film resulted in laughable photographs by today's standards, then we could dismiss the S100 as a photographic tool. Instead, I'm so pleased with the images the S100 is giving me that my dSLR is gathering dust. And if I can say that about a sensor 1/5 the size of the sensor inside every 4/3rds camera going, surely I can say even better things about Olympus' latest no-holds-barred 4/3rds offering, the E-M5?

Perhaps the way to look at it is this: We all know the bell curve and what it signifies. The more powerful the digital sensor sub-system within a digital camera (or the film in a film camera) the further to the left and right extremes of the bell curve that sensor will take you. Not just shots in afternoon sunshine with the sun over your shoulder, but contra-jour shots. Not just contra-jour, but twilght. Not just twilight, but available darkness. It takes sheer sensor size to push a camera's reach to the furthest outliers of the imaging bell curve. The bigger the sensor, the bigger the camera body and lenses at any given focal length. If your camera is too big, that starts to limit your ability to work on those bell curve extremities every bit as much as lack of DR or lack of high ISO. War correspondents don't use an IQ180 digital back on an Alpa technical body for a reason. Ditto sports photogs, if for somewhat different reasons.

If your uses for a camera cluster around the big central bulge of the photographic bell curve, you have an easy time shopping for camera equipment. To the degree that you want or need to push the envelope in any of many directions, such as shot-to-shot speed, stealthy size and silence, carry-all-day low weight, huge DR, available darkness, physical robustness, weather sealing, etc., to that degree you need to get particular about particular specs and particular details. To dismiss the E-M5, or any camera out of hand, due to inherent IQ limitations is only defensible if you do a type of shooting that actually requires more DR, more low light cleanness, etc., than the camera in question can produce.

Circling back to the beginning of this post: If the E-M5 can give me substantively better IS plus substantively more responsiveness in such a small camera body, I'm suddenly quite interested. The fact that it's IQ is guaranteed better than that of my S100 P&S that I use daily makes it all the more interesting. The fact that it has sort of a pseudo-retro styling is of no great interest. The fact that it has a full adult control set with Olympus' endless depth of menu-driven customization is far more than icing on the cake. I personally, am not in the market for a new camera, being far too happy with the dSLR + P&S combo I already have. But I can see the E-M5 being of considerable interest to a whole lot of people. And if the promise of the new IS and AF hold true, I can also see it being responsible for some Pulitzer Prize winning photographs bye the bye.

Up next: my thoughts on the Nikon D800. Sure to be a hoot, if not a holler, so stay tuned...

Return to top of page

 

Entry 5, 06 Feb: More on the OM-D

The OM-D is a new line of cameras Olympus will officially announce on Wednesday, but in the meantime has been systematically leaking details about in what it believes is good interest-generating policy.

There are two aspects of a certain generation of all-metal, all-manual 35mm cameras that make those of us who have used them long for the good old days. One was the interface simplicity; and I went on at great length in my previous post about the OM-D about why that isn't coming back any time soon. The other aspect is the sheer physicality of those cameras, including size, weight, and solidity. It's this physicality aspect that Olympus is clearly aiming to riff on. The 35mm OM series was definitely at the pinnacle of that design eminence. If you've never used a camera like an Olympus OM or Nikon FM, then any fuss surrounding the OM-D will probably pass you by. Similarly, if you've never used a Leica M style rangefinder, then the fuss around the Fuji X10 and now the X Pro 1 will likely pass you by too.

Best I can say is that we're talking about a camera as a physical object that your hands want to handle; plus controls that have a sensuality that your fingers want to manipulate; plus a view finder with a big bright optical image that your eyes want to look at. To the degree that the OM-D lives up to that standard, it will speak loud and clear to a certain segment of the camera market.

Unfortunately, cameras are not just sculptures but also have to take pictures. Now that buying a camera locks us into the equivalent of a single size, type, and brand of film it's of course even more important that a new camera deliver on that front as well. So when we hear that the OM-D will have a 4/3rds sensor inside we've already got a bias against it that it will have to overcome ... if it can. This is the same thing happened with Nikon's 1 series release. The smaller-than-APS-C 4/3rds sensor size makes perfect sense when the object is to build a smaller-than-APS-C camera system. But in this case minimal size is not the goal.

