manual vs auto focus

[transcend]

I have a 17-40L on my 10d, recently i have been noticing images seem sort of soft. I ran a few quick very unscientific tests, and it seems that the AF almost 100% of the time, focuses a bit more softly then by hand.

If i manually focus i can get it razor sharp, otherwise i cannot.

This was on a tripod, 180/sec, iso 200, flash on and flash off. No chance of hand shake (which i normally attribute to this, but which doesnt appear to be the case as it happens at all shutter speeds.

Is this normal and am i just paranoid? Or should i look into having it calibrated?

[FlipsidE]

The camera and/or lens could be front focusing or back focusing. Unfortunately, I don't have the links to point you toward the tests, but there are front focus and back focus test available.

FlipsidE

[robertwgross]

My camera and long lens were having a "softness" problem starting just a few days ago. Finally, I looked at it and the viewfinder eyepiece diopter setting had gotten tweaked. As soon as I re-tweaked it, the problem went away.

---Bob Gross---

[EXA1a]

I ran a few quick very unscientific tests, and it seems that the AF almost 100% of the time, focuses a bit more softly then by hand.
Please don't mix up the terms "soft" and "out of focus". You are talking about a potential misfocussing problem, right?

[transcend]

I ran a few quick very unscientific tests, and it seems that the AF almost 100% of the time, focuses a bit more softly then by hand.
Please don't mix up the terms "soft" and "out of focus". You are talking about a potential misfocussing problem, right?

Possibly a misfocus problem, possibly just the lens being soft and completely normal. This is an L lens however, i expected much better performance then I am getting, especially when at middle apertures. Wide open i expected it to be a bit soft.

I have front abd back focus tests now, so i will give it a shot this afternoon

[EXA1a]

I ran a few quick very unscientific tests, and it seems that the AF almost 100% of the time, focuses a bit more softly then by hand.
Please don't mix up the terms "soft" and "out of focus". You are talking about a potential misfocussing problem, right?

Possibly a misfocus problem, possibly just the lens being soft and completely normal. This is an L lens however, i expected much better performance then I am getting, especially when at middle apertures. Wide open i expected it to be a bit soft.

I have front abd back focus tests now, so i will give it a shot this afternoon
In your original post you found differences between MF and AF shots, and that's misfocusing for sure.

[jalafer]

The camera and/or lens could be front focusing or back focusing. Unfortunately, I don't have the links to point you toward the tests, but there are front focus and back focus test available.

FlipsidE

A very simple focusing test.
Put your camera on a solid tripod in front of a wall.
Put a printed sheet on the wall (a wall with painted paper would be Ok).
Make one picture in AF, then turn the camera on MF, focus manually and make the second picture.
Check the results in your computer screen, if MF is better than AF, then your camera is frontfocussing. Take into account that you can have this probelm with one lens and not have ot with another.

[Scottes]

if MF is better than AF, then your camera is frontfocussing.

But if AF is better this test means nothing, because the diopter could be off.

[PacAce]

The camera and/or lens could be front focusing or back focusing. Unfortunately, I don't have the links to point you toward the tests, but there are front focus and back focus test available.

FlipsidE

A very simple focusing test.
Put your camera on a solid tripod in front of a wall.
Put a printed sheet on the wall (a wall with painted paper would be Ok).
Make one picture in AF, then turn the camera on MF, focus manually and make the second picture.
Check the results in your computer screen, if MF is better than AF, then your camera is frontfocussing. Take into account that you can have this probelm with one lens and not have ot with another.

You might do better with this test if you set the camera and tripod so that the camera is facing the wall at a 45 degree angle. Then set the lens to AF and shoot a couple of frames. This way, you'll be able to better see if the camera is back or front focusing or if the lens is just soft to begin with.

[jalafer]

You might do better with this test if you set the camera and tripod so that the camera is facing the wall at a 45 degree angle. Then set the lens to AF and shoot a couple of frames. This way, you'll be able to better see if the camera is back or front focusing or if the lens is just soft to begin with.[/quote]


In my opinion is better if you face the wall direct. If your lens is front focussing (as it happened with an old 70-210 in my case) you will realise inmediately because the MF picture will be well focussed.
Of course if you do it at 45º and put a mark on the paper to be foussing at, the you can see is your camera is front or back focussing.

[transcend]

You might do better with this test if you set the camera and tripod so that the camera is facing the wall at a 45 degree angle. Then set the lens to AF and shoot a couple of frames. This way, you'll be able to better see if the camera is back or front focusing or if the lens is just soft to begin with.


In my opinion is better if you face the wall direct. If your lens is front focussing (as it happened with an old 70-210 in my case) you will realise inmediately because the MF picture will be well focussed.
Of course if you do it at 45º and put a mark on the paper to be foussing at, the you can see is your camera is front or back focussing.[/quote]

I will be using the focusing chart i printed out online, and the camera at a 45 deg angle. Trying to see what it is doing. If it is front focusing, i take it it is a simple enough calibration at a canon service center? Any idea on cost?

