For the cacao lover [permalink]

Well, here’s some of the very best chocolate on the market from ex-winemaker John Scharffenberger. Tour the chocolate factory in Emeryville if you get a chance, but make a reservation well in advance, and don’t look askance at any proffered samples.

The “Milk Nibby” is my favorite, and a real diet-breaker as I can personally attest, and it makes a damn good s’more as well. Hearty backpacking food I say. The 82% is for serious cacao fans, and might fairly be called a health food.

Leica service followup-followup [permalink]

Well, I must give Leica an “A” for effort (see yesterday’s entry for background). Yesterday,they pulled my 90/2 APO-Summicron-R ASPH from the pending shipment to Germany, and early this morning I received an email regarding the lens. They had conferred with their staff about the need to send it back, and held the shipment to Germany until this morning, waiting for my answer on what to do with the lens.

I telephoned, and it was explained that to eliminate the accumulated internal dust (one of my requests), the lens would have to be opened on the factory line, which was involved enough due to the lens construction to require special equipment. I was also quoted a worst-case estimate of$200-$300, along with an implied best effort to turn it around in a shorter time frame. That’s hardly cheap, but given the effort and labor involved in working with such a tight-tolerance lens, the cost must be kept in perspective. When restored to factory-fresh specifications, the 90/2 APO ASPH is about as close to the limits of performance as has ever been built for a 90mm f/2. It quite probably was already optically at specifications, but the internal dust was enough that it concerned me for some strongly backlit shooting I have in mind, especially in infrared. That is one weakness of the 90/2: it could be sealed better against the ingress of dust.

So the 90/2 is headed back to Germany, and I look forward to shooting it upon its return. I had previously shot it (see review), and it shows enough potential that it’s worth the effort. For an example image, see Wash and Dry on the Experiments page. What I don’t know for sure is whether the performance will be maintained if converted to Nikon mount (so that I could exploit it on both Nikon and Canon).

Why so much effort for a lens that requires an adapter to shoot on Canon EOS? First, the 21MP Canon EOS 1Ds Mark III demands such lenses to exploit its full-frame sensor, especially if one plans to shoot in the f/2 - f/4 range, as I do. Second, the 90/2 offers some unique qualities in the way it renders out of focus areas, particularly specular highlights and bright lights. Third, it should offer high performance in infrared. Fourth, “they” just don’t build ’em like this anymore. With the exception of the Coastal Optics 60mm f/4 APO macro, lenses that perform to this level are few and far between. Nikon and Canon don’t offer a single apochromatic lens, though some designs are well-corrected for color.

Without A/B testing, I cannot say that the Zeiss ZF 100/2 Makro-Planar (or even 85/1.4 Planar) might not offer very similar performance, speaking strictly in terms of resolution and contrast; I already know that I prefer the color balance of the Zeiss ZF 85/1.4 Planar. However, the Leica 90/2 APO is apochromatic, and so out-of-focus areas render in some very interesting ways by comparison, a trait I noticed when shooting at night. Lenses are artistic tools, and I’ve grown to appreciate the nuances of individual lenses. Never assume your preferences are those of others, though.

The future of resolution, 36 megapixels? [permalink]

In the “long” term (the next 2-3 years!), 30-40 megapixel DSLRs will emerge, a completely pointless development unless pixel quality improves substantially and optical performance rises to the challenge. Even today, the 21MP Canon EOS 1Ds Mark III demands exceptional optics to exploit the sensor, especially near the edges and in the corners. Worse, with those same exceptional optics, diffraction begins to degrade image contrast by f/8, so “wiggle room” is not great.

Resolution is certainly fun. This 21-megapixel image was taken with the Coastal Optics 60/4 APO macro on the Canon EOS 1Ds Mark III. Loads of detail. Could the Coastal Optics 60/4 deliver even more resolution at good contrast on a 36 megapixel sensor? It seems likely.