Olympus' previous digital Pen introduction mimicked the size and styling of their earlier film Pen line, which indeed used a half-scale film size to achieve smallness. Physically, the OM-D mimicks a 35mm SLR camera. This is what we now call full frame, and full frame is what the appearance of the OM-D suggests. Instead, Olympus (pretty much out of necessity) opts for a sensor size that is smaller even than APS-C, let alone FF. And thus Olympus cuts the light gathering area by 40% from 370 sq mm to 225 sq mm, which can only exact a price in that most critical of all factors: the ratio of signal to "noise" (photons recorded versus pixels that record no or incorrect information about the scene).

The good news is that the sensor inside the OM-D is going to be 16 MP. That's plenty of resolution for a handheld camera and as much as most of us are ever likely to want even for tripod work. Those as know say Olympus makes good glass, so resolution is not likely to be an issue.

But that 40% loss in light gathering area is not a handicap that can be overcome when it comes to dynamic range and low light sensitivity. Of course, simple surface area is only one factor that affects light-gathering capability. An APS-C sized sensor with poor performance might well lag behind a 4/3rds sensor with better performance. Trouble is: that's not going to happen. About the best scenario would be for Olympus to have scored a coup by getting Sony (a direct competitior) to manufacture the 16 MP 4/3 sensor for Olympus using Sony's latest sensor technology. The up-front cost for Sony just to engineer and cut this different sensor size for the first time would probably kill the deal. But beyond that the sensor would have to be specially designed for the sorts or read-outs that Olympus requires. In practice it's far more likely that Olympus will continue to go hat-in-hand to Panasonic. And Panasonic has already demonstrated that they have no interest in sharing their latest generation sensor. technology with direct-competitor Olympus. Olympus gets hand-me-downs or gets nothing.

Assuming the above analysis is correct, then we already know the ballpark into which the OM-D sensor will fall: in DXOmark terms it will have a base DR of about 11 stops, an ISO score of about 700, and a colour depth score of about 21 bits. This compares to Sony's current APS-C sensor which ranks at 14 stops DR, 1100 ISO, and 23-bit colour depth.

Now for many areas in the photographic microcosm this differential is not critical. So long as you give the OM-D sensor a reasonable amount of light it's going to repay you with 16 MP of very usable IQ. Shoot in daylight, or in a studio, or use flash and you're doing fine. When you start shooting twilight, night concerts, bars, restaurants, etc. in available light, you're going to notice the difference. Even so, one thing the DXO numbers don't measure is something we may ironically call Quality of Noise. One thing Olympus excels at is freedom from the great nemesis of the low light shooter: shadow noise banding, which is the very devil to eliminate in post-processing. Another thing Olympus excels at is consistently producing a notch more pleasant in-camera JPEG colour than the competition. Or so I'm told – they all look equally wretched to me.

Above, I mentioned a big, bright optical view finder as one of the hallmarks of the 35mm OM series and it's brethern. We already "know" that the OM-D will incorporate the guts of Olympus' VF-2 EVF as its finder. I have yet to own and use an EVF camera, but my reading tells me this is a very decent EVF, all-told about as usable as the new-fangled OLED one currently Sony touts. So most likely no one who disdains EVFs is going to be converted, and no one who loves them is going to totally disappointed. Speaking of the OM-D's finder, one thing we don't know is the purpose of the large pentaprism-style hump on top of the EVF. It clearly doesn't house a pop-up flash.

Return to top of page

 

Entry 4, 03 Feb: Dale's mirrorless crib-sheet

For the past several months we've been bombarded by a relentless stream of new mirrorless camera announcements. Many include a new lens mount, which is stupefying, not to mention something that didn't even happen during the film-to-digital transition. It's all rather helter-skelter but it seems pretty clear by now that we're in the middle of a full version transition in photography. Digital photography version 1 was built around usually finderless compact cameras with small sensors; plus APS-C and FF dSLRs. Digital photography version 2 looks to be built around camera phones and EVF mirrorless. EVF is a technology that is progressing rapidly but isn't completely functional as yet.

IAC, I've read the specs and hype about each new model, just like anyone else. So in self-defense I've started compiling a list of models and my initial evaluations of each. I compiled this for my own use in trying to keep track of what each of these models is about when reading blogs, forum posts, etc. If it proves useful to anyone else all the better. I have neither interest nor expertise in matters related to either flash or video recording, so you'll have to add those details for yourself. Also: my perspective will only be useful to you if you insist of full control of the picture-taking process, and see mirrorless as being primarily a vehicle for snapshots and decisive moment work, rather than for traditionally tripod-based shooting.

Note: The second line under each model header gives the megapixel count followed by the DXOmark results for cameras they've tested.

