Canon 35mm f/2 IS resolution comparison
/forum/topic/1173850/0



Fred Miranda
Registered: Dec 31, 2001
Total Posts: 18132
Country: United States

Canon 35mm f/2 IS resolution test
Roger Cicala just posted a resolution test of the new Canon 35mm f/2 IS and compared it with the Canon 35mm f/1.4L, Canon 35mm f/2 and the Sigma 35mm f/1.4.

Here is an excerpt:
"The new 35mm f/2 IS and the classic 35mm f/1.4 L do much better in the corners, with the 35 L (stopped down to f/2) clearly better than the new f/2 IS. But the Sigma does better than any of them."

Read the entire article titled: Another 35mm Lens for Canon

In Stock at Amazon.com: Canon EF35mm f/2 IS USM



Tom Dix
Registered: Jun 29, 2010
Total Posts: 1711
Country: United States

Thanks for posting Fred. Thank you Roger for The comparison.
Interesting how well the Sigma performed. I love my 35L, yet, am now interested
In the new Sigma and other Sigma lenses that will follow.
Sincerely hope the QA and the af are solid on the new Sigma lenses.



vsg28
Registered: May 07, 2012
Total Posts: 1227
Country: United States

Thanks for the link, Fred. I am amazed at Sigma's performance- good for them reaping the benefits of having risked making fast AF primes when no other 3rd party maker does them for Canon/Nikon. I am now VERY excited for the new 120-300 OS.



jamato8
Registered: Dec 24, 2005
Total Posts: 2213
Country: United States

So if Sigma can do it, so can Canon. What is the deal?



skibum5
Registered: Jan 21, 2005
Total Posts: 16899
Country: United States

Hmm that Sigma really is looking like quite the lens (wonder a bit about AF though, maybe a bit more misses means real world shots blurrier on avg shooting wide open with super low DOF?). Not that i'm in the market for any more camera stuff now, plenty spent out on it.


I hope the new 50mm IS is good. I sort of almost wish it wasn't planned for IS, since it means $$$ and perhaps worse IQ. But who knows.



Snopchenko
Registered: May 19, 2010
Total Posts: 2350
Country: Russia

vsg28 wrote:
Thanks for the link, Fred. I am amazed at Sigma's performance- good for them reaping the benefits of having risked making fast AF primes when no other 3rd party maker does them for Canon/Nikon. I am now VERY excited for the new 120-300 OS.

I believe the Sony users will be the ones who benefit the most - the Sony 35/1.4 lens is dated mechanically and is reported to have atrocious IQ at wider apertures. And a lot more expensive to boot.

jamato8 wrote:
So if Sigma can do it, so can Canon. What is the deal?

I may believe that the 35/1.4 II that is going to be released eventually will be better yet than the Sigma... at 3x the price.

skibum5 wrote:
Hmm that Sigma really is looking like quite the lens (wonder a bit about AF though, maybe a bit more misses means real world shots blurrier on avg shooting wide open with super low DOF?). Not that i'm in the market for any more camera stuff now, plenty spent out on it.

So far the reports seem unanimously positive.


I hope the new 50mm IS is good. I sort of almost wish it wasn't planned for IS, since it means $$$ and perhaps worse IQ. But who knows.

All the IS primes have been tested to match or exceed the 20+ year old non-IS ones in IQ. What makes you think the 50 is going to be any different? Can't agree more with the $$$ part though.



Pixel Perfect
Registered: Aug 16, 2004
Total Posts: 19928
Country: Australia

Snopchenko wrote:

I may believe that the 35/1.4 II that is going to be released eventually will be better yet than the Sigma... at 3x the price.



It's hard to see what improvements the 35L II could bring to the table. How sharp do you want it? Maybe the 35L II could produce smoother bokeh. At some point it's a law of diminishing returns just like high end audio speakers. I'd find it hard to pay any more even if the 35L II were better. Maybe if the 35L II had IS and was still $1300-1400 it might be tempting, but otherwise the Sigma is going to be value king for the level of performance it delivers. Also I'd expect the Sigma prices to drift downward in the next 6-12 months making it even better value.



thw2
Registered: Dec 27, 2004
Total Posts: 2931
Country: N/A

I looked at Photozone single lens copy test results for Canon 35 f/1.4 vs Canon 35 f/2 classic, both tested on Canon 5D Mark 2. At f/2, the results are:
35 f/1.4 center 3629 edge 1685
35 f/2 center 3122 edge 2516
So, the f/1.4L has great center performance but AWFUL edge performance even when stopped down to f/2. The classic f/2 lens has much better edge performance.

Roger's f/2 results are:
35 f/1.4 center 830 corner 490
35 f/2 center 835 corner 240
Here, both lenses are equally good in the center but the classic f/2 demonstrates horrible edge performance.