At 36MP, an aperture no smaller than f/5.6 will be the limit at which image quality is reigned in by diffraction—which is not to say that perfectly wonderful images won’t be made down to f/11 or even f/16—it’s just that optimal contrast will be reached no later than f/5.6 with a top-grade lens on a 30+ megapixel camera.

So my belief is that as we move to the world of 30-40 megapixel full-frame DSLRs, only the world-class lenses will prove satisfying in the restricted f/2 - f/5.6 range. For that reason, it’s a strategic move to acquire world-class optics that might or might not be available in the future (and buying used at a steep discount is fine when care is taken). Already, alleged gems like the Cosina Voigtlander 125/2.4 APO can no longer be had. Who’s to say whether Leica will continue to make their 90/2, 180/2.8 and 280/4 APO models? Or that Coastal Optics will make their 60/4 APO macro for years on end? Such lenses are a long-term investment, not to be purchased on a whim, but not to be taken for granted, either.

Leica service followup [permalink]

Following up on yesterday’s entry— first, please note that only certain special lenses like the 90/2 APO APO-Summicron-R require service in Germany. The Leica 90/2 is unquestionably one of the pinnacles of modern lens design, and requires exceptionally high-precision equipment to build and service, equipment that apparently resides only in Germany (and quite probably the expertise to use that equipment also). I’ve added a small note to my review, btw.

By comparison, you can buy a brand-new Nikon or Canon lens, and take your chances. My turnaround time from Nikon/Canon (known from numerous lens experiences) is generally 2-4 weeks; they have USA-based service centers. However, I once sent a brand-new Nikon 360/f9 view camera lens in for service that required 3-4 months to service in Japan.

Yesterday’s entry received a response from Leica about being “very very disappointed that you went to the blog without consulting us”, perhaps because I was informed before I sent the lens in that it “might have to go to Germany” to be serviced. In addition, Leica did not promise any specific turnaround time. Leica was honest, and I don’t fault them for not quoting an up-front turnaround time. Furthermore, I spoke with Leica today, and they had already pulled my 90/2 from the outbound items for Germany (to re-check it), to double-check whether it really needed to go to Germany (my service request was for it to be performing at 100% of specification, and to remove some internal dust). That’s responsive service, and speaks well for Leica.

Discussing this with Leica’s VP of Marketing, there is approximately 2-3 weeks of transit time to Germany from the USA (including customs), and perhaps 7-10 days of processing time in the USA. Many products can be serviced in US, but I learned that apochromatic (APO) lenses have to actually go back onto the production line due to their exceptionally precise tolerances, where even a tiny misalignment can defeat the exquisite optical calculations. So all of this is understandable, but the fact remains that even one month is a long time, and 3-4 months is very long indeed, though the 3-4 month figure might be a case of “underpromise and overdeliver”.

As a fixed principle, I do not and will not accept special treatment in exchange for not writing about something. A corollary is that I never “preflight” editorial comments, reviews, etc with any manufacturer. I publish first, and if a manufacturer finds any factual errors or relevant “flip side”, then I address those appropriately.

I do hope that Leica expedites the handling of my 90/2, but it won’t change what I’ve written here. I’ll report on the outcome—while hoping that Leica addresses the turnaround-time issue.

A product consists of both the physical object and the service and support. I don’t think any customer would consider a 3-4 month turnaround fair or reasonable. Even one month is 3 weeks too long for a working professional.

It’s all too easy for companies to react negatively (or ignore) constructive criticism, instead of understanding that they lose sales by not understanding their customers, that they worsen the situation by not tackling it head-on, and that they would be far better off sending profuse thanks to anyone willing to provide such invaluable customer feedback. The bottom line for success: apply Miller’s Law to any customer feedback:

To understand what another person is saying, you must assume that it is true and try to imagine what it could be true of.

Try it sometime—it’s very easy to not apply (I’m good at that), but when applied rigorously, the results can be astonishing. What often happens is the converse, Lloyd’s Law:

To misunderstand what another person is saying, assume that it is wrong, and try to reject it for any reason whatsoever.