Canon

G1X

• 117 x 81 x 65 mm (4.61 x 3.19 x 2.56"), 534g
• MP:14.3 Clr:21.7 DR:10.8 ISO: 644
• Body on the big side.
• Quasi-4/3 sensor size (1.5" = 18.7 x 14 mm) sensor) with 1.85x crop factor.
• Should have excellent image quality within confines of 4/3-like DR and sensitivity.
• Non-interchangeable lens going to be a turn-off for many including the fast primes crowd. This does have an upside: lens will be highly tuned to sensor; no sensor cleaning events; no temptation to buy and tote multiple lenses.

Fuji

X100

• 125.5 x 74.4 x 53.9mm (5.0in. x 2.9in. x 2.1 in.), 445g
• MP:12.3 Clr:22.9 DR:12.4 ISO: 1001
• APS-C sensor, 12.3 MP.
• Fixed 35mm-e focal length lens but perfectly matched to finder.
• Hybrid optical + EVF finder.
• Adult control set.

X10

• 117.0 x 69.6 x 56.8 mm (4.6 x 2.7 x 2.2 in.), 350g
• MP:12 Clr:20.5 DR:11.3 ISO: 245
• 2/3" Exmoor sensor, 12 MP
• Small sensor but not pocketable.
• The Fuji sensor thing, like Foveon, is big on hype but most of the benefits kick in when resolution is dropped from 12 to 6 MP. It's also a nightmare for raw developers to cope with.
• Reasonable control set.

X Pro 1

• 139.5 x 81.8 x 42.6 mm (5.5 x 3.2 x 1.7 in.), 450g
• APS-C X-Trans (non-Bayer) sensor, 16.3 MP
• New, all-electronic lens mount.
• Attractive but largish compared to other mirrorless cams. Clearly an attempt to duplicate the Leica M body dimensions and retain the beloved range finder aesthetic.
• Non-Bayer pattern sensor with no AA filter promises slight bump in resolution
• Looks to be a good control set, but only 2 stops exp. comp.
• Apparently, X100 users have mixed feelings about the hybrid finder.
• The in-body finder magnification switching thing is clever. Only two mag. choices, but switching to EVF mode fills finder with image from any focal length.
• Lack of in-body IS is biggest down check for me; less an issue for those who love fast primes, but still that crop factor issue for legacy FF lenses.
• Emphasis on prime lenses is not my style but appropriate for majority of people attracted to the RF design.
• I suspect the relatively high price point has a lot to do with establishing cachet. It's not a camera that's ever going to appeal to the mainstream, so Fuji is deliberately targeting a lower economy of scale business model.

Nikon

V1

• 113 x 76 x 44 mm (4.45 x 2.99 x 1.71") , 383g
• MP:10 Clr:21.2 DR:11 ISO: 346
• Body and lens too big given 2.7 crop factor
• 10 MP may now be too few for many; but DR at least is comparable to rest of pack.
• Lack of controls.

Olympus

OM-D / E-M5

• 122 x 89 x 43 mm (4.8 x 3.5 x 1.69"), 425g
• MP:10 Clr:? DR:? ISO: ?
• E-M5 is first model of new OM-D line, essentially the "pro" complement to the Pen line, using the same lens mount. The OM-D line is not SLR but is instead mirrorless. Body styling will resemble Olympus' OM film SLR line, as you can see here, but is somewhat smaller and stubbier than the 35mm OM bodies.
• Finally Olympus gets a 16 MP sensor. Although this is manufactured by Panasonic and is likely related to the sensor used in the Panasonic G3, based on past experience, they will squeeze the last little bit of image quality out of it.
• Unlikely to have APS-C or better DR or sensitivity but will come closer than Panasonic.
• World's fastest AF, but will other performance numbers be up to par for action shooting?
• 5-axis IS could be significant breakthrough.
• Weather-proofed body and kit lens.
• EVF similar to existing VF-2 but with faster updating.
• As always, Olympus provides an excellent suite of adult controls.

E-PM1

• 110x64x34mm (4.33x2.52x1.34"), 263g
• MP:12 Clr:21.0 DR:10.3 ISO: 500
• Smallest mirrorless camera, smallest kit zoom.
• All-around solid feature set.
• Low price point.
• 12 MP sounds low but any more is wasted for hand-holded shooting.
• Somewhat limited DR and sensitvity.
• User complaints of IS problems proved to confined to shooting with optional EVF (ditto for E-PL3).
• Controls adequate for Av shooting.
• Full customization.
• Relatively inexpensive optional EVF.
• Superb lens selection.
• No orientation sensor.
• Non-OLED 460 LCD.

