Canon Rebel XT Autofocus Question

Disclaimer: I've done a site search and a google search and haven't found what

I'm looking for. Therefore, I ask for the learned ones' help.

<p>

I've been shooting with a Rebel XT for about 3 years now. I haven't really

given much thought to AF results over that period; however, back in October I

went to a motorcycle race at Road Atlanta. While shooting, I was chimping a bit

along the way, and noticed that a I was getting more sharp blue and purple bikes

than I was yellow and white (among other light and dark colors). When I got

back and dumped the photos on my computer, I confirmed this suspicion. While

some of the photos may have been blurry from panning/motion blur rather than

failure of the AF, I think the ratio of sharp dark subjects to sharp light

subjects confirms my suspicion.

<p>

My opinion is that the Rebel XT is getting better AF on darker subjects (bikes)

than lighter ones as a result of the increased contrast against the gray track.

<p>

I'm curious about a deeper technical understanding of this AF behavior, if, in

fact, my opinion is correct. If I'm wrong, let me know and give me your further

thoughts.

<p>

In case you're interested, visit my gallery from the <a

href="http://ericreagan.smugmug.com/gallery/3724894/1/213757166#213757166">2007

WERA Grand National Finals</a> for the subject samples discussed above. Note,

however, that the gallery does not include the numerous out of focus shots

discussed above.

<p>

I await your thoughts.

<p>

Thanks,

<p>

Eric

Without looking at ALL your photos... they looked pretty even as far as which percentage of yellow vs blue are sharp (ie. 3 (yellow) looks pretty good... 14 (dark) looks a bit soft).

 

If you truly are seeing this dilemma... my guess would be that the lighter bikes are less contrasty than the dark ones. I am not sure where on the bik you focused also.

 

All these things factor into it. It is going to be hard to tell from 1 event... you will learn from a few more similar situations especially with different lighting situations.

If you follow the links from <a href="http://www.zen20934.zen.co.uk/photography/Canon%20AF%20System.htm#Testing">Lester Wareham's Focus Test</a> page (at the top) you will find a description of how the Canon AF system works. This has the most detail I have seen on the subject. Apparently it's also in Canon's "Lenswork III" document.

Taking into account the camera viewpoint; angle of subject to the camera; the background colour and lighting and the foreground colour and lighting, and assuming that doesn`t change much throughout the gallery: on my understanding of how the Canon AF functions there is substantial evidence to assume White and Yellow bikes would be more problematic for the AF than those painted Purple and Blue.

 

WW

Are you using AI Servo focus with centre point only active? I suspect that the key point is that with a darker bike there tends to be a high contrast boundary element for focus to latch on to - e.g. the number decal in white, or a specular reflection from the handlebar. I doubt the background has much to do with it, as it will not be covered by the focus point most of the time unless you mess up your panning.

>>> I doubt the background has much to do with it, as it will not be covered by the focus point most of the time unless you mess up your panning. <<< (MU)

 

Yes, I though about that, for a long time (and I assumed that only the centre point was being used), and I expect it is as you wrote when the centre of the red square was on the bike 100% of the time.

 

 

But then I thought:

 

1. About how (much of) the focussing action is NOT limited to inside those little red squares and then;

 

2. I thought with how often, at the beginning of a pan, we are slower than the action, and catch up to it.

 

Hence, even with `even` panning, there would have been a point in time where the centre square was on the background (or foreground) and then it snapped into focus on the darker bike.

 

Then with AI Servo . . . handle bars and decals etc.

 

Thinking further upon it, I think the real answer might be in these two parts, but both are dependant upon the fact of the bike being dark.

 

WW

Aaron: Sorry for the misunderstanding, but I weeded out the blurry shots before uploading them. I meant for the gallery to be a sampling of the subject matter, not specific examples of the blurry shots.

<p>

Glen: Thanks for the resource references. This is the type of info that I was searching for.

<p>

Mark U: I was using AI Servo. I varied the focus points from middle to one left or right of middle depending on where I was at on the track.

<p>

William: I began panning well before releasing the shutter. I feel like was pretty steady with the focus points on the shot. However, maybe you are on to something.

<p>

Thanks for the replies guys.

<p>

Eric

http://www.ausphotography.net.au/forum/archive/index.php?t-1799.html

 

http://doug.kerr.home.att.net/pumpkin/Split_Prism.pdf

 

http://doug.kerr.home.att.net/pumpkin/AF_accuracy.pdf