We should all hope that Leica takes the foregoing to heart, because the presence of multiple viable lens/camera manufacturers benefits everyone regardless of brand. Leica is a venerable company with a long history, and with world-class offerings simply not available elsewhere. Leica has recently experienced ownership turmoil (read the news). With luck, that will sort itself out with a newer and stronger and more realistic Leica emerging.

My short-term suggestion for Leica: fix the turnaround time issue. It’s not my concern that the Leica experts and their advanced equipment reside in Germany; that’s Leica’s logistical problem to sort out. Leica could expedite shipping, provide a loaner, etc; there are all sorts of tactical options to mitigate the impact, along with a longer-term strategic plan to fix the problem.

My medium-term suggestion for Leica: expand your market by doing what Zeiss is doing with the ZF line: make the world-class “R” lenses available in (at least) Nikon mount. There is simply no company today that addresses the high-end “no compromises” lens market. (The Coastal 60/4 is a “no compromises” lens, but it’s a singleton offering).

My long-term suggestion for Leica: the Leica R10 digital needs to emerge; the M8 rangefinder is definitely not for everyone. An R10 digital should be 16-20 very high quality megapixels, and there should be some new AF lenses to go with it. A Leica M9 rangefinder, with full-frame and Live View is also sorely needed.

Leica service: 3-4 months [permalink]

I recently acquired a used Leica 90/2 APO-Summicron-R at a very nice price (see my review). The lens seemed to be in good working order, but for such an exotic lens, I wanted to make sure it was operating at 100% of specifications, so I sent it in for a checkup. About 10 days later I received a notice in the mail that it’s off to Germany with an expected 3-4 month turnaround time.

How can any company survive making its customers wait that long, especially when they sell products for 3-5X the cost of the Japanese brands? Leica makes stunning world-class lenses, but a product is both the physical object and the support behind it. For Leica to succeed in the new world of digital photography, customer support must be much more responsive.

Canon shooters can shoot many (but not all) Leica R lenses on their EOS 1D/5D bodies using an adapter; that is my usage, though I am considering having the 90/2 APO permanently modified to Nikon mount by S.K. Grimes.

Canon EOS 1DsM3 — detail galore [permalink]

To balance Wednesday’s entry on the noise and color of the Canon EOS 1Ds Mark III, what excites many people are those 21 megapixels, which offer a fantastic amount of detail, provided that top-performing optics are used, such as the Zeiss ZF 50/1.4 Planar used in the image below (or even better, the Coastal Optics 60/4 APO macro).

I predict that most of the Zeiss ZF line will “deliver the goods” to around the 36 megapixel mark, so I consider them a worthwhile investment for today and tomorrow. A 36-megapixel camera would offer an image of about 7350 X 4900 pixels using 4.9-micron pixels. 36 megapixels is not so far-fetched with Sony’s 24MP sensor on the horizon (D3X?). Per-pixel quality is not likely to impress at such high pixel densities, but the resolution will have its applications. Keep in mind that such a camera will be diffraction limited around f/4 or so, so forget about “f/8 and be there”. Even today at 21MP, it can readily be seen that the Coastal Optics 60/4 APO macro peaks at f/5.6 due to diffraction; a 36MP camera will drop that another stop or so, depending on the sensor design.

The EOS 1DsM3 offers unprecedented image detail from a DSLR, and only a drum scanner with the finest-grain slide film could hope to offer more usable detail. The Nikon D3 simply cannot approach this level of detail, so if resolution is what floats your boat, then the EOS 1DsM3 is your camera. But the overall visual impact (color, contrast and lifelike rendition) simply does not match that of the Nikon D3. Still, there’s no denying the vicarious pleasure of seeing more after taking the picture than being there in person. Call it the “handheld surveillance camera”.

Black and white shooters won’t find color or noise to be troubling at ISO 1600, in fact the results are very nice indeed. These images were converted straight from Canon’s Digital Photo Professional, using the “Monochrome” Picture Style. For this style of shooting, I love how the Zeiss ZF Lenses render the out of focus areas. The 35/2 Distagon offers particularly lovely results.