E-PL3

• 110x64x37mm (4.33x2.52x1.46"), 313g
• MP:12 Clr:21.0 DR:10.3 ISO: 500
• Same as E-PM1 but slightly thicker body due to moveable LCD.

E-P3

• 122 x 69 x 34 mm (4.8 x 2.72 x 1.34"), 369g
• MP:12 Clr:20.8 DR:10.1 ISO: 536
• Same as above but with full physical controls.
• Size and weight of body are not smallest but not the largest in class.

Panasonic

GX1

• 116x68x39mm (4.58x2.67x1.55"), 313g
• MP:15 Clr:20.8 DR:10.6 ISO: 703
• IS in lenses so lenses either large or have no IS.
• Panasonic has a history of really poor noise handling, including high-ISO banding. I'll have to investigate one of these days whether they've overcome that or not.

Pentax

Poor Pentax! This is a company I love the spirit of but feel sorry for given their handicaps. Getting Sony to give them the 16 MP APS-C sensor while still current was a coup. Getting someone to give them a usable EVF seems to be a different matter. Hopefully, they'll be allowed to play ball with microOLED.

Note: see reader response to above paragraph below.

K-01

• 122 x 79 x 58 mm (4.8 x 3.11 x 2.28"), 561g
• Uses Pentax's existing APS-C lens. Thus, camera body has to be thick since these lenses were designed to focus on a point behind a (now-non-existent) mirror box.
• No finder! LCD panel only.
• Heavy, fairly big.
• AF likely to be slow.
• Will have best-in-class image quality.

Sony

NEX5N

• 111x59x38mm (4.37x2.32x1.50"), 269g
• MP:16 Clr:23.6 DR:12.7 ISO:1079
• IS in lenses so lenses either large or have no IS.
• Lacks adult controls.

NEX7

• 120x67x43mm (4.72x2.64x1.69"), 400g
• MP:24 Clr:24.0 DR:13.4 ISO:1016
• IS in lenses so lenses either large or have no IS.
• Small lens selection esp. when looking to match 24 MP sensor.

 

See also: DPReview's Mirrorless Round-up.

---

Responses

I noticed your comments about Pentax and their new K-01. While I personally have no interest in this format and this particular camera, I think your comments about its Sony sensor are misplaced. That 16 MP sensor was sold to both Pentax and Nikon. Sony sells its sensors to many Japanese camera companies (many P&S) and it is purely a business proposition. As you know, it is called outsourcing and components in all cameras are made by different companies for a profit through various business agreements. Canon does and they all do it. Sony happens to have huge R&D facilities along with large production capacity to market their components (in this case their sensors) to other firms.

No firm is "giving" their components to "Poor" Pentax. It is strictly business and Japanese companies have a long tradition of doing business with one another. As of April 1st, Ricoh will be merging its imaging into just one division under the brand of Pentax – Ricoh Imaging Company. This restructuring will bring significant resources to Pentax and place them in a very competitive position going forward. There is absolutely no doubt in my mind that Pentax will continue to make fine products of which the K5 is a superb example. New lenses are already in the can among other initiatives. They will continue to be a player for a long, long time as well they should.

Cheers,

Guy

Return to top of page

 

Entry 3, 30 Jan: Behind the scenes with some recent pictures

I had intended to write this post encouraging people to pursue excellence at just one craft in their lives, whether dill pickling or egg tempera painting. I've long felt this elevates a person out of shallowness and opens up a whole universe of sensitivity to subtlety and depth. But when I actually set about trying to explain why that is a better way to live than the more casual, nothing-in-excess, good-friends-good-times approach that most people pursue ... I realized I was just trying to foist my own prejudice on to other people. So scratch that idea.

Instead, let's go visual. Below are four pictures I'm currently infatuated with; all taken with a Canon S100 pocket camera I reviewed on this site recently. Of course, blurry pictures – even deliberately blurry pictures – are not going to be to everyone's taste. I don't usually slather a layer of verbosity over my pictures, expecting them to stand firmly on their own two feet. But in this case let's peek behind the curtain to see The Artist At Work. (Drum-roll, please.)

As always, click on any picture with a light border for a larger version.

