This post has been temporarily due to contradictory results. Once I get this sorted out with more careful testing, I will repost the findings. My apologies for any confusion.

After about a year of trying, I have failed to get good looking black-and-white images from my digital workflow. With that in mind, I recently bought my first 35mm film camera in years, a Canonet GIII QL17 rangefinder. Unfortunately, that camera had a stuck shutter release and stripped light seals, so I bought another one. The second was also plagued by issues, this time related to the pressure plate scratching film. Yet, the few rolls of film I put through these cameras were enough to convince me that even C41 B&W film (Ilford XP2) developed and scanned at Costco had more of the look I have been pursuing than I can get out of my Canon 5D and Alien Skin Exposure. Here's a shot of my son Oliver off the first roll. He was so excited to meet Santa while my wife was shopping for a purse =).


The issues finding a good Canonet prompted me to look elsewhere. I considered buying a Bessa R3A with a Nokton 40/1.4, but that's just a little more than I'd like to spend on a film camera right now. Although I enjoyed the rangefinder experience with the Canonets, the most cost-effective way for me to go with film right now was the SLR route. I already own a few nice lenses for the Olympus OM system (purchased for use via adapter on the Canon 5D), so I went ahead and purchased the Olympus OM 2n film SLR camera.


My new camera should be here soon. Hopefully I will have better luck with this purchase!

Back when I had a Canon 30D, I was struggling to find the right normal lens. The optically-excellent Sigma 30 was not autofocusing reliably, and the Canon 28/1.8 paled optically by comparison. I asked Joe Mama, whom I knew from the DPR forums, whether it would be better in his opinion to buy a Canon 35mm f/1.4L lens or spend a bit more to replace the 30D with a 5D and 50mm f/1.4. The 50mm f/1.4 lens costs roughly $800 less than the 35L and gives similar field-of-view coverage on a 5D to what the 35L provides on a 30D. I was thinking that $800 would go a long way towards making up the price difference between a 30D and a 5D.

Joe, having gone from a 20D and 35L solution to a 5D and 50/1.4 combination, recommended the latter without hesitation. I followed his advice, and right away it felt as though I had made the right choice. Every lens I owned at the time seemed to get significantly sharper when mounted on the 5D rather than the 30D. Thus, when Dave Chow (user name "chowy" in the DPR forums) asked nearly the same question on the forums, I was there with Joe to give him similar advice about selling his 30D to buy a 5D. Like Joe and me, Dave was very happy with the performance of the 5D once he had made the move. Unlike me, Dave had much of the best glass Canon had to offer prior to changing bodies. With the 5D and 30D in hand, he presented in the forums the following series of comparisons. All images shown below were taken by Dave Chow. Thanks again to Dave for letting me share these here.

First up: Canon 5D and 50mm f/1.4 versus Canon 30D and 35mm f/1.4L. These two systems provide a similar field of view, with the 5D system being slightly wider. The 30D system provides about the same depth of field with the 35L at f/1.4 as the 5D does with the 50 at f/2.2. Here is the test scene with the two systems at those settings (Click image for intended viewing size):


I can't tell much difference between them. The 35L bokeh is a touch more appealing (less of a hard edge), but that's a subtle distinction. Let's pixel peep. Here they are at 100% (Click image for intended viewing size):


The 35L is damn impressive wide open, isn't it? Still, I believe the edge in resolution goes to the 5D system here, particularly when the two images are compared at the same output size (30D image upres'd or 5D image downres'd). If right about now you are thinking that I should be comparing these two camera systems at the same f-number rather than at the same DOF, or the same focal length rather than the same field of view, then you are bringing up a topic that I have chosen not to address in this article. Joe Mama has prepared what I consider to be the definitive essay on those issues.

Next pair: Canon 5D and 135L versus Canon 30D and 85/1.2L. Once again, these two systems provide a similar field of view. The 30D system provides about the same depth of field with the 85L at f/1.2 as the 5D does with the 135L at f/2. Here is the background for the test scene:


Here they are with the subject close and the background nearly completely blurred (Click image for intended viewing size):


In this comparison, the 30D system has rendered a slightly more pleasing background blur. Dave shared a second comparison set for these two systems in the original thread, but I am omitting that one from this article because I believe the outcome with regards to bokeh was affected by backfocus on the part of the 30D system. Although I cannot provide the 100% crops for these images, a number of reviews have established that the 135L is sharper at f/2 than is the 85L (either version) at f/1.2.

Finally, we have the standard zoom shootout: Canon 5D and 24-105/4L IS versus Canon 30D and 17-55/2.8 IS. Once again, these two systems provide a similar field of view with the 5D system covering a somewhat wider as well as moderately longer field of view. In this particular comparison, the 30D system was used at 55mm with the 5D system at roughly 90mm to provide a similar field of view. Each lens was shot at maximum aperture, which gives the 5D system a slightly more shallow DOF. Click the image for the intended viewing size.