Target audience [permalink]

I am struck by the irony here—the hard-working people (and bums) who ride the San Francisco Muni bus aren’t likely to be able to afford the lifestyle this woman exudes, let alone health care, a knee replacement, or a bevy of physical therapists (click image to read the text), so who exactly is this ad targeted at?

It’s gotta hurt to be reminded about what you can’t have when you’re worried about your next paycheck, though after your bus ride home you can connect your iPod to your new iMac to recharge while you check out the web site in your copious spare time while enjoying take-out from the city’s finest. Never fear, California Pacific Medical Center is “With you. For life.”—cash upfront please, and please add 5% for self-serving public relations overhead.

Handheld shooting—can it be sharp at 1/4 sec?! [permalink]

I’ve found from experience that handheld shooting with large and heavy cameras can yield sharp images down to ridiculously low shutter speeds, provided that one accepts that 2-4 frames might be needed to yield one frame that is acceptably crisp.

By “large and heavy” I mean cameras like the Nikon D3 or Canon’s 1Ds Mark III. I dislike cameras like the Nikon D200 for low light shooting; less mass (and perhaps a less well damped shutter) means a lower success rate under marginal conditions, or so I’ve concluded from years of experience, at least with Nikon DSLRs. In addition, the handheld technique I’ve developed relies on stabilizing the larger camera bodies against my face, something not really feasible and/or quite awkward with the smaller camera bodies.

I wonder...the amount of vibration I can feel (through my hand) with the Nikon D3 is notably less than with the Canon EOS 1Ds Mark III. I haven’t determined if this actually results in a higher percentage of sharp images in the 1/4 - 1/60 range— a good topic to add to The Sharpest Image at some point.

Today’s example: I took two frames, both at 1/4 second (yes, one quarter second!) at ~/f8, handheld at ISO 1600 with no support (just free-standing). The successful frame is not critically sharp, but it’s awfully good for 1/4 sec handheld, especially keeping in mind that the 1DsM3 is a 21MP camera; the image would stand up to fairly high enlargement (though the chroma noise of the Canon EOS 1DsM3 is really very disappointing compared to what the Nikon D3 can deliver).

The point here is that sharp images can be made at very low shutter speeds—this is not a one-time freak accident. By shooting 2-4 frames, chances are good that one of them will be sufficiently crisp for even fairly large prints—I do this all the time, though more commonly in the 1/15 - 1/60 range.

Canon EOS 1Ds Mark III — low light shooting [permalink]

My Nikon D3 is in for service, so I used my 2nd choice today: the 21MP Canon EOS 1Ds Mark III. First, let’s keep things in perspective: the images that the 21MP Canon EOS 1Ds Mark III makes at ISO 1600 are so much better than film ever managed—we’ve come a long way indeed.

But in reviewing the images I shot today in the interior of a victorian-style house in San Francisco, I was struck by the marked inferiority of the 1DsM3’s images to those I’m used to seeing from the Nikon D3 (the “poor man’s medium format back”). The amount of chroma (color) noise and its subjective appearance really grabs my attention compared to the Nikon D3 results, not to mention the muddy tones and color of the 1DsM3. Even the midtones suffer from ugly chroma noise—yuck! This kind of stuff shows up in the clarity and depth of the overall image, making it look muddy; it’s not about the actual pixels view but about the overall visual impact.

You can read all the “objective” reviews you like with their comparison patches and the dithering claims of “a stop or so” advantage with the Nikon D3. Yet my emotional response to the 1DsM3 images is one of disappointment; the images lack the life and realism of the D3 images, and the chroma noise jumps out in even the medium tones. We all see differently, so your conclusions could differ, but that’s my take on it and I have no axe to grind here since I own both brands (and shoot the same Zeiss ZF Lenses on both). To be fair, the 21MP images require less enlargement than the Nikon D3 images (for the same size print), but I don’t think that overcomes the nuances, and the hot pixel is unacceptable for a US $8000 camera.