Fig. 1 is pretty much as shot except for contrast, mainly to deepen the shadows for more impact. This is a good example of what drew me to pursue camera blur. The scene if shot straight, is a typical suburban/residential street on a sunny day. Perhaps pleasant but certainly dismissible. To my eyes the blur adds grace and poetry ... perhaps an acquired taste. ;) If I were a gallery staff person trying to drum up a sale, I'd point out that the cheerful late afternoon sunlight+long shadows mood is in no way obscured by the stylistic device of the camera blur. If more were needed, I'd then point out the compositional energy generated by the bow-like sweep of the sidewalk crossed by the arrow-like thrust of the driveway and white car pointing out-of-frame to the left.

Doesn't quite work for you? Oh well: over here we have some nice contemporary impressionism canvasses by a little old lady living in Tulsa... ;)

If camera blur can rescue an otherwise bourgeois scene, can it do so on a drab rainy day? For me at least, Fig. 2 answers the question in the affirmative.

But unlike Fig. 1, Fig. 2 required extensive post-processing. Fig. 3 shows what the image looked like after initial adjustments in the raw converter. Not a whole lot to love at first blush is there? As with a newborn babe your first reaction to the slimy, red-face, wrinkled blob might well be "No thanks! Let's send this one back and try again". The trick (at least where art is concerned) is to isolate individual faults and correct each one in turn.

Contrast curves: For those not up on image-editing. A contrast curve is simply a way to make the darks darker and/or the lights lighter in all or part of a picture.

The first thing that had to go was the blank glaring area that is the entire upper right quadrant. Since the sky is not actually blown out, I was able to apply successive contrast-exaggerating curves to exploit the values that were present, if just barely. Then I needed to change the sky colour to something that rhymed with that of the concrete. There is also a magenta cast to Fig. 3 that is more apparent in Photoshop than in the JPEG. Trouble was I wanted parts of that cast, so had to selectively remove it where I didn't want it. Finally, the grass is a bit amorphous, requiring a combination of contrast curves, plus a really huge dose of LCE (USM-based local contrast enhancement) applied to the whole image.

Fig. 4 was taken on a commuter train trip a few minutes before pulling into the main terminal in downtown Toronto. You can see a fine cross-hatching in the foreground gravel from the train's motion. But the train was moving too slowly to blur the rest of the scene, so I had to do that myself. About the only post-processing required was to reduce brightness and exaggerate contrast in the sky area above the buildings. I'm a contrasty kind of guy, but the orange of the brick buildings adamantly resisted all futzing. And, yes: I did try to de-tilt the image so the rail runs parallel to the frame. But that just sucked all energy out of the composition; so the tilt stays and I'll meet anyone who disagrees in the alley behind those orange buildings after the show.

Fig. 5 asks the ultimate question: does one even have to leave the house to find a fertile source of usable subject matter? (Or: just how lazy can a guy be and still do photography?)

As Fig. 6 makes clear, Fig. 5 is the result of even more extensive handy work. For one thing the hexagonal window is several stops blown, and since this was necessarily a handheld shot with added camera motion, doing a second shot exposed for the outdoor scene was more than a challenge. I made my best stab, then used Photoshop's shape transform feature to match the distortion of the second shot to that of the first.

Up close and personal: If you click on Fig. 7 you get an unreduced look at a section of the image file. I think you can see that, even though the S100 is a tiny camera with a pipsqueak sensor inside, if you give it anything like enough light it's capable of returning the favour with a sumptuous helping of nuance, both of colour and of detail. Yes: there is a fine speckling of luminance-only noise reduction to be seen in the shadow areas. That's purely my choice made during the raw development stage. It's automatically removed if you shoot in-camera JPEGs; and it's easy enough to remove in raw development if you don't want it.

For the rest of the image, after adding some global contrast to deepen the shadows, I worked quadrant by quadrant. This involved especially darkening the upper left and lower right quadrants, but also darkening the pale green cover on the bed to stop the lower left quadrant from being one fat blob, while keeping it light enough to rhyme with the upper right quadrant. The result of these changes loses much of the bright airy feeling that was part of what drew me to this exposure in the first place. As best I could figure out that was a necessary sacrifice to make the picture work as a whole.

Or does it work? Your comments explaining how I'm a pretentious airbag with delusions of artistic merit will be invaluable in helping me crop back the kudzu-like excesses of my ego.

