OK, I did another comparison. I placed the 35L, 35mm f/2 and 24-70L @ 35mm at about 0.35 meters from the target (5D2 + tripod). Lighting/exposure the same for all frames.
This is the full frame shot. 100% (almost - flickr resizing) bokeh crops are taken from the card to the right side of the mannequin:
So when it comes to quantity/quality of bokeh when using these lenses up-close:
-the 35L is the best option at and under f/2.0
-all three lenses are very similair at f/2.8
-the 24-70L seems to be the best option at f/4 and smaller
Daan wrote: I did this comparison (not a test) to appreciate the differences between these lenses. I own (and use) all three of them. Comparisons like this give me a better insight which lens(es) to take for which task/job at hand. In other words... I did the comparison for myself. But I am willing to share the results (and put some effort into that) so that other can -MAYBE- benefit from it. If you have something against that, please move along (instead of putting somebody else on the soapbox). Thank you
You're right. I guess we all have our own ways to blow time.
Nice Daan. I've done a few similar things, though less formally. I too am fascinated by the differences betweeen these lenses.
For instance, I have both the 24L II and the 24 2.8, and I have observed that the 24L II tends to render backgrounds softer overall at the same aperture than the slower lens. It's as if there is a more rapid transition from the in-focus to the out-of-focus, which almost looks like shallower DOF.
This was a surprise to me, as while I had understood that the quality of the bokeh could vary, I had thought that the amount of blur was a strict function of the focal length and aperture. But apparently lens design can play a role too.
garyvot wrote:
This was a surprise to me, as while I had understood that the quality of the bokeh could vary, I had thought that the amount of blur was a strict function of the focal length and aperture. But apparently lens design can play a role too.
Yeah, lens design definetely plays a big part in this. Below are shots of each lens @ f/2.8 (5D2 + tripod didn't move for each test shot). Notice how the 24-70L seems to magnify the foreground in comparison to the other lenses, while the background is very similair. This has -no doubt- something to do with the bokeh advantage of the 24-70L over the other lenses when used up-close (from f/2.8 and smaller). An advantage that the 24-70L quickly looses when evaluating the bokeh when shooting at longer focus distances (where the 35L has the advantage at comparable apertures).
Thanks for these comparisons Daan. I think they are very helpful. For me and my tastes I don't fine either of the 35s to have very good bokeh. One of the reasons that getting a 35L isn't very high on my list. If they make a MKII version I hope the pay particular attention to this issue.
Steve Spencer wrote:
For me and my tastes I don't fine either of the 35s to have very good bokeh. One of the reasons that getting a 35L isn't very high on my list. If they make a MKII version I hope the pay particular attention to this issue.
I don't really disagree with you, but I will say the 35L is capable of rendering lovely backgrounds when shot wide open or nearly so. It's a look that cannot really be duplicated with an f/2.8 zoom lens, no matter how nice the zoom's bokeh is. Here's one at f/1.4 (5D Mark I)... (Note the lack of veiling flare in this shot. I have a very good copy of the 35 f/2, but it would suffer badly by comparison in this situation.)
garyvot wrote:
I don't really disagree with you, but I will say the 35L is capable of rendering lovely backgrounds when shot wide open or nearly so. It's a look that cannot really be duplicated with an f/2.8 zoom lens, no matter how nice the zoom's bokeh is. Here's one at f/1.4 (5D Mark I)... (Note the lack of veiling flare in this shot. I have a very good copy of the 35 f/2, but it would suffer badly by comparison in this situation.)
Bokeh and wide angle lenses is always a tricky thing... But I think Canon managed to make the 24L II somewhat better in this regard than the original 24L. So, it can be done... (disclaimer: I don't have any direct experience with either the 24L or 24L II, just heard this through the grapevine).
But anyway, the 35L should be glued to f/1.4-f/2.0... with a severe penalty for those who uses it beyond f/2.8
To me the strength of the 35L lies in its wider angle + larger aperture. Perfect to isolate a subject in its environment. Bokeh at f/1.4-f/2.0 is quite good/pleasant. Beyond that it ranges from average to non-pleasant IMO.
The 35L isn't perfect. But I agree, it allows me to get shots I can't get with other lenses. Not even with the 35mm f/2. But Canon should put some effort in improving bokeh quality (at smaller apertures), CA's and (extreme) corner sharpness when designing the 35L II
I like these tests because they are relatively objective and the focal length is standardised so interpretation can be more easily made.
But as for the 35L in general, compared to other lenses (of other focal lengths), it's really hard to compare, so RE: the comment a few posts back, I'm not sure it's fair to say the bokeh isn't very good compared to other non 35mm lenses
I don't know how to interpret bokeh anymore, there are so many variables that affect the look of the OOF areas. I don't even know if it is the blurry backgrounds that I like, or good bokeh that I like. Things like the focal length and subject to background distance make sure a huge difference to the rendition of OOF areas, it's hard to interpret what is good bokeh and what is just blurring of backgrounds.
For instance, most people would agree that the 85L has better bokeh than the 35L but does it really?
Dawei, there is a difference between the amount of blur and the characteristics of the blur.
Things like donut shapped OOFH's and double contours are considered non-pleasant because they can be very distracting. OTOH, if you are after such a "specific" distracting look (for whatever resons), those characteristics might be desirable