Canon EOS 1Ds Mark III — another “brain fart” [permalink]

I previously reported (Feb 20) on sporadic exposure problems with the Canon EOS 1DM3 and 1DsM3. Well it’s happened again. This image below is another example from the 21MP Canon EOS 1DsM3. OK, so I had dialed in +1 stop, but Evaluative Metering still was off by about 3 stops.

No nothin’ [permalink]

Of course the beach would be trashed quickly, but wouldn’t it be fun to drive onto the beach, pitch a tent, light a fire and enjoy some brewskies? Or maybe just s‘mores if it’s with kids. In the “old days” people used to be able to go to a park like Yosemite, drive into a meadow and have at it. The world has changed...

Nikon 70-200/2.8 VR on Nikon D3 [permalink]

I discovered that the Nikon 70-200 VR offers awful performance on the full-frame Nikon D3 when focused near infinity towards the long end of the zoom range (200mm). I reported this to Bjørn Rørslett (naturfotograf.com), who subsequently confirmed my findings.

The problem is so severe that stopping down to f/8 or f/11 (or even f/16) is insufficient to overcome it; it appears that the 70-200VR simply cannot “cover” the 36 X 24mm frame adequately. Its optical performance is without a doubt the very worst I’ve yet to see on the D3 from any lens, Nikon or otherwise. And because I see symmetric results over the frame, optical alignment issues cannot be blamed. Rather, it is fair to conclude that it is under-engineered for a full-frame camera, or at least the D3’s sensor. Truly, this is a pathetic result for a pro-level zoom.

The foregoing applies to the long end (200mm) when focused near infinity. At other settings, the 70-200VR is very sharp over most of the frame (eg at closer distances and/or shorter focal lengths), and is otherwise a very fine lens. It remains an excellent choice for DX cameras, but this optical defect leaves a huge “hole” in the lens lineup for Nikon D3 shooters.

For rapid action and certain other uses, a 70-200 autofocus zoom is the only viable option. However, for other purposes, the Zeiss ZF 85/1.4 Planar and 100/2 Makro-Planar not only offer across-the-frame sharpness wide-open, but superior contrast and color as well. Why not read Zeiss ZF Lenses and consider exploiting your D3’s sensor fully? Regrettably, the ZF line currently includes nothing longer than 100mm. Get Zeiss ZF lenses at B&H Photo.

Nikon D3 off to Nikon service [permalink]

Having continuing problems with eye focus (see Feb 17 entry), I’ve sent my D3 in to be checked out by Nikon. I could not obtain accurately focused images with a variety of lenses using manual eye focus—even with the D3’s “focus assist” in agreement with my eye. Autofocus has been accurate however.

With previous Nikon DSLRs, such a failing would have been a show-stopper, but the new Live View feature offers a superior alternative for tripod use. However, I shoot handheld with manual focus a great deal (eg Zeiss ZF Lenses), so the D3 had to go in to get this corrected. Now I hope Nikon fixes the problem, not a “given” of course.

Fiat money, inflation and “liquidity” [permalink]

This blog entry has nothing to do with photography...or does it? Why is Canon increasing US prices? (Leica has already done so, numerous others can’t be far behind). Are the higher prices due to “inflation*”, a mysterious force that even the Federal Reserve struggles to comprehend?

Today’s brutal financial environment makes The Creature From Jekyll Island—A Second Look at the Federal Reserve more relevant than ever. If you think “inflation” is some mysterious force of nature or simply “higher prices”, then cure that profound ignorance today*—learn how your money is steadily embezzled. It is not an accident that Creature is not mandatory reading for college students today (I wish it had been for me). And it’s just as relevant for those using a currency other than the US dollar.

* Inflation is an expansion of the money supply, often leading to higher prices (the law of supply and demand also moves prices). An increase in the money supply beyond the real wealth being created results in more paper money chasing the same goods, hence higher prices eg a currency worth less and less in terms of gold, oil, camera equipment. Only the government has the power to inflate (increase) the money supply, eg running the real or electronic currency printing presses.