Return to top of page

 

Entry 2, 26 Jan: Standardizing the digital camera interface

In the previous post I took the imminent appearance of a new model retro camera as a springboard to launch on a rant against current digital camera interface design. And especially on the lack of consistency from one model camera to the next. What would a digital camera interface that is both consistent and usable look like? It just so happens I have the answer all figured out. ;)

The things you absolutely have to control in any camera since the Kodak Brownie are aperture, focus, focal length if zoom, shutter speed, and sensitivity/gain/ISO. Add flash, if the camera has a built-in flash. In a digital camera you also need a whole host of computer controls, such as playback, delete image, and invoke menu.

Looking first at the physical picture-taking controls, the following principles apply.

1. Many of these controls toggle between multiple functions. Yes: I understand that the current fad is to have a separate, dedicated control for each function, but this approach has serious disadvantages. It makes the camera more intimidating to the new user; It makes it more difficult for your fingers to learn to operate the camera without you having to look at camera instead of at the scene. It makes it more difficult or impossible to retain interface consistency when moving from larger to smaller camera bodies. Look how much less intimidating Fig. 2 is than Fig. 1, yet all the functionality you need is available without diving into a menu.
2. The function assignment you choose for a given control is retained until you change it. The function assignment doesn't change automatically when you turn the camera off then on again or when you change to a different mode dial position. For example, if you set the top right control wheel to control ISO, it doesn't revert to controlling shutter speed each time you turn the camera off then on or if you change the mode dial from P to A or whatever.
3. Visual feedback re which function assignment you've chosen and what you're doing to that function appears in the view finder and/or back monitor. For example, if you set the top right control wheel to control shutter speed, you see a portion of the shutter speed choices (such as 15 .. 30 .. 60 .. 120) in the finder and/or back monitor, and of course the current shutter speed setting is marked.

The above schematic is meant to represent any size and style camera body from pocket compact up to dSLR. Therefore, the view finder hump and even the PASM wheel may or may not be present.

Picture-taking controls

1. Shutter release. Leave the shutter release right where it is, top right. No problem there. But just like the zoom toggle on a point & shoot, circle the shutter release with a lever that toggles between any further shutter-related options, such as 2 second timer, 10 second timer, continuous fire, bracketing, bulb, and intervalometer. Of course the current choice shows in the finder or on the LCD.
2. On-off-complexity button. Pressing this recessed, LED-lit button once turns the camera on and puts it in the complexity level the user last used the camera at. Further presses toggle between three complexity levels: full-auto mode, hand-holdy mode, and full-control mode. The button changes colour or brightness or some such to indicate which mode is in effect. In full-auto none of the exposure or lens control controls have any effect; essentially you press the shutter release and do playback, period. In hand-holding mode some at least of the exposure and lens controls have limited, easy-to-understand effects. In full-control mode the usage button is unlit and returns complete control potential to the user.
3. Lens ring. Put aperture, focus, and focal length controls on the lens. Not due to some mystical form-follows-function mantra; but because many attachable lenses already have such controls on the lens barrel. For attachable lenses this simply means retaining the current design of an aperture ring at the base, followed by a manual focus ring, followed by a zoom ring if needed. For an integrated lens, the lens ring of the Canon S90, S95, S100 series is perfect. In playback mode this wheel is used to magnify pictures. In addition, make the lens ring depressible if possible, otherwise add a button right beside the lens ring. Either way this toggles the lens ring's function from aperture to focus to focal length control. For cameras that offer both auto and manual focus the toggle sequence goes: aperture, AF, MF, focal length. If a lens that already has any of aperture, zoom, etc. rings, the camera-based lens ring only offers those functions the lens does not provide.
4. Right control wheel. Put shutter, ISO, and exposure comp. (if wanted) in a single control wheel on the top right of the camera. This could be directly on top as shown in Fig. 2, or it could be partially embedded as is common with most recent dSLRs. Again: either the wheel is depressible or an adjacent button (labelled 4a in the illustration) toggles between these variables and feedback appears in the finder and/or monitor. There is no particularly compelling relationship between shutter speed, ISO, and exposure comp. It's simply better for the hand to learn the location of one control without looking rather than two or three. In playback mode this wheel is used to scroll between pictures.
5. Exposure method. Also beside the shutter/ISO wheel put a button that toggles between total scene exposure and specific subject (whatever is at the focal point) exposure. (I mentioned this briefly in the previous post and will likely explain it more in a future blog post.)
6. Mode dial. If any major control has to go due to space constraints, ditch the mode dial and put its options on the control panel (see below) or as a button that one presses multiple times to toggle between options. Otherwise, standardize it's location at top left (optionally recessed to prevent accidental position changes). The familiar MASP options are specified, but detents for hand-holding options such as scene modes, picture effects, etc. become user custom setting recalls when the camera is in full user control mode. For this reason I've labelled these extra positions more abstractly as simply 1, 2, 3, 4. Selecting one of the numbered positions causes the appropriate information to be displayed in the finder or LCD, so the user is not confused and doesn't need to memorize anything.