In this case, the 5D image has more bite, producing nicer colors with greater contrast, though Dave pointed out that the 30D had +1/3 stop higher exposure compared with the 5D system. The 24-105 has also done a slightly better job here with the bokeh, though the slightly different DOF is a confounder. At any rate. neither of these lenses is going to win any awards for bokeh.

Here's what I learned from these comparisons:
1) The 5D system and 30D system in these comparisons produce very similar images, and the cost of the two systems is comparable. The 5D, 50/1.4, 135L, and 24-105L would cost a combined $4000 on Amazon right now. The 30D, 35L, 85L, and 17-55 would set you back $4600.
2) Based on these limited data, the 5D&50/1.4 and 5D&135L combos may do a slightly less good job handling out-of-focus regions when compared with the 1.6x crop system equivalents.
3) Lens testing is a bitch. Okay, so I already knew this, but it bears repeating.

Other points to consider:
1) Current crop offerings do a lot that the 5D can't do.
2) The 5D is probably due for an update soon. I expect that the 5D replacement + 50/1.4 + 135L will likely cost about the same as a 40D + 35L + 85 L though.
3) The 5D combos can go shallower (and faster) than the crop equivalents. Not shown above, for example, was the 5D&50/1.4 at f/1.4, because the 30D and 35L can't produce an equivalent photo (again, see Joe's essay for discussion).
4) The 50/1.2L yields nicer bokeh than the 50/1.4, so if the slight bokeh edge to the 30D/35L combo in the above comparison was troubling, the 50L is always an option for the 5D user.
5) The 5D combos, in general, capture slightly more detail than 1.6x crop systems with equivalent settings.
6) Petteri's site has a nice wide zoom shootout between Canon full frame and crop systems.

It has been widely held that the Four Thirds system offers better corner sharpness than does traditional 35mm photography, so as a user of both systems I thought I'd do a simple test.

Systems compared:
1) Canon 5D (35mm sensor) with EF 28mm f/2.8 prime. This is a cheap lens with sharper corners at 28mm than the more expensive Canon 28mm f/1.8 copy I previously owned. Obviously this lens does not represent the ultimate in corner sharpness available for the 5D (that would require some very expensive glass), but it competes well with the much more expensive L zooms at this focal length (confirmed by SLRGear testing). Look here for a comparison of 13 lenses at this focal length tested on a 5D. You can see that the 28/2.8 is no slouch.
2) Olympus E-410 with 14-42mm zoom (kit lens)

In terms of lens and body, I was somewhat unfair to the Four Thirds system here. The 14-42 is outperformed by the higher cost Olympus zooms (also confirmed by SLRGear testing), and the E-410 has less per-pixel sharpness than the Panasonic DMC-L10 or Olympus E-3.

The 5D was at ISO 800 with the 28mm lens stopped down to f/11
The E-410 was at ISO 200 with the kit lens at 14mm and f/5.6
This was done so that the two systems would produce images with the same depth of field, diagonal angle of view, perspective, and apparent exposure given the same shutter speed and print size.

Shutter speed with both systems was 1/250s. Unfortunately I left my tripod at my parents' home, so both systems were shot handheld. I picked the best of four shots in each case. In each case, all four shots came out the same. I can reliably handhold a 28mm lens at 1/250s.

For processing, I converted from proprietary RAW to DNG in Lightroom and then processed in C1 4 beta using default sharpening (same for both cameras) and zero noise reduction.

More points to consider: 1) To compare these images, they should be examined at the same print size. I did not upres the E-410 photo or downres the 5D photo because the method used to do either could potentially introduce bias. Of course you can do this on your own. I printed them for comparison. 2) C1 doesn't have a specific color profile for the E-410. Color differences here are not representative of these camera models IMO. 3) These two lenses do not represent all lenses available for their given systems (duh). Again, this is a cheap prime designed decades ago versus a cheap zoom designed recently.

Click here for the two images with full-res version and EXIFs available.

Here they are as resized by Flickr. You can see the smearing in the extreme corners of the Canon shot even at this size (Click image for the intended viewing size):


In contrast, corner softness is barely apparent on the Olympus image at this size (Click image for the intended viewing size):


Here's a representative corner crop (Click image for the intended viewing size):


This crop shows the greater overall resolution and accutance of the particular Canon system used in this comparison (Click image for the intended viewing size):


My observations: 1) Softness in the extreme corners was clearly present on the Canon image and much more subtle with the Olympus image. 2) The overall detail rendered by the Canon system was greater. 3) Noise levels were comparable.