Canon price increasing coming May 1? [permalink]

I received an “early bird” notice today via email from my local camera store Keeble & Shucat (below). If you’re in the San Francisco Bay Area, K&S is the premier store, with sales, rentals and photo processing. Only a handful of camera stores carry the range of products they do. (Bear Images is also a good choice, and it’s only half a mile from K&S).

Welcome to your early-bird notice for the upcoming Canon Days Sale! If you are a Canon user...you will not want to miss out on this sale and the events surrounding it. Our KSP newspaper ads will hit all the bay area papers on Tuesday. If you prefer to beat the rush, starting today KSP will honor the sale prices. KSP will be paying the California sales tax on Canon cameras, lenses and Canon printers that are inventory. There is one exception to the sale KSP will not be paying the sales tax on the Canon 1DsMarkIII body camera body. We have received notice from Canon that there will be a price increase across the board on all of their products on May 1st. This will be a great time for you to take advantage of the tax savings before the price increase goes into effect.

Note the part about a price increase coming on May 1 from Canon (emphasis added). It will be surprising if Nikon, Zeiss and other foreign companies don’t soon follow suit.

Asha Holi [permalink]

I stayed upwind of this celebration (April 6) at Stanford, put on the by the Asha Stanford group, not wanting the dust in my eyes, lungs or lens! Click for more.

Digital infrared video (mountain biking) [permalink]

Digital video from a point & shoot can be very interesting in infrared. I hung my converted Canon G9-IR around my neck and let it dangle—obviously the mounting system needs some improvement, but the proof of concept is clear. If you know of a system to mount a small point & shoot camera on a bike helmet, please let me know—I have some cool videos in mind.

Anyway, I want to show the video on YouTube, but all I get is a generic “Failure” after it uploads and processes it (I tried twice). As they say in internet parlance, WTF?

Samsung 1 terabyte hard drives—not recommended [permalink]

I recently returned some Samsung 1TB (1000 gigabyte) hard drives (specifically “SAMSUNG Spinpoint F1 HD103UJ 1TB SATA 7200 RPM 32MB Buffer Hard Drive”). Out of 4 drives, one was consistently 9% slower than the others, dropping striped RAID speed by much more than 9%; mismatched drives are the kiss of death to performance for a striped RAID.

More importantly, performance was highly variable with the Samsungs on both the internal Mac Pro SATA ports (original 3.0GHz Mac Pro) and external ports on both Sonnet E4P and FirmTek eSATA cards. I exchanged the drives for the consistent and high-performing Seagate 1TB drives (see Feb 20 blog entry). Your results might vary on other computers and/or with other eSATA cards, but the Seagate ST31000340NS drives have performed flawlessly for me. See also Hard Drives. In short, I strongly advise against choosing the Samsung drives; I wasted hours getting the damn things to not perform, and “eating” a 7.5% return fee too.

For this computer geek, 1 terabyte (TB) = 1024 gigabytes = 1024 X 1024 X 1024 kilobytes. But hard drive vendors consider a terabyte 1000 X 1000 X 1000 bytes, which means that on Mac OS X, a 1TB drive formats to about 931 base-2 gigabytes; Mac OS X uses base-2 math (powers of 2) so that a kilobyte is 1024 bytes, a megabyte 1048576 bytes, etc. Don’t worry about it, but don’t be surprised either. Even accounting for the base-2 vs base-10 difference (only 2.4%), a one-terabyte drive still formats to only about 953GB on Mac OS X, a discrepency of 5%. A drive cannot be used without being formatted, but doesn’t 1TB (1000GB) sound more appealing than 953GB? That’s the explanation, really.

Shown below is my TimeMachine backup disk, a 2-drive striped RAID. A single drive would be better for reliability, but unfortunately my data now just (barely) exceeds 1TB.

