Yes, Canon's hybrid IS works

paddler4

There has been a lot of debate about whether hybrid IS will help with macro. I got my 100mm f/2.8 L today and tested IS before dinner. These are quick and dirty shots, but they convinced me.

Based on other lenses, with my crop sensor camera, I would expect to get a reasonable percentage of sharp photos only at 1/(1.6*focal length). I occasionally can manage 1/focal length with my 60mm macro.

These are all 100% cropped, or 1:1 in lightroom parlance. All were shot handheld with a 50D under fairly dim available light. That's why some of them are very noisy--I had to shoot some at ISO 1600, and when you blow them up to 100%, the noise is substantial. I shot RAW and did nothing to them except import them into lightroom, turn off all sharpening, crop to 100%, and export directly to smugmug, with export sharpening turned off. In one case, I upped the exposure. That's it. All were shot with autofocus, center point only.

1/30 f/5.6, seated. This is a hair on a Christmas cactus


1/60, f/8.0, seated. A quarter. Not parallel to focal plane. focused on the 8


1/60 f/5.6, edge of quarter. Seated


1/40, f/6.3, text from The Economist. Exposure adjusted in LR. Focused on the O. Standing, stooped

Lord Vetinari

Looks pretty convincing - comparison with it off and on would perhaps be more telling ?

Brian V.

paddler4

Brian,

Indeed, this was just some first impressions, not a real test. However, before I bought, I read a bunch of reviews, and one--unfortunately, I can't remember which--took the time to do this right: shooting multiple shots at each shutter speed, with IS on and off, and tabulating the percentage sharp in each case. If I recall, my reading of their graphs indicated that they were getting about a stop and a half at minimum focal distance. That seems to fit with these and some others I did not post. I think if I did the test right, I might be getting two stops.

Dan

DQE

paddler4 wrote:

Brian,

Indeed, this was just some first impressions, not a real test. However, before I bought, I read a bunch of reviews, and one--unfortunately, I can't remember which--took the time to do this right: shooting multiple shots at each shutter speed, with IS on and off, and tabulating the percentage sharp in each case. If I recall, my reading of their graphs indicated that they were getting about a stop and a half at minimum focal distance. That seems to fit with these and some others I did not post. I think if I did the test right, I might be getting two stops.

Dan

I don't know how to make reliable conclusions without comparisons from a carefully controlled experiment. Otherwise, it's just a subjective impression which may be correct or incorrect. I know since I have such impressions all the time myself!

Here's my suggestion: make about 10 photos with and without the IS at 1x magnification. Have several people rate them for sharpness and detail without knowing which was made with IS, and tabulate the data. Posting a representative photo from each group at various exposure times would be very interesting to many other macro photographers. If you attempt this experiment, be sure and pick a subject with lots of fine detail, and figure out some way to use exactly the same positioning, focusing, and framing for each photo so that these differences won't confuse things.

It's usually quite difficult to make an evidence-based conclusion, but it's always worth it if it's done correctly.

I'm sure you'll enjoy your new gear and that you've made good choices. I would personally guess that IS will be most useful for hand-held, low natural light macro photos and of course non-macro photography.

I hope my comments and suggestions are helpful!

paddler4

Of course, you're right--that's how to do a serious test. I have done a number of systematic tests of other issues and posted them one place or another. Didn't do it this time for two reasons: they have already been done and posted by others on other sites, and I already bought this lens, so it is too late for me anyway! I just wanted to share an initial experience, which is that I was able to get some sharp shots at shutter speeds that would have been pretty much out of the question for me without IS. I don't have particularly steady hands. I did a bunch more shots with and without IS after posting, and I got a total of two that were reasonably sharp without IS, and lots with. But I am not trying to persuade anyone.

GOLDENORFE

have you tried any shots with flash yet?
i know my 70-300, the " IS" doesnt work well with flash or tripod.
i have to turn it off
phil

paddler4

Phil,

Older versions of Canon's IS are supposed to be turned off when you use a tripod (not monopod). With this one, Canon is less definitive: "depending on the kind of tripod and shooting conditions, sometimes it may be better to turn of [IS]." That's too vague to be helpful, but then again, with a tripod, there is no need for IS, so I'll just continue turning it off.

When I get some time, I will try some flash shots. If I learn anything interesting, I'll post it.

Dan

paddler4

Phil,

tried some flash shots, handheld, with IS on. 1/200, f/14. No problems. But then again, the short flash duration would take care of most hand motion.

Dan

Ekaj

It's interesting that your "findings" contradict the makes of the lenses themselves. No offense, but I will trust every other statement that VR/IS is next to worthless for macro use.

paddler4

Ekaj,

I don't know whether you intended it, but the quotation marks make for a nasty tone. In any event, you have your math and facts wrong. What does NOT help at macro distances is VR/IS that compensates for angular displacement. What SHOULD help at macro distances is IS that compensates for vertical displacement parallel to the focal plane, which only Canon's hybrid IS has. Just do the geometry and you can see this. Canon makes precisely this claim--only for this one lens--and has charts posted showing how it works. Several systematic reviews have confirmed the effect, although if I recall, some say the effect is a little less than the 2 stops that Canon claims. I just wanted to post a few initial examples to make it concrete. So, read twice, type once.