Again, this is just one comparison. While I found the results interesting, more lenses and bodies from each format would have to be compared to reach significant conclusions.

One of the things I've noticed about images from my Olympus E-410 is that they display less "per-pixel" sharpness than those from my Canon 5D. I can attribute this in part to the fact that I use some nice primes with the 5D, whereas I use only the "Standard" class of Olympus glass on my E-410. The weak anti-aliasing filter on the 5D is another likely factor. However, I have wondered whether the Canon's larger pixels were a bit easier on the resolving ability of the lens, since I noticed a step up in sharpness going from the 30D to the 5D. By comparison, all Four Thirds sensors have quite a high pixel density, and I wasn't sure whether Four Thirds glass could deliver the resolution to match. The recent Panasonic DMC-L10 review at DPReview has nicely addressed this issue. Compared to the E-510, which shares the same sensor as my E-410, the L10 sensor appears to deliver superior per-pixel sharpness. See the comparison here. All crops in that comparison were processed identically from RAW, and the excellent Olympus ZD 50mm macro was used on both bodies. The L10 crops show the sort of resolution I have come to expect from the 5D, so I have concluded that the Olympus glass is up to the task. Hopefully the next generation super-compact body from Olympus will have a similar sensor to the one in the L10, In my case, it doesn't matter much given my choice of lenses. The 14-42mm kit lens, which I love, is well matched to the abilities of the E-410 sensor.

16-9.net pitted the new Canon 14mm f/2.8L II against the new Nikon 14-28/2.8 at 14mm. Each was mounted on the new Canon flagship, the 1Ds Mk III, the Nikon using a lens mount adapter. I'm not going to say anything about the results. You just have to see it for yourself. Wow.

Pavel Kudrys posted some GRD II vs GX100 image crop comparisons at Ricoh Forum and was kind enough again to share the source files. This time, he processed the DNG files using dcraw, which is probably the best method one can use to compare the native files. Definitely worth checking out!

Rumors of a compact 25mm (50mm equivalent angle of view) lens for Olympus Digital have been circulating in the DPReview forums. I don't know whether there is any reason to believe them other than the fact that this is what I've been wanting ever since buying the "remarkably small" E-410. The 14-42 zoom is very good, but my preference is to shoot primes. A 17.5, 20, or 25mm compact f/2 ZD lens would make my PMA!

Update - More on this rumor can be found here.

In March, Fujifilm released the Klasse S, a compact 35mm film camera featuring a high quality 38mm f/2.8 lens and fully manual controls. It's also available with a 28mm lens (Klasse W). At $750 shipped on Ebay, the price is a bit steep. The similar Rollei AFM 35 (Fujifilm Klasse) was available at less than half the price not long ago. Nonetheless, I'm strongly considering the new Klasse S. Now if only Fuji would get around to making a Klasse D (Digital)!

Sounds funny, doesn't it? "Compact Cameras with Inherently Deep DOF" describes just about every "point-and-shoot" camera one can think of, yet "P&S photography" isn't the subject of the "4/3rds and Small Sensor Cameras" forum at GetDPI.com. Rather, this new forum has been created to foster discussion of the serious use of relatively small-sensor cameras, in particular those that allow manual control. Thus far, much of the discussion has been focused on the Ricoh GRD/GRD II and GX100, but it is expected that discussion will be cross-brand. The forum is off to a very good start, and hopefully it will flourish. After considering a number of sample images and brief user reviews of the GRD II, it was this comparison posted by Mitch Alland that finally convinced me to go for it.

Addendum: A new forum has been created there for Four Thirds, distinct from the Small Sensor Forum. This makes more sense to me.

I just started a tumblelog for my daily ramblings. A public, personal journal of sorts - mostly non-photography related. Feel free to visit and say hello ;-).

LetsGoDigital has posted a detailed preview of the new Ricoh GR digital II. The preview is accompanied by some very positive editorial comments about the pre-production unit they have in hand. One interesting comment in the preview is that the GRD II uses "the same CCD image sensor as that of the GX100 camera." I haven't seen Ricoh confirm this anywhere, but the GRD 2 RAW files I have seen do have a similar quality to GX100 RAW files. Lars Jeppeson, an owner of both cameras, made a similar observation in the Ricoh Forum and posted some examples of processed ISO 80 and ISO 400 RAWs.

Most if not all of those who own both the GRD and GRD II seem to feel that the latter produces significantly better image quality than the former, particularly at higher ISO. According to the LetsGoDigital preview, the GR lens is said to have been left unchanged with the new model.

Phil Holland has posted an excellent review of Canon's latest flagship, the 1Ds Mark III. Even if you don't have $8000+ to spend on a camera body, this is an interesting read. You can find it here.