Canon G9 infrared [permalink]

The Canon PowerShot G9 is a nice little (brick) of a camera for infrared, converted for me by maxmax.com. However, one of the G9’s irritations is that Canon’s Digital Photo Professional won’t operate on the CR2 (RAW) files produced by the G9, and Adobe Camera Raw seems to be incapable of using the “as shot” white balance properly. So that leaves the nuisance problem of having to use Canon’s consumer-grade (and incredibly sluggish) ImageBrowser software, which does a decent job, but not a very good job at reducing digital noise while maintaining detail.

The self-defeating decision to offer 12 megapixels in a point and shoot tiny-sensor camera is not limited to Canon, and those pixels should be discounted by at least 60% when compared to my Canon EOS 5D-IR, but that’s par for the course with all point and shoot digital cameras, including the petite Fuji F30, which doesn’t offer the flexibility of the G9, nor RAW file support. It’s a darn shame the G9 isn’t offered with the same size sensor, but in a 6MP version. Very little would be lost, and a great deal gained in overall image quality. What’s next, 24MP pinky-sized sensors? Yuck...

Windy Hill (below) is a landmark in this area due to its grassy expanse devoid of trees, with the hills to the north and south being tree-covered. Please contact me if you know the geologic or environmental conditions that cause the lack of trees, and I’ll happily share that with my readers.

The false-color look is overused with infrared, but in this case I prefer it over black and white with quick-and-dirty processing, though black and white might offer a superior result with additional effort—not worth it for this documentary photo.

This is an ideal time of year to photograph landscapes in infrared in California—the trees are leafing out and the new foliage positively glows in infrared. Too bad I’m cooped up in the office every day...

The choice of image rendition is yours—learn about all the possibilities in the diglloyd Guide to Digital Infrared Photography. Scripts are included that generate image variants, useful for ideas for further exploration.

Gaze [permalink]

What lies therein? A horse in a stall has plenty of time to think.

Sheer magic [permalink]

I commented yesterday on the Canon EOS 1Ds Mark III, and after shooting the Nikon D3 later that day after the 1DsM3, my perceptions have only been reinforced further. The D3 offers, by far, the best image quality ever seen in a DSLR. Its images have a dimensionality, smoothness and transparency that simply isn’t there with the 21MP Canon EOS 1DsM3 (though Canon’s 10.1MP EOS 1DM3 is a lot closer, a camera I recommend for those sticking with Canon).

Like magic, the D3 also overlooks the faults of mediocre optics and makes them worth shooting again; this is not true of the EOS line. Go figure—maybe it’s only “big fat pixels”, or maybe it’s Nikon electronic magic or a “killer” sensor, but it is the case that many of Nikon’s older AI-S lenses perform visibly better on the D3 than on any previous Nikon DSLR. Such lenses include the 24/2 AI-S, the 28/2 AI-S, the 35/1.4 AI-S, and a variety of others. Keep that in mind if you want to build a lens collection—lots of used equipment can be had for very reasonable prices. Of course, my preference remains with the Zeiss ZF Lenses, for a variety of reasons, but the performance gap narrows with the D3 magic.

If you’re “on the fence” about the Nikon D3 vs the Canon EOS 1Ds mark III, absolutely and unequivocally get the D3 unless you really need megapixels (detail). The Nikon D3 is The Poor Man’s Medium Format Digital Back (see my reviews). You heard it here first. What remains to be seen of course is whether Nikon can maintain D3 quality in a 20+ megapixel DSLR. In the meantime, there are not many subjects that demand more than 12 high-quality megapixels, so consider your needs carefully and take the extra $3K and put it towards Zeiss ZF Lenses or the Coastal Optics 60/4 APO macro. Get Zeiss ZF lenses at B&H Photo.

Zeiss ZF Lenses [permalink]

Zeiss ZF Lenses is well worth reading not just as lens review, but for understanding lenses and lens testing and image quality in general, as the comments from readers attest. Why not order it today?

Experiments [permalink]

I’ve added a new experimental image to my Experiments page. I like the color duality of this image, and I used a few Photoshop layers to make something interesting.