Computer controls

Standardize an arrangement of a row of perhaps four buttons on the back left of the camera beside the LCD monitor (if any) to invoke computer control. One arrangement that works for me is:

 7. State toggle (record->playback->control panel)

 8. Task 1 (e.g. video record->protect image->page 1)

 9. Task 2 (e.g. focus lock->edit->page 2)

10. Task 3 (e.g. exposure lock->add to group->page 3)

11. Task 4 (e.g. WB->trash->page 4)

The top button (7) toggles the function of the lower four buttons by changing the camera's operational state. For example: when the user presses the top button once the camera switches from record to playback, which in turn switches the button 9 from functioning as video record to protect current image. If the user presses button 7 again the camera's operational state switches from playback to the control panel array. Pressing button 9 now causes the control panel display to jump to its second page of options.

The actual assignments of functions to buttons 8, 9, 10, and 11 are completely customizable by the user within the menu system. Therefore the leftmost edge of the LCD displays an icon or abbreviation parallel to each button to display its current assignment in the camera's current state. This puts to rest the universal irritation that the camera you just bought has, for example, a dedicated white balance button, which you never use, but lacks a dedicated mirror lock-up button, which you would use daily.

Recall, however, that the above only describes what buttons 7 through 11 do when the camera is in full-control mode. In full-auto mode these buttons only provide their playback assignments. In hand-holding mode they provide some subset of picture-taking enhancements at the camera manufacturer's discretion. I'd guess centering around scene modes and pictures effects.

Notice the lack of function/Fn buttons, and of a menu button. These are no longer needed. But there are obviously other control needs I haven't assigned dedicated buttons to but which manufacturers may want to add as separate physical controls in certain models, such as the common red video record button, live view, and flash. The more the location and type of these controls are standardized the better.

Control panel



Fig. 3: Control panel example from Canon S100
up/down arrows move left column selection
left/right arrows change selection value

My little S100 pocket camera has a very efficient control panel system, one of several arrangements that could be used. Press the control panel button (labelled FUNC on the S100) and a column of commonly used options appears along the left edge of the LCD, while a row of options for the currently selected column entry appears along the bottom. To the right of the LCD screen there is a four-way arrow controller (labelled 12 in Fig. 2) surrounding a centre set button. This familiar arrangement is used to navigate the control panel but can and should be omitted if the camera has a touch-screen LCD. An intelligently chosen set of options here reduces the need to dive into the more extensive legacy menu system. There doesn't need to be a separate physical button to invoke the legacy menu maze – that should be the last option on the control panel. Replacing the four-way controller with a touch screen reduces the appearance of complexity even further. Further: once the control panel has been invoked, the row of buttons on the left can be used to switch between screen-fulls of options to reduce scrolling.

The key here is standardization. Camera manufacturers need to voluntarily join an industry standards committee to hammer something very like the above proposal than voluntarily implement it on all future models. Over time the system would become less relevant, assuming an endless parade of new features. Fine. Revisit and revise the joint standard when needed. The above proposed arrangement reduces the number of control points without going to a drastic extreme. This in itself makes a camera less intimidating to the novice. Standardizing to some such system further reduces the user's sense of complexity due to familiarization gained over time. Again, notice how much simpler the control layout in Fig. 2 looks compared to Fig. 1. After you've learned it through using any one camera, pick up any other camera with the same layout and you're good to go in seconds.

Am I crazy in the head? No doubt; but write back to tell me in what way, exactly. For my next daring feat I'll explain why you should care about quality even though today's boys and girls just wanna have fun. Stay tuned.

Return to top of page

 

Entry 1, 23 Jan: The OM-D and the concept of retro digital

Olympus used a teaser campaign with great success in 2009 to build interest in its launch of its first micro-four-thirds camera. Now, in the weeks leading up to a trade show starting Feb. 8 Olympus has started a new teaser campaign to build interest in another new product line, the OM-D.