So, what can we conclude from the above differences? Testing methodology matters or that LensRentals only has bad copies of the classic 35 f/2?



alundeb
Registered: Nov 06, 2005
Total Posts: 4561
Country: Norway

thw2 wrote:
I looked at Photozone single lens copy test results for Canon 35 f/1.4 vs Canon 35 f/2 classic, both tested on Canon 5D Mark 2. At f/2, the results are:
35 f/1.4 center 3629 edge 1685
35 f/2 center 3122 edge 2516
So, the f/1.4L has great center performance but AWFUL edge performance even when stopped down to f/2. The classic f/2 lens has much better edge performance.

Roger's f/2 results are:
35 f/1.4 center 830 corner 490
35 f/2 center 835 corner 240
Here, both lenses are equally good in the center but the classic f/2 demonstrates horrible edge performance.

So, what can we conclude from the above differences? Testing methodology matters or that LensRentals only has bad copies of the classic 35 f/2?


Curvature of field?

Photozone in their method refocus for the corners. I don't think LensRentals do, but Roger Cicala will probably be here in a minute to answer that

http://www.lensrentals.com/blog/2012/09/why-arent-the-damn-numbers-the-same



RCicala
Registered: Jan 09, 2005
Total Posts: 2956
Country: United States

alundeb wrote:
thw2 wrote:
I looked at Photozone single lens copy test results for Canon 35 f/1.4 vs Canon 35 f/2 classic, both tested on Canon 5D Mark 2. At f/2, the results are:
35 f/1.4 center 3629 edge 1685
35 f/2 center 3122 edge 2516
So, the f/1.4L has great center performance but AWFUL edge performance even when stopped down to f/2. The classic f/2 lens has much better edge performance.

Roger's f/2 results are:
35 f/1.4 center 830 corner 490
35 f/2 center 835 corner 240
Here, both lenses are equally good in the center but the classic f/2 demonstrates horrible edge performance.

So, what can we conclude from the above differences? Testing methodology matters or that LensRentals only has bad copies of the classic 35 f/2?


Curvature of field?

Photozone in their method refocus for the corners. I don't think LensRentals do, but Roger Cicala will probably be here in a minute to answer that

http://www.lensrentals.com/blog/2012/09/why-arent-the-damn-numbers-the-same


Testing methodology for certain. I don't refocus for corners, we use the best center focus and take the corner results as they lie.

There's also testing distance (small variable) and chart differences (depending on what you analyze, corners may be 85% to 95% of the way to the actual corner. Finally the 4 corners are never exactly the same - you could use highest, lowest, average, for that value. I use average, not sure what they use.

Bottom line - this is why we need to look at several peoples reports and balance them to get an idea of what is reality. I usually end up putting out the first report, just because I get early access. But then I'm like everyone else, looking to see what others say to confirm or modify what I saw.



andyjaggy82
Registered: Jan 25, 2006
Total Posts: 1371
Country: United States

I love that 3rd party lenses are kicking Canon lenses in the butt... and they are still cheaper. Good news for everyone.



gdanmitchell
Registered: Jun 28, 2009
Total Posts: 9831
Country: United States

The charts somewhat confirm why I continue to be so happy with the performance of my old 35mm f/2 lens - small and humble, but a fine performer.

Also, the chart demonstrates that charts don't tell the whole story! Something I know very well from using the 35mm f/2 (old model) quite a bit is that the very furthest corners can be a little soft - though usually not enough to even notice. However - and what the chart does not tell us - is that the drop-off in sharpness between the measured center point and the measured corner is not linear. Instead, the 35mm f/2 remains very good until just before the far corners, at which point the decline in sharpness kicks in. Other lenses - some of which I own or have owned - showed a decline in sharpness that is more gradual and starts further into the frame.

Such charts that inform us of performance at two points in the frame (center and corner) provide a lot of interesting information, but there is also a lot that they don't tell us. They are a bit like reporting "the weather in the United States" by telling us what it is like only in Anchorage and in Dallas.

And, as always, we are reassured by this report that there are a whole bunch of 35mm primes for our Canon cameras that produce truly excellent image quality, and that the choice among them is more about things like cost, size, intended use, and so forth... than about the single data point of "sharpness."

Dan

(Of course, as always, it is smart to not assign too much significance to the results of a specific test, especially given the history of tests that can be quite different from one another. It is also really important to recognize comments like "A kicks B's butt" for the hyperbole that they represent.)



gdanmitchell
Registered: Jun 28, 2009
Total Posts: 9831
Country: United States

thw2 wrote:
35 f/2 center 835 corner 240
Here, both lenses are equally good in the center but the classic f/2 demonstrates horrible edge performance.


Actually, I think this is corner performance, not "edge" performance. See my previous note about the nature of the roll-off in sharpness on the old 35mm f/2.

Dan