Canon EOS 1Ds Mark III comments [permalink]

The day’s shooting at ISO 200 and 400 reminds me yet again that the chroma noise and murky color of the Canon EOS 1Ds Mark III is mediocre by comparison with the Nikon D3. More and more I’m concluding that the 1DsM3 is a overpriced dud whose only worthwhile attributes are its high resolution and dynamic range—but a dynamic range also impaired by shadow noise. I’m even beginning to be a bit pissed off at spending US$8K on what is clearly not Canon’s best effort. To be sure, the 1DsM3 is a really good camera, but not an outstanding one (in terms of pixel quality), and at US$8K it’s simply overpriced by $3K for what it delivers. See also my March 13 comments.

What is it?

A simple device that is not so simple.

Reviews you can learn from [permalink]

The May 2008 Shutterbug has a “review” of the Canon EF 14mm f/2.8L II, consisting of the usual warmed-over rehash of marketing materials, along with a few insightful generalities like “delivered crisp images with good color”. Uh huh, that’s got me sold on this US$2000 lens. Virtually every lens available today can meet that qualification! The reviewer is without a doubt quite ignorant about the actual performance characteristics of the 14/2.8L. Does it perturb anyone but me that there is almost no reliable and objective information in today’s photographic magazines? The one magazine I really do recommend is Photo Techniques.

As far as how the Canon EF 14/2.8L II does perform, I document those characteristics in my review, which is a screaming bargain if you’re considering the 14/2.8L II, since you’ll not only learn whether the 14/2.8L II is appropriate for you, but you’ll also understand its peculiarities, which can otherwise take time to understand.

A friend loaned me the Canon EF 14/2.8L II for my review for two months. He has decided that it’s one too many lenses to carry, and it’s now for sale—near mint condition. Contact me if interested.

Back to the Shutterbug review—among other gems, the reviewer bemoans the fact that “a circular polarizer can’t be accommodated”. This should be your tip-off that the “review” is another rehash of marketing materials, and that the reviewer’s photographic expertise is dubious, to put it kindly.

Maximum polarization is at 90° from the sun; away from that angle polarization drops off quickly, which leads to gradient of brightness from no change at 0° to maximal change at 90°. This is fine with lenses of 35mm focal length or so and longer, and down to about 24mm or so with careful framing, but even with a 25mm lens (see example), polarization can be “iffy” for this reason, and by 20mm it’s really inadvisable with most subjects. The Canon 14/2.8L II covers a vast 114° degrees, making its angle of view incompatible with polarization under most circumstances.

Coastal Optics 60mm f/4 UV-VIS-IR APO Macro [permalink]

I’ve added a lot of material to my review of the fantastic Coastal Optics 60/4 UV-VIS-IR APO macro, including more examples, spectral transmission charts, comments from the lens designer and more.

Hyperspectral imaging [permalink]

Just for fun: hyperspectral imaging. See also the March 31 entry.

Infrared camera conversions [permalink]

Have some caution when converting a camera for infrared use. First, read the diglloyd Guide to Digital Infrared Photographybefore you pick a camera. Then pick your vendor for a camera conversion (if conversion is the route you choose). I’ve used maxmax.com and highly recommend them; I’ve had half a dozen various cameras converted there, including two DSLRs and 4 point and shoots. Site owner Dan really knows his stuff.

Dan recently sent me examples (pictures) of two conversions he redid for customers dissatisfied with conversions from unspecified “other vendors”.

First example: the exquisite precision of using a belt sander to adjust the focus; so much for plane-parallel sensor alignment (critical for digital sensors). The job done by  “major west coast player” according to Dan.

The second example is a plastic filter, showing an obvious lack of flatness (waviness). According to Dan: “This is from an East Coast vendor. The customer said he was told to shoot at F10 or higher (to compensate for the wrong refractive index).”

While my Nikon D70 was converted using a plastic filter, the one in my D70 is “hardened” and doesn’t show any waviness. Also the vendor took months to do and redo the conversion to solve a filthy sensor problem. But it was a painful experience.

It’s important to not tar all vendors with the same brush. But equally important is realizing that quality can vary, so satisfy yourself before you have a camera converted.