Sharp & Sharp; Canon EF's compared to new Lumix S-lenses

[ermesjo]

When I retired my Canon gear in late September 2024, switching from Canon EOS 5Dii (released in September 2008, bought in 2011), over to Lumix S5iiX, I kept my five EF lenses: The 100mm 2.0, 50 1.8, 17-40 4.0 (released in May 2003), the 24-70 2.8 and finally the 70-200 4.0. All of them "L" except for the 100 and 50. When I closely observe the quality of my EF 17-40 on my S5iiX, blowing pictures up, I must admit that the lens isn' THAT sharp. Yes, it has that "movie-like" feel to it, slightly chromatic aberrations that give pictures a tiny-little-bit of "mushy" extra color from center to corners. But the lens is not "tack-sharp". Another issue with the Canon EOS 5Dii, is that over ISO 800, pictures starts to be quite grainy. And of course, filming in HD, which was a revolution for Canon in 2008 (many pop/rock videos was filmed with the EOS 5Dii back in the days) - was the start of the DSLR's as hybrid cameras. The Lumix S5ii / S5iiX is in another league here. RAW, open gate, native ISO, hundreds of options in the video menu, Netflix-approved. Hybrids are here to stay.

The lenses from film to digital:

1. The advances in technology has raised the bar. Many lenses from the 1990-2000, was designed for the 135 film age, when the digital sensors replaced film. For instance was Zeiss lenses for Hasselblad discontinued and Fujinon lenses brought into the new digital area. Zeiss was legendary, but the lenses from Japan very much sharper. I have a Hasselblad 500 C/M with the Zeis 80mm f/2.8 myself. Very sharp (much because of the 4x larger film area compared to an 135 film), but the 80mm will not stand the test compared to the requirements of digital sensors. The resolution factor in the lenses have to be good enough to resolve the pixel density of the digital sensors. Today you've got the Hasselblad XCD 55mm f/2.5 as the "standard" normal lens, but the optical performance is in another level compared to that of the film age. (The 80mm Zeiss dates back to the 1950's with only 1-2 improvements).
2. I read an article that suspected that the Leica Summicron 50/2.0 had similar optical design as the Lumix 50/1.8 (when take into account no. of lenses, aspherical units, weight, etc.). I guess that the Lumix S-lenses lineup (also the S Pro - editions) are at a very high level. (Leica products has high price as part of the technical performance - LOL).

So, when we compare the Canon EF 17-40 f/4.0 to say a Lumix S 14-28 f/4.0-5.6, 20-60 or an S 18-40, or S 18 f/1.8 ? If the lenses where athletes like 100 meter speed runners in an Olympic stadium, how many 100's split of of a second will the old Canon lag behind our new Lumx S'es or S Pro's ? Sharpness, chromatic failures, breathing, flare, centers, corners, ... ?

​

 

[pdk42]

I ran Canon gear until 2013 when I switched to Olympus. I had a 5dii and a few lenses (17-40, 24-105, 70-200/4, 100-400, 50/1.4). The 17-40 was really never that sharp, esp in the corners (even at f11). The Olympus 9-18 which I "replaced" it with was definitely better, even though that lens isn't the best UWA in the m43 catalogue. For sure, the 14-28 I run now for the Lumix gear is a lot better than both - and it's not even an expensive lens.

The Canon 24-105 was a bit better than the 17-40. The 70-200 was very good and the 100-400 was decent enough, but not exceptional. The 50/1.4 was great when it focussed accurately, but the 5dii's AF wasn't great. And for sure, that 21Mp sensor was a long way behind modern FF sensors.

