Home · Register · Join Upload & Sell

Moderated by: Fred Miranda

New fredmiranda.com Mobile Site

New Feature: SMS Notification alert

New Feature: Buy & Sell Watchlist

FM Forums | Canon Forum | Join Upload & Sell

2 end

Archive 2005 · Imatest: How steady are you?

gfiksel
Offline
• • • •
Upload & Sell: Off

p.1 #1 · p.1 #1 · Imatest: How steady are you?

So, how steady are you? What is your handheld limit?

I used an excellent program, IMATEST (w.w.w.imatest.com) by Norman Koren to find an answer to that for myself. Not that I haven’t known it already but I wanted to express it in a somewhat quantitieve way. For those unfamiliar with IMATEST, it’s a program that helps to determine the resolution of a camera and lens. Basically, you take a picture of a perfectly straight edge and the program analyzes the images and determines how blurred the edge is.

Well, the blurriness comes not only from a finite resolution but also from vibration of the camera if you handhold it. Therefore, it’s a useful tool for determining how shaky your hands are.

So I took a base shot, with the camera on a tripod and then took the camera off the tripod and continued shooting increasing the exposure time from 1/1000 sec to 1 sec. To cover that huge range I had to change the lighting level (I had a few halogen lights that I could turn on and off), the ISO, and the aperture. The ISO was changing from 100 to 400, thanks to a low noise level of 20D, and I was trying to keep the aperture between f/4.0 and f/8. I used prime lenses EF 50mm/1.4, 85mm/1.8 and 135mm/2.0L . The range of apertures from f/4.0 to f/8.0 is a sweet spot for them; the lenses resolution reaches its maximum and does not change much within this limit.

So here's how the resolution changes with the exposure time for the 85mm/1.8 lens (top plot). Never mind what the Y axis scale is, it’s not relevant for my purpose. But you can see that for short exposures the resolution is practically the same as the for the base shot on the tripod but it starts plummeting after 1/60 sec. So this is my handheld limit for f=85mm lens – 1/60 sec.

Then I went through the same procedure for the 50mm and for 135mm lenses. The curves looked almost like the one for the 85mm only the transition happens at different exposures. For the 50mm it’s 1/40 sec and for 135mm it’s 1/100 sec. That’s what the bottom plot shows. I also plotted the famous 1/f rule. Well, I am slightly better. What about you?

I know, I know I should go out and take some pictures. Maybe now I will

Apr 08, 2005 at 07:02 PM

Ken Tanaka
Offline
• • •
Upload & Sell: Off

p.1 #2 · p.1 #2 · Imatest: How steady are you?

Very interesting, Gennady. I've never seen such an analysis. It looks like a fun experiment.

There was a similar thread on a Leica forum early this year. Of course shooting with a Leica M is a very different game than with, say, a Canon 1 body. But there were some folks who could get a good M shot at shutters as long as 1/2 sec.

I can pretty consistently get a usable shot up to a shutter speed of 1/10 sec as long as I'm relaxed, well-balanced and stationary. Of course this is also a function of the lens' focal length, the camera/lens weight, etc.

It's worth noting that some of the most famous and memorable photographs of all time are far from razor-sharp.

Apr 08, 2005 at 07:17 PM

Monito
Offline
• • • • • •
Upload & Sell: Off

p.1 #3 · p.1 #3 · Imatest: How steady are you?

Excellent testing. Note the up and down nature of the resolution in the low speed handheld area. This is due to the statistical nature of shakiness. If you were to take several shots at each shutter speed and average the resolution, you would probably get a smoother curve.

Apr 08, 2005 at 07:18 PM

gfiksel
Offline
• • • •
Upload & Sell: Off

p.1 #4 · p.1 #4 · Imatest: How steady are you?

Ken, thanks.

Monito, I agree. This scatter at low speeds also explains why one low speed handheld shot can be almost usable while the next one is total garbage

Apr 08, 2005 at 07:22 PM

Bobster2
Offline
• • • •
Upload & Sell: Off

p.1 #5 · p.1 #5 · Imatest: How steady are you?

I agree about taking several shots at each shutter speed, I think that's essential, but instead of taking an average, another possibility is to pick the best from each set.

In a "real world" situation, say when you're trying to handhold for a landscape and get the best possible quality, you will probably have the opportunity to take several shots and pick the best. So I think it's a good idea to do the testing the same way.

Apr 08, 2005 at 07:52 PM

mclaren20
Offline
• •
Upload & Sell: Off

p.1 #6 · p.1 #6 · Imatest: How steady are you?

With the 35/2 I have gotten shots off at 1/2+ seconds. If I'm relaxed and using my eos3, then its pretty easy. With all other lenses I can usually go about a stop under whatever the focal length rule is.

Apr 08, 2005 at 07:55 PM

Gary Petersen
Offline
• • • • •
Upload & Sell: Off

p.1 #7 · p.1 #7 · Imatest: How steady are you?

When everything goes just right I did this. It's a 4 second hand held exposure at the 28mm equivalent setting on my D7i.

Apr 08, 2005 at 08:51 PM

Monito
Offline
• • • • • •
Upload & Sell: Off

p.1 #8 · p.1 #8 · Imatest: How steady are you?

