Irwin Puts, 400D versus 5D, 10-22 vs 17-40, etc?

[fast_primes]

Yes, I do know that generally speaking, the 5D will best the smaller 1.6 dslrs.

However, I've heard Canon users chafe at the lack of superb quality Canon wide

angles--L or no L lenses, with some going so far as to track down and use Zeiss

FF wides. So is it possible, that the highly regarded 10-22 EF-S zoom coupled

with a 400D, can match or exceed the performance of a Canon 5D+ Canon wide angle

zoom or prime (the 24F2.8)?

 

In fact, Irwin Puts states flatly that the 60mm EF-S Macro will outperform the

50F2.5 Macro on a full frame digital. Have others observed this? How does the

400D/Sigma 30F1.4 fare against a 5D+50F1.4 EOS combo, or the 400D/60EF-S Macro

versus the 5D/100EOS Macro?

 

I have not seen Irwin's comments discussed in this forum.

[gdanmitchell]

Can your provide a link to his comments?

 

Thanks.

[fast_primes]

<i>"A case in point is the introduction of the new Canon macro lens 2.8/60mm that offers vastly improved performance compared to the older 2.5/50mm macro lens for film based photography. Canon explicitly states that the lens cannot be used on large image areas (be they silver halide or solid state). Canon therefore implies that it is impossible to deliver the equivalent quality when having to cover larger capture areas."<br><br>

"A Canon 2.5/50mm macro lens and a 16 Mp 24x36mm sensor will bring less image quality than a Canon 2.8/60mm and a eight Mp solid-state sensor with an area of 17x 24mm."</i><br><br>

 

Link: <a href="http://www.imx.nl/photosite/comments/c014.html">Why the full size (35mm) solid-state imaging sensor will be a pipe dream. (june 24, 2005)</a>

[fast_primes]

Also, apologies for mispelling Erwin's first name! It should be Erwin--not Irwin!

[steve_dunn2]

<cite>Canon explicitly states that the lens cannot be used on large image areas (be they silver halide or solid state). Canon therefore implies that it is impossible to deliver the equivalent quality when having to cover larger capture areas.</cite>

 

<p>If that's an example of the logic upon which his conclusion is based, why would you pay attention to him?</p>

 

<p>That's absolutely not the reason why Canon explicitly states that the lens cannot be used on "large image areas." The reason is that the lens is designed to project an image circle large enough for a 1.6-crop body; if you were to modify it to mount on a 1.3-crop or full-frame body (assuming that there's enough clearance behind the rearmost element for the mirror), you'd find that the image circle isn't big enough to cover the whole frame. Pure and simple. Nothing to do with quality, and anyone who has used the 50/2.5 will tell you there is absolutely nothing wrong with the quality of images it produces.</p>

[bobatkins]

<strike>Putz</strike> Puts is clearly missing the point.

<p>

Anyone who has actually made large prints from a 16 MP 1Ds Mk II and an 8/10 MP XT/XTi will probably have quite a different opinion.

[jbq]

If the reasoning was correct, small-sensor point-and-shoots would have the best image quality, and medium-format backs would have the worst one.

 

As for the comparison between the 50/2.5 and 60/2.8, photozone found that the center sharpness of the 50 (on 350D) is within 5% of that of the 60, that the 50 slightly beats the 60 at the edges of the 350D frame. If you consider that the 60 only covers 40% of the surface that the 50 covers and that the 60 costs more than 50% more, the 50 on a full frame is actually a very good deal in my opinion.

[dogbert]

To put it another way, small frame lenses need to be sharper to deliver equivalent image quality becuase the smaller images from the sensor require greater enlargement. For this reason while the 60 /2.8 may be sharper than the 50 /2.5 on a 350D, the 50 on FF will easily outperform the 60/ 2.8 on a 350D/400D.

 

It is well know that large and medium format lenses are mostly not very sharp at all relative to 35 mm. They don't need to be because the images they produce don't require as much enlargement in the first place.

[Brad_]

<I><s>Putz</s> Puts is ...</i><P>

 

Mighty feisty, come on down to the leica forum!

[John Crowe]

While I think Puts goes a little overboard I do tend to agree with a few of his points. I would love to see an XTi with a Canon EF 200mm f2.8 L put up against a 1DsII with a Canon EF 300mm f2.8 L. Even more interesting would be the XTi with a Canon EF 400/2.8 L against a 1DsII with the 400/2.8 L and 1.4x converter. While not a technically perfect test both scenarios would provide similar views on their respective formats and provide real world choices between the two systems. I believe the smaller format could very well produce superior results in telephoto photography.

 

 

I suppose that with the test I recommend above I am actually hinting at the idea of using higher quality glass on the smaller format to try to out perform lower quality glass on the larger format, as Puts suggests as a possibility. At least with telephoto lenses.

 

 

Super wideangle photography (wider than 20mm on 35mm film) gets a little more murky since the existing glass for full frame DSLRs was developed over many decades, while the development of superwide optics for 1.6x factor bodies is very new and so far the best optical design techniques have not been employed in any of these new lenses. Here, I suspect, the full frame bodies, at least for now, will reign supreme.

 

 

Unfortunately 1DsII users are too busy making money to pay for their systems to fully test them against lowly cameras like the XTi, so we may never know how they stack up.

[jbq]

I have to agree with John only at the edge case where it's not practical to get a long enough equivalent lens on the full-frame body, and that's only because at the moment the smallest pixels are available on the smallest sensors. While this happens in a few cases, those aren't common.