 

[ermesjo]

Yes, I have the similar experience. My Canon 17-40 is OK, but smears it's "chromatic" kind of all over the place. But looking at pictures, they got character, meaning the lack of perfect sharpness is acceptable. When I bought the EF 24-70 f/2.8 I thought I had acquired the optimum optical, but this lens has disappointed me. I have blown up two posters with canvas on my walls in 2 by 3 meters taken with the EOS 5Dii. The one taken with my 70-200 f/4.0 and the other with the 24-70 f/2.8. The telezoom is what I will say is "tack-sharp" (the really best one), but the 24-70 is just not really (optimal) sharp. In use, it also has some "breathing" problems (sharpness is not even from center to corners, you must adjust and it zooms a little bit when focusing). I regard the 70-200 as best (even at today's standard), then the 17-40, and at last the 24-70. My 100mm f/2.0 is my favorite, often used when I had paid jobs too. The plasticy (even the mount is plastic) EF 50 f/1.8 is a classic lens with the "gauss" 50mm optical formula. Not sharp in the corners at f / 1.8 to 4.0 but with "personal signature" otherwise very sharp in f / 4.5-8.0. I expect that the Lumix S/S-Pro lenses are much better. I "robbed the bank" when I bought my Lumix gear in September, but hope to save money for the 14-28mm later this year. My motivation for switching to Lumix is that I think Canon does not deserve my money. You've got to buy a new R-model (both the R5 and R6 lack behind video specs, and Lumix is on par regarding autofocus with the MK2-editions) and finally you have to buy ne RF lenses as well. I think the new Canon lenses are unnecessary expensive and also very bulky. It does not help that the camera bodies are small. (Why use a Canon EF/RF adapter on a Canon, when you can use a Sigma EF/L-mount adapter).

 

[GeorgeHudetz]

As you have found, it depends on the lens. Some DSLR lenses hold up quite well on a 24 MP sensor; others do not. I have an older (DSLR era) Sigma 12-24 F4 "HSM" Art lens that actually looks quite good on my S1R, even using HR mode. At least, below 22mm or so.

Another thing to be aware of is that those older, film-era lenses were made with a different set of compromises, as I understand it. Particularly with regard to distortion, since it's basically impossible to correct distortion in a darkroom. But, digitally, it's relatively easy to correct basic barrel and pincushion distortion automatically, so the lens designers allow for distortion in their design in order to optimize some other parameter. Although one could argue some of the modern lens designs push this too far. Same with vignetting. The new sensors have almost no noise when pushed several stops, so the designers are not overly worried about vignetting on modern designs. So it is said.

I've also heard people say that current lens design emphasizes test results (particularly: sharpness/resolution) over final IQ, but I'm not going to argue that one way or another.

 

[Travis Butler]

Another thing to be aware of is that those older, film-era lenses were made with a different set of compromises, as I understand it. Particularly with regard to distortion, since it's basically impossible to correct distortion in a darkroom. But, digitally, it's relatively easy to correct basic barrel and pincushion distortion automatically, so the lens designers allow for distortion in their design in order to optimize some other parameter.

Yeah... for example, the author of the Lens Genealogy articles (from the Lens Design post) said: "Therefore, the distortion found in two very sharp retrofocus lenses I shoot with (The Zeiss 21mm and Nikon 24mm f/1.4) may have been a conscious choice of the designer to maintain sharpness, not a design flaw."

Although one could argue some of the modern lens designs push this too far. Same with vignetting. The new sensors have almost no noise when pushed several stops, so the designers are not overly worried about vignetting on modern designs. So it is said.

A hypothesis I have is that since coating improvements made it possible to add lots of corrective elements to a lens, designers will deliberately push a lens formula in the tuning stage to optimize a performance characteristic - even if that adds an aberration - because you can then add a corrective element to fix that distortion. Which is one way you end up with lenses that have lots of corrective elements, are very well-corrected... and big, heavy, and expensive.

I've also heard people say that current lens design emphasizes test results (particularly: sharpness/resolution) over final IQ, but I'm not going to argue that one way or another.

I've heard the same thing, yeah. As someone who prioritizes color/tonality over razor sharpness (at least as long as the lens meets minimum sharpness levels), this is not optimal for me. ^^;;