Fig. 1: Olympus OM-1 35mm film camera, circa 1973

This will also be a 4/3rds sensor camera, and will have a body style very similar to Olympus' classic OM series of 35mm film cameras, just as its miro-four-thirds Pen series is styled after its film era Pen models. Those who are old enough to have done photography with pre-electronic 35mm cameras are very nostalgic for their all-metal construction, perfect size, and elegantly simple interface. This era of cameras – which included the Leica and Contax rangefinders, the Nikon F and FM series, the Pentax K1000, the Olympus OM series, etc. – were truly at antipodes to today's digital computers with attached optics. The near-universal design paradigm for a digital camera is to throw in as many features as possible, then let the user customize its function as much as possible to match the user's needs. The resulting complexity, at least during the first few days after purchase, has seemingly come to equal that of a major business software product, such as Microsoft Word.

Rumored-but-likely specs for the OM-D:
– 16 MP 4/3 sensor (variant of the unimpressive Panasonic GH2 sensor?)
– 200 to 25,600 ISO
– OM series retro styling
– 5 axis in-body image stabilization
– Ultra-fast AF
– 1.44 million dot EVF
– Weather-sealed magnesium body with leather covering
– 425 gm weight including battery

With the digital Pen series and now the OM-D Olympus is seeking to capitalize on that nostalgia by imitating the look of by-gone cameras. Fuji is following suit (and Leica goes much further than just appearances with its M8 and M9). The dissonance, however, between what Olympus is doing and the film cameras they are modeling is profound. The OM-D will have an electronic view finder housed in its apparent pentaprism bulge. There will be the inevitable LCD monitor on the back. There is unlikely to be a film advance lever, for what purpose would it serve? Yet the manual film advance was an integral part of the all-manual shooting experience. Instead, there will be the inevitable mountain of features – bracketing, auto exposure, image review, JPEG conversions options, etc. Forum posters are already starting to speculate on the first OM-D model's auto focus performance.



Fig. 2: Olympus E-P1, back view. 17 controls visible.

There just is no sensible way to marry the interface simplicity of an all-manual film camera with a digital sensor. You need a battery to power the computer that drives the sensor. Given you have a battery, may as well throw in an in-camera exposure meter. May as well throw in an illuminated aperture, shutter speed, and frame count display in the finder. No 3 inch LCD on the back for image review? Count your sales on your fingers if you do that. Probably true even if you have an electronic finder and image review can be shown there. Either way you need buttons for at least playback and deletion. And so it goes: What no bracketing? Gasp! No AF? Horror and shock.

Back-to-fundamentals sounds good in theory. In practice you lose sales for each feature you exclude. The systematic corruption of user-expectations dates back to the last three decades of the last century. During those decades we saw what was first a mechanical device, not dissimilar in spirit to a wind-up wrist watch, systematically morph into a high tech jet aircraft plasticky gizmo. The 1990s film camera had auto focus, auto exposure, auto film advance, auto ISO setting, etc. The process was so gradual that few users noticed that each new "Look what I can do!" feature came with a dollop of "Look what else you have to learn how to use".

The solution to this problem is three-fold and very unlikely to be implemented:

1. Standardized interface. Ever drive a car with a joystick instead of a steering wheel? Or with the brake pedal to the left of the gas pedal? If I've driven a Ford Taurus for the past ten years I nevertheless expect to be able to get into a current-model Mazda and know how to start the engine and drive it off the lot. Digital cameras, like electronic film cameras before them (and unlike automobiles) take few lives, so manufacturers have been free to compete on interface layout as well as feature introduction. The result has been chaos. This plays into the manufacturers' hands in that my hard-won familiarity with one manufacturer's interface is non-trivial reason to not switch brands.
2. True AE. Auto exposure, including pattern metering, has been with us for decades. Besides pattern/evaluative metering, we now have the legacy centre-weighted and spot metering modes, plus exposure compensation to make up for the limitations of the technology, plus AE lock, plus exposure bracketing, etc. All this replaces the single cadmium diode in the view finder with an analogue needle display that was first introduced about 40 years ago. Part of the reason photographers were actually able to take pictures with such a "primitive" mechanism back then (and almost inconceivably with no chimping available) was due to the huge dynamic range of neg film, plus its shoulder/toe forgiveness of mistakes – a forgiveness sorely lacking in digital capture.
3. Hideable complexity. A properly designed interface completely morphs from full automatic to semi-automatic to full manual depending on the setting of one standardized control. We are seeing hints of this principle being applied in a rather hit-or-miss fashion. No one has taken it to anywhere near its logical conclusion.

What the combination of the above three principles would look like in practice I'll leave for my next blog entry. Stay tuned.

Return to top of page

* * *

2011 blog entries