Canon 20d focus issues

danman

There has been a lot of talk recently on dpreview about canon 20d and focus issues (e.g. see this thread), so I decided to run a test of mine last night while it was quiet at work. I would be grateful of your thoughts on my results.

One of the main problems that others have identified is one of incorrect focusing - in particular forward focusing.

I have tested 2 lenses - the Sigma 18-50 f2.8 EX and the Canon EF 85mm f1.8 USM. The results of my 45 degree focus test can be seen on this page.

I am pleased that the Canon prime looks ok throughout the range from 1.8, but concerned that the Sigma seems to forward focus at f2.8 but is fine from f3.2.

Do you think that I should send the Sigma back or is this acceptable? I'm hoping that the results with the Canon prime mean that the 20d itself is OK - what d'ya reckon!?

Cheers,
Danny.

photocat

I am pleased that the Canon prime looks ok throughout the range from 1.8, but concerned that the Sigma seems to forward focus at f2.8 but is fine from f3.2.

What is forward focus? I am a beginner and have no clue of what you are talking about. How about some explanation forward focus for dummies?
Pretty please with cherries on top???

danman

Sorry - it was new to me until a few days ago! I think it means that the lens/camera focuses a distance closer to the camera than where you actually pointed the focusing sensor (I pointed it at the X). Camera was on a tripod.

Danny

photocat wrote:

I am pleased that the Canon prime looks ok throughout the range from 1.8, but concerned that the Sigma seems to forward focus at f2.8 but is fine from f3.2.

What is forward focus? I am a beginner and have no clue of what you are talking about. How about some explanation forward focus for dummies?
Pretty please with cherries on top???

photocat

danman wrote:

Sorry - it was new to me until a few days ago! I think it means that the lens/camera focuses a distance closer to the camera than where you actually pointed the focusing sensor (I pointed it at the X). Camera was on a tripod.

Danny

Could it have something to do with wrap around? I have a function on my camera that I have no clue off, but is about wrap around with focus area if I recall well.
Your pics are a phenomene that I never thought of before. I tend to think it is the lens and not your camera.

ian408

If you use the Canon software, you can see the focus point the
camera selected. Is the focus point on the X?

danman

Yes!

ian408 wrote:

If you use the Canon software, you can see the focus point the
camera selected. Is the focus point on the X?

Michiel de Brieder

Dan, I'm curious
As to the size of your crops. The 20D's focus points represent a much larger focus area. So it could be that the focus point you were using, was snapping on the gridline instead of on the x-mark because it is also very contrasty. Have you repeated the test several times with the same result @ f/2.8? If that is so, then you would indeed have a bit of front focus.
As a side note, what I'm always wondering of... I haven't tested my 20D on focus, I just shot some pics in the way I always do. Now it could be that it has a little front-focus or back-focus, I don't know, but in real life I'm not troubled by it (yet)...
I hope this post is of some use to you
Cheers!

damonff

I've been reading a lot of bad stuff about the 20D...

Michiel de Brieder

damonff wrote:

I've been reading a lot of bad stuff about the 20D...

I've been having a very good time with my 20D....
20D =

Steve Cavigliano

Michiel de Brieder wrote:

I've been having a very good time with my 20D....
20D =

I've got to agree with Michiel. I also been having a super time with my 20D. I've used 7, or 8, lenses, and shot close to 2000 pics, and the images always seem to be in focus. Well, other than the ones I know were caused by operator error
No "lock up's" and no Err99's from/when changing lenses (and I do alot of lens changing). I had one Err99 when reviewing 4gb's worth of pics, in the camera. This is a really good camera and I wouldn't hesitate to recommend very highly.


Hey Danny,
Were you seeing issues with your Sigma, before you did this test? I mean, were you seeing oof images, or having any focusing problems with it? If not, I wouldn't worry too much about your results. If you were having issues, then by all means send the lens in to Sigma. These tests can be misleading. As alrerady mentioned, the focus points are bigger than they appear inthe VF. So higher contrast areas can cause the camera to pick the "wrong" in-focus area. My Sony's do this too. But, if you are seeing focus problems when shooting, with this lens only, then the lens may be at fault.

Anyhow, good luck and I hope all works out well

Steve

lynnma

This is'nt really related to the focus question but I didn't want to start a new thread really.. it is related to the EOS-20D tho.. is the CMOS on the Drebel (which I have) vastly different to the CMOS in the 20D?? and if so, anyone have any ballpark percentage of how much? and... why is it different..

Michiel de Brieder

Lynn, about the CMOS
Here's a quote from dpreview

Phil Askey wrote:

8.2 megapixel CMOS sensor
The first and probably most significant change is an all new 8.2 megapixel CMOS sensor which has the same dimensions (albeit for half a mm) as the EOS 10D's six megapixel CMOS sensor. Canon also claims improved sensitivity thanks to new microlenses and a new on-chip 'noise elimination circuit', frame rates are up as this sensor has a 4-channel read-out (double the EOS 10D).

Now the 300D uses the same sensor as the 10D, so the difference is the same. In addition to having an improved sensor, the 20D also has a new imaging processer DIGIC-II. It is speculated that the new sensor also has much influence on the improvements of the 20D over the 10D/300D.
Hope this helps somewhat, or confuses some more
Ciao!

PerezDesignGroup

I've noticed soft-focus issues with my 300d. But it's just the camera. It seems everything I shoot needs a USM of 150-1-0. I know it's not the lens because I throw the 50mm 1.8 on my EOS 35mm and I got crisp shots like the one below (which is completely unprocessed).


full size

Michiel de Brieder

soft-focus vs FF/BF

PerezDesignGroup wrote:

I've noticed soft-focus issues with my 300d.

Hello Perez,

There is a difference between front focus and back focus vs soft focus. Canon's DSLRs are very much concentrated on user flexability. Canons DSLR shots really sharpen up well with USM. So you have a choice of soft vs sharp (also, the in-camera paramters can take care of sharper picture straight from the cam) The photos may seem soft at first, but with sharpening...... I really enjoy the prints that can be had with a DSLR (and I also shoot analogue from time to time). At 20x30cm there is really no difference (that was with a D30, I've still got to print a 20D shot )

danman

Focus issues
Thanks Michiel at al. I haven't done tons of repeats and in fact just getting on with it and using the camera has not really shown up any significant problems in practice. At the start I think I was shooting with too slow shutter speed. I think 1/1.5f is a better guide for digital SLRs - at least in my hands!

I am a bit worried about the fact that that particular lens is the only one which showed apparent focus problems and Sigma have said they'll look at it. Initial testing with a 70-200 and now 2 other Canon ones was fine under similar situation.

Cheers,
Danny.

Michiel de Brieder wrote:

As to the size of your crops. The 20D's focus points represent a much larger focus area. So it could be that the focus point you were using, was snapping on the gridline instead of on the x-mark because it is also very contrasty. Have you repeated the test several times with the same result @ f/2.8? If that is so, then you would indeed have a bit of front focus.
As a side note, what I'm always wondering of... I haven't tested my 20D on focus, I just shot some pics in the way I always do. Now it could be that it has a little front-focus or back-focus, I don't know, but in real life I'm not troubled by it (yet)...
I hope this post is of some use to you
Cheers!

Stan

As posted in the Camera shake thread, I think it has greater relevance in this thread though, If anyone is interested, I found the link to the depth of field calculation programme Here