The combination of the Canon 17-55/2.8 EF-S and 30D announcements seems to have surprised a lot of people, including me.

I am not at all surprised that Canon has continued to strengthen the credential of EF-S for advanced amateurs and even many "pros who have to pay for their own gear": a fast, apparently high quality f/2.8 standard zoom, roughly equivalent to the 24-104 f/4 for DOF control and speed, and spot metering at last. (Why has Canon omitted spot metering on all previous "amateur" models, while other DSLR brands put it in even entry-level models? A habit that goes back to its film SLRs; lack of spot is my main peeve with my Elan II.)

Canon's investment in the 17-55 f/2.8 IS EF-S strongly suggests that it is not planning to relegate EF-S to "entry level only" in the foreseeable future. The lack of an "L" designation is probably purely a marketing decision, Canon keeping the designations "L" and "professional" only for their 35mm format offerings. The 17-55/2.8 certainly has "L" pricing.

What did surprise me a little was not increasing the pixel count, leaving Nikon and even the Sony R-1 with MP marketing bragging rights, even though the difference between 8MP and 10MP is of very little real significance.
I suppose that this could simply show that even Canon now plans on getting two or three years out of each sensor design, with smaller "mid-year model updates" in between, as in the auto industry. After all the 20D sensor is only 1 1/2 years old while there were 2 1/2 years from the 6MP D60 sensor to the 8MP 20D one.

I think companies have to be wary about competing with themselves. A 12mp 30D would have been too close on the heels of the 5D and 10mp would have represented too small an increment as you've implied. My guess is the 12mp offering will come next time round and not before the 5D's 16mp or 18mp successor.

I think Canon also has to be wary of producing a similar format camera to the Nikon D2X and D200 which does not have at least equal image quality (and of course, desirably greater image quality) in respect of resolution and noise. In a sense Nikon have a slight natural advantage here because their sensor is slightly larger than the 30D's.

I was wondering if canon could maintain their noise lead with a 10mp sensor. It is their claim to fame sensor wise.

To me, the len isn't all that suprising. Canon has sold a great many 20Ds and Rebels (of both series) and, despite many people's feelings otherwise, there is nothing inherently wrong about the 1.6 crop lenses. Many people, serious and not, will stick with a 1.6 body for a long time. Why not cater to those wishes? This is exactly the lens that I've been waiting for for my rebel and the 20 or 30 D that will replace it one day. My only quibble is the price, but I guess for a fast, wide IS zoom that is to be expected. I will be interested to see how this lens fairs.

That's what Dell coupons are for.

I am very skeptical of this frequent suggestion that Canon, or any company, can afford to hobble a product like the 30D in order to protect sales of a far more expensive, lower selling item like the 5D.

Some numbers I have heard are that Nikon has increased D200 prduction levels to 40,000/mo and still cannot keep up with demand so is adding even more production capacity, while Canon has stated a production level of something like 10,000/mo for the 5D and the price has already been cut a bit, suggesting that supply is matching or exceeding demand at the original price. I would also guess that the less expensive 30D has the potential to outsell the D200. Thus the revenue potential of having the 30D compete successfully with the D200 looks more important than protecting those 5D sales.

Anyway, no matter how many pixels Canon could have given the EF-S format 30D, surely the 35mm format 5D would always sell well to those who (a) are willing to pay twice as much, and ( seek the advantages of the larger format in higher DR, higher abberation/diffraction limited resolution, lower minimum DOF, etc.

BJL,
Hobbling appears to have happened with some Canon products. I'm thinking of the 300D which dropped certain features of the 10D resulting in some 300D owners resorting to 'hacks' to activate the features.

But in this instance, I don't think 'hobbling' is the right word and it wasn't what I was suggesting. Don't large companies like Canon have secret 'road maps'? They have products under development in their secret R&D labs and a planned order in which they expect to roll out the new products. There's no doubt a certain degree of flexibility is there to meet changes in market trends. They can bring some products forward a bit, maybe sometimes, but not change the whole road map without great upheaval.

I was personally a little surprised that the 5D which was announced after Nikon's D2X was already showing great commercial success, was not a 12mp 30D. Such a model would have competed more directly with both the D2X and the current D200, but one has to assume that the 5D was well under development long before the D2X was released. One can also assume that Canon has a 12mp 1.6x cropped format camera well under development and possibly already at the prototype stage.



That's understandable. Isn't Canon offering a greater number of DSLR models than Nikon? The public is better served by having some 'real' choices and Canon is already providing that greater choice. (Have a heart, BJL. We don't want to hammer Nikon into the ground ).



Not necessarily. Many of us cannot afford the luxury of buying several camera bodies that vary only slightly in performance and relative advantage. I'm one such person. Having made the decision to get the 5D I would have little interest in a 12mp 30D because it would not be able to do much that the 5D cannot. (I'd be a bit dubious about the ability of those extra 4mp, over the 20D, to eke out greater resolution from the long end of a 100-400 zoom). Likewise, if Canon had planned to release a 12mp 30D last October instead of the 5D, I would certainly have bought it and would have subsequently shown little interest in a later 5D. It seems that Canon's strategy of model roll-out has certainly resulted in my spending more than I might otherwise have.

Although it was only a parenthetical remark, I just thought I'd comment on it.