Bobster2 wrote:
I agree about taking several shots at each shutter speed, I think that's essential, but instead of taking an average, another possibility is to pick the best from each set. In a "real world" situation, say when you're trying to handhold for a landscape and get the best possible quality, you will probably have the opportunity to take several shots and pick the best. So I think it's a good idea to do the testing the same way.

Good point, but sometimes one gets only one shot, so there would really be three stats to track: the best of a set, the average or median (two separate stats, choose one), and the worst. By evaluating the worst of each set, you could tell what speed you would have to use to shoot in order to guarantee a good shot when you can only take one. By evaluating the average, you can see the "expected value" (statistical term) of the speed you can expect to get a good shot (say 50 percent of the time). By evaluating the best of each set you can find what speed you can shoot at when you have the luxury of shooting a set.

Apr 08, 2005 at 08:54 PM

gfiksel
Offline
• • • •
Upload & Sell: Off

p.1 #9 · p.1 #9 · Imatest: How steady are you?

Gary Petersen wrote:
When everything goes just right I did this. It's a 4 second hand held exposure at the 28mm equivalent setting on my D7i.

Unbelievable!!! It's 50 times slower than I would be able to.

Apr 08, 2005 at 08:58 PM

Canon 10D
Offline
• • • •
Upload & Sell: Off

p.1 #10 · p.1 #10 · Imatest: How steady are you?

Did u take into account the crop factor of the 20D when determining the f (mm) shown in your graph?

Apr 08, 2005 at 09:03 PM

Monito
Offline
• • • • • •
Upload & Sell: Off

p.1 #11 · p.1 #11 · Imatest: How steady are you?

Another thing about sets is that if you put the camera on continuous shoot mode, you often get a good shot in the second or third one. The first shot is too tense, but the body relaxes and a subsequent one is steadier.

Apr 08, 2005 at 09:04 PM

gfiksel
Offline
• • • •
Upload & Sell: Off

p.1 #12 · p.1 #12 · Imatest: How steady are you?

Canon 10D wrote:
Did u take into account the crop factor of the 20D when determining the f (mm) shown in your graph?

No, I took the lens f without crop factors.

Apr 08, 2005 at 09:06 PM

Bobster2
Offline
• • • •
Upload & Sell: Off

p.1 #13 · p.1 #13 · Imatest: How steady are you?

> there would really be three stats to track

Exactly! The best is to have three curves on the graph.

Apr 08, 2005 at 09:06 PM

gfiksel
Offline
• • • •
Upload & Sell: Off

p.1 #14 · p.1 #14 · Imatest: How steady are you?

Bobster2 wrote:
Exactly! The best is to have three curves on the graph.

Well, there goes my weekend

Apr 08, 2005 at 09:08 PM

Monito
Offline
• • • • • •
Upload & Sell: Off

p.1 #15 · p.1 #15 · Imatest: How steady are you?

It is a series of tests we should all undertake to really get a grip on our capabilities. I'll add it to my list

Apr 08, 2005 at 09:29 PM

gfiksel
Offline
• • • •
Upload & Sell: Off

p.1 #16 · p.1 #16 · Imatest: How steady are you?

Monito wrote:
Excellent testing. Note the up and down nature of the resolution in the low speed handheld area. This is due to the statistical nature of shakiness. If you were to take several shots at each shutter speed and average the resolution, you would probably get a smoother curve.

On the second thought, I think some of the scattering comes from the intricacies of the Imatest analysis.

For the lens resolution test the edge blur is close to gaussian so all the MTF transfer functions are smooth and it's easy to determine MTF10, MTF50, etc.

For the shakiness test the edge blur is more like linear and the MTF transfer functions have noticeable oscillations, hence the scatter.

Apr 08, 2005 at 10:12 PM

edtang
Offline
• • •
Upload & Sell: Off

p.1 #17 · p.1 #17 · Imatest: How steady are you?

Very interesting. Do you have any Image Stabliazation lenses that you can test with this program? I would be very interested to see actual data about IS lenses and how much sharpness you can gain from it.

Apr 08, 2005 at 10:15 PM

gfiksel
Offline
• • • •
Upload & Sell: Off

p.1 #18 · p.1 #18 · Imatest: How steady are you?

Sorry, I don't have IS lenses.

But you can download a trial version of Imatest and check it out. It's really a nice program. I routinely run all my new lenses through this test, so I would have a reproducible and unbiased basis for comparison.

Edited by gfiksel on Apr 09, 2005 at 07:30 AM GMT

Apr 08, 2005 at 10:22 PM

andrew_rs
Offline
• • •
Upload & Sell: Off

p.1 #19 · p.1 #19 · Imatest: How steady are you?

1/f is a myth. I can hand hold at any speed. Check out this example: 125mm at 1/8th sec.


	http://ella.slis.indiana.edu/~arsteven/pics/arbor2.jpg

Maybe I should lay off the coffee.. lol

Apr 08, 2005 at 11:17 PM

HinduG
Offline
• • •
Upload & Sell: Off

p.1 #20 · p.1 #20 · Imatest: How steady are you?

andrew_rs wrote:
1/f is a myth. I can hand hold at any speed. Check out this example: 125mm at 1/8th sec.

HAHA

Apr 08, 2005 at 11:39 PM

1 2 end
FM Forums \| Canon Forum \| Join Upload & Sell 1 2 end

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.