[fourfa]

take Puts's writing about digital issues with many grains of salt. the guy is famous for testing lenses with tech pan film and a grain microscope, I have no doubt he's good at that. his digital articles I've read really miss the point - he gleefully crops to high heaven in a test of resolution, he's never heard of sharpening, he doesn't seem to understand how to use RAW formats, and field of view equivalence is foreign to him. for instance in this case he needs to compare the EF-S 60mm 2.8 to the 100mm 2.8 macro. In another test he attempts to compare the 5D and the Leica M8 by mounting a 75mm Summicron to the Leica, and a 24-105 zoom to the Canon apparently thinking "eh, close enough"

[gluteal cleft]

From what I've read (which may or may not be accurate), the 100mm macro on a full-frame beats the EF-S 60mm on an APS-C (and would give you about the same FOV), and the 17-40 f/4 on a full-frame beats the 10-22 on an APS-C, even in the corners.

 

That's not to say that the EF-s 60mm and EF-s 10-22 aren't fine lenses, they are.

[andrew robertson]

Puts has shown that he is still getting his feet wet in digital. Some of his conclusions are

quite amusing and dead wrong. Look at his early digital reviews to see what I mean.

[robin_sibson1]

"Vastly superior"? I don't think so. If you look at the Photozone tests (OK, OK, all the usual caveats) you will see that, with both lenses tested on a 350D/XT, (a) the 50/2.5 has even lower distortion than the 60/2.8, both described as "absolutely negligible", (b) the 50/2.5 is a clear winner regarding vignetting/falloff - not suprising since it is designed for FF, © the 60/2.8 has marginally better resolution, but the differences may well be at the level of sample variation, and (d) the CA profiles follow different patterns, with the 50/2.5 almost independent of aperture and the 60/2.8 increasing as the lens is stopped down. Overall, either a dead heat or a slight edge for the 50/2.5. Of course, that's not telling us what happens with the 50/2.5 in the part of FF outside the APS-C area, but that's not too sensible a comparison. As "fast primes" recognises, it is the 100/2.8USM on FF that compares directly with the 60/2.8 on APS-C - unfortunately, there is no APS-C short macro. All we have from Photozone is the performance of the 100/2.8 across the APS-C sensor, and that again looks pretty comparable, but with even lower distortion and considerably better CA. If image quality is the only consideration, the 100/2.8USM on FF looks as if it should out-perform the 60/2.8 on APS-C - at least, you'd be sticking your neck out a bit to assert the contrary in the absence of proper comparative tests.

 

Back to the real world - for anything except possibly very critical flat-subject copying work, the optical performance of all of these lenses is simply not a limiting factor - I use all three on a 20D and previously used the two FF lenses on film.

 

Comparison between the 10~22 on APS-C and the 17~40 on FF is a more interesting question. Such reports as I have seen suggest that the FF combination has the edge, but not by a large margin - I know from personal experience how well the 20D+10~22 combination performs, so nether the ranking nor the margin are any surprise.

 

Looking at the other responses on this thread suggests a possible answer to your question about why there is little discussion on this forum of comments by Erwin Puts - it's hard to engage with stuff that, in the immortal phrase of a distinguished physicist, is "not even wrong".

[yanglee]

I would rent/borrow the equipment and make an assessment yourself, then purchase the one you are happy enough with without breaking your wallet. The web biases whiners who tend to complain more than they praise. My experience with the 5D with L lenses is that they are quite good in pixel resolution (I assume that's what you meant by "performance"), equivalent at least to most well-scanned variety of MF. Good luck.

[mark_ryan2]

I've actually been contemplating going the other way and getting a wide angle lens for a 6x9 Rangefinder camera and scanning the negatives.

 

I find that with my 17-40 on my 20D that wide angle pictures, especially things like natural scenes with banches or grass (random details) the detials get muddled.

 

I figure with all that negative realestate, I'll capture the details and then scan them out of the negative. That is if the "6D" doesn't have 16+MP

[steve_spencer2]

William Castleman has an interesting short review that compares the 17-40L, 16-35L on a

full frame to the 10-22mm on a 20D. You can see it at the following link.

 

www.wlcastleman.com/equip/reviews/16-35/index.htm

 

The 17-40 is generally sharper in centers and the corners and has less CA, but the 10-22

has less distortion and vignetting.

[desmond_kidman]

All readers who are tempted to give much merit to Erwin's writing should go back a few years and sample his writings from various periods to see how his train of thought is truly without direction. He is self contradictory on many occasions and jumps to technical conclusions with truly convoluted logic.

 

The truth is that nobody knows what the eventual format size at the top performance levels will be. It could be larger, smaller, or somewhere in between. Anybody who claims to know, or even have a well-educated guess, about how the optical/sensor/producibility tradeoffs will stack up a year down road is seriously deluded. Exceptions might be folks at the upper levels of R&D at Canon and Nikon.

[conraderb]

the first few paragraphs that I read from erwin sounded interesting until I came to this part:

 

A case in point is the introduction of the new Canon macro lens 2.8/60mm that offers vastly improved performance compared to the older 2.5/50mm macro lens for film based photography. Canon explicitly states that the lens cannot be used on large image areas (be they silver halide or solid state). Canon therefore implies that it is impossible to deliver the equivalent quality when having to cover larger capture areas.

 

 

this guy and needs to take a basic course in mechanics! what a moron. if he doesn't understand the difference between an EF and an EF-S style lens mount, I don't understand how anyone can take his word on anything else...