To me, it certainly seems like the 20D is pushing the boundaries of usefulness for the 100-400 already.

The sample I tested was apparently yielding its best resolution around f/13 to f/16, BTW.

Jani,
Such reckless and extravagant statements need nipping in the bud.

It's true there's a 'law of diminishing returns' at work here. For each proportional increase in sensor resolving power one cannot expect an equal and proportional increase in 'system' resolving power without also an increase in the resolving power of the lens used. This of course does not generally happen. Canon keeps releasing new camera bodies with greater resolving sensors but is apparently not able to provide new lenses with a similar degree of improved performance.

However, the evidence before my eyes does not lead me to conclude that sensors have reached the stage where they can be said to have 'out-resolved' lenses. All you have to do is check out the MTF charts at photodo to see that all good lenses can out-resolve any sensor that's currently available, and sometimes by a big margin. Even the 100-400 IS zoom at 400mm and f8 would appear to at least equal the performance of the 20D. By that I mean, if the lens can deliver 50-60% MTF at 40 lp/mm (the photodo chart) then it can probably deliver 20-30% MTF at 60 lp/mm which is the resolution limit of the 20D.

Since some people have in the past accused me of being too theoretical and impractical, I'm going to put my money where my mouth is, or 'walk the talk' as the Yanks say. The following series of 300-400% crops (the 5D needed greater enlargement) might look like I discovered a new planet in the solar system. But in fact it's merely a piece of a used, medium grade, orbital sanding disc which I stuck on a piece of plywood.

I could show you some crops of a Norman Koren test chart indicating the superior resolving power of the 20D over the 5D (using the 100-400 zoom), but I know some of you are only interested in 'real world' results and have a great disdain for test charts.

Click to view attachment

For those of you who might feel it necessary to question my procedures, or think maybe I didn't focus properly, I provide the following notes.

(1) The 100-400 lens was mounted on a sturdy tripod. I used a remote cord and MLU even though the shutter speed was mostly above any MLU concerns.

(2) I kept shutter speeds as close as possible for both cameras at the same f stop by varying ISO, and of course I did not use IS.

(3) Having focussed on the target to the best of my ability, I kept autofocus off for all shots and used the same focussing for all shots. I did not remove the lens from the camera but the camera from the lens.

(4) The RAW images were converted in ACR with auto-settings off, and luminance smoothing and sharpening at zero. I later tweaked the images in 'levels' only, to get their general appearance similar.

Ray,

That is both impressive and informative. Thanks for sharing it. As a lowly 10D user, I have lusted after both the 20D and the 5D (one look through the 5D optical finder convinced me that it was what a real finder should be like). But, since what I am most interested in is the picture that comes out, I'm now thinking the new ho-hum 30D (or is it 3D0?) might be my next upgrade. Or a 20D, if the prices drop enough.

Would you care to do a similar A-B test at the wide end, with the 20D vs. the 5D? I guess I'd want to see center and near the edge at several apertures, again with matching final image size.

Eric

Since when is the 100-400 even regarded as a particularly sharp lens?

Testing is pretty silly, you're not going to duplicate the test results in the field very often anyway.

For some reason so many people are so interested in these esoteric tests when all these cameras are already at a point where utilizing their performance in the real world is a real challenge.

Eric,
If I have the time and weather permitting I might repeat the test at 180mm which is supposed to be the focal length at which this zoom is sharpest (according to Photodo). But to do a test properly really does take a lot of time. In the above instance I did not want to test for a number different focal lengths whilst the camera was attached to the lens because doing so would have introduced the possibility of focussing differences between all the compared shots.

BenInMA,
You appear to have completely missed the point. Did you not read my post?

This test was partly in response to Jani's statement that he did not think the 100-400 zoom would be good enough to provide any worthwhile resolution advantage with a 20D. This is therefore a test to demonstrate that even the 100-400, which has a reputation for being soft, is still good enough to deliver that extra resolution that the 20D can provide.

Also, these tests were carried out in the field. The piece of sanding disc was shot from a distance of 10-12 metres outside. I would have liked to have got further away but there were bushes and trees at my back which made it too difficult. The used and roughened-up sanding disc represents any fine grained texture you might find anywhere, and the fine lines on a test chart (not shown) which are clearly delineated with the 20D but not with the 5D would represent the fine feathers and hairs sticking out from the fluffy down of a medium sized bird shot from the same distance.

A friend of mine recently upgraded from the 20D to the 5D. He said the image quality is better in all respects from the 20D--but he only does landscape photography, and uses no zooms except something like a 17-35, etc. He attributes this to the superior sensor and the increase to 12.8 MPs, since the processor is the same in both cameras. At 20x30 print, he said there is just no comparison between the 20D and the 5D. After all, you're getting about 30% more pixels with the 5D. However, as stated copiously in other threads, you are losing zoom--which he didn't care for, since he does 99% landscape. In another thread, there is a link to why Canon did not increase the pixel count with the 30D--sensor size. They thought it would be too noisy.

At this moment, I'm in the middle of extracting the 5D sensor and fabricating its reset into my 20D. Then I'll have what we all wanted--a 20D with 13MPs--I jest.

Well, that is simply not a true statement, as I've just demonstrated .