[lmelendez]

If you suspect that your camera is back focusing... what do you do? you just use MF all the time :? ?

How can you calibrate that?

[robertwgross]

If you suspect that your camera is back focusing... what do you do? you just use MF all the time :? ?

How can you calibrate that?

You can't, but Canon can.

A focus problem might be related to the camera body being out of spec, or it might be related to the lens being out of spec. Canon can test each piece independently, correct as necessary, and then put them together to make sure the combination is still right.

---Bob Gross---

[transcend]

If you suspect that your camera is back focusing... what do you do? you just use MF all the time :? ?

How can you calibrate that?

You can't, but Canon can.

A focus problem might be related to the camera body being out of spec, or it might be related to the lens being out of spec. Canon can test each piece independently, correct as necessary, and then put them together to make sure the combination is still right.

---Bob Gross---

yup, luckily there is a service center 20 mins form my house. I will take both in on monday.

[transcend]

here is the result of the test.

Taken at 45 degs (as per marking on tripod, may be slightly off).
40mm, f 4.5, 3 foot distance aproximately. On 17-40L

I would assume that at 40mm/4.5 DOF would be pretty decent, would being off of 45deg even a tiny amount make the test worthless? It looks like it is back focusing a fair amount.

Also, as for the diopter, i just tested it and my manual focusing indoor in not so great light is not good. WAY off for now, i wear a fairly bizarre Rx pair of glasses tho (oddly shaped eyeball due to concussions). I matched it up so that AF looks sharp, when looking through it, but obviously this wont work if AF is off.,

http://www.transcendmagazine.com/random_pics/focus_test.jpg

this shot is from straight on

http://www.transcendmagazine.com/random_pics/focus_test_straight.jpg

[RDKirk]

You might do better with this test if you set the camera and tripod so that the camera is facing the wall at a 45 degree angle. Then set the lens to AF and shoot a couple of frames. This way, you'll be able to better see if the camera is back or front focusing or if the lens is just soft to begin with.

Do the head-on test first. There are a number of ways the 45-degree test can steer you wrong. If the camera passes the head-on test, you need not even dip into treacherous waters.

[PacAce]

here is the result of the test.

Taken at 45 degs (as per marking on tripod, may be slightly off).
40mm, f 4.5, 3 foot distance aproximately. On 17-40L

I would assume that at 40mm/4.5 DOF would be pretty decent, would being off of 45deg even a tiny amount make the test worthless? It looks like it is back focusing a fair amount.

Also, as for the diopter, i just tested it and my manual focusing indoor in not so great light is not good. WAY off for now, i wear a fairly bizarre Rx pair of glasses tho (oddly shaped eyeball due to concussions). I matched it up so that AF looks sharp, when looking through it, but obviously this wont work if AF is off.,

http://www.transcendmagazine.com/random_pics/focus_test.jpg

this shot is from straight on

http://www.transcendmagazine.com/random_pics/focus_test_straight.jpg

To be honest with you, I don't think your lens is off at all. Your "target" line seems to be pretty in focus to me.

[RDKirk]

I would assume that at 40mm/4.5 DOF would be pretty decent, would being off of 45deg even a tiny amount make the test worthless? It looks like it is back focusing a fair amount.

First, judge all images as 8x10 prints of the full image. Canon's tolerances are based on the DoF at 8x10 hardcopy enlargement, no larger. Canon has published at least two documents urging users not to judge AF at high enlargements on the monitor.

Second, remember that the camera may focus on ANY contrast in within the area seen by the sensor (at random, and not necessarily consistently). On the 20D, that area is NINE times the size of the marked central square, so it includes many more lines than the one you focused on. The camera may lock onto any of them and still be within Canon's tolerances.

Third, the camera may focus anywhere within 1/3 of the DoF around the intended plane of focus (using the central AF sensor at f2.8 on the 20D, anywhere within the DoF in in other cases). So in that way, it may show "front focusing" or "back focusing," yet be within Canon's tolerances.

Also, as for the diopter, i just tested it and my manual focusing indoor in not so great light is not good. WAY off for now, i wear a fairly bizarre Rx pair of glasses tho (oddly shaped eyeball due to concussions). I matched it up so that AF looks sharp, when looking through it, but obviously this wont work if AF is off.,


I've pretty severaly myopic myself, and now I'm also presbyopic. The camera viewfinder is designed to place a virtual image about 1 meter in front of your eyes--your eyes are actually trying to focus at that distance, so you should wear whatever glasses you need to focus at 1 meter.

To adjust the diopter settings:

Remove the lens and point the camera toward a large, evenly lit area (a blank, light colored wall will do). Look through the viewfinder at the AF marks. That's what you want in sharp focus.

Turn the wheel all the way to an extreme so the marks are strongly defocused. Then turn the wheel QUICKLY to sharp focus. Don't try to focus slowely and carefully, or you eye will adapt to an out of focus condition. You may have to do it several times, going back to a strongly OOF extreme, then going quickly to correct focus.