10D Autofocus

[Panza]

I guess this has been dabated abit around here but I tried a search and found no info.

How much is the Autofocus improved on the 10D compared to the D30 ?
Especially in low light.
and does the active focus squares in the 10D in the viewfinder turn red like on the Eos5 ?

[Sendide]

Have no Idea on the D60, but switching to digital first time with 10D, I was kind of desapointed.... , front focusing problems and not very accurate in low light, needs flash focusing light . but still, 10D remains a wonderful camera in a lot of other criteria.
god luck

khalid

[leony]

The focusing in low light is slightly improved. As far as disappointments, there was some hearsay about front/back focusing problems - more talk than problems. There were some cameras with a problem - about 10% of the "Mine has a problem" actually did. As was explained previously, the reason is that most "problem" people did not have experience with AF before and the ones that did did not know how to work it right.

Accuracy in 10D is just as good as that in EOS-3, Elan, Rebel, EOS-1 - virtually any EOS body of modern design. What is easier is looking at things 100% blown up.

When you look at a file from 10D in Photoshop with 100% magnification, you're actually looking at an equivalent (from 35mm film) of a print 11x17 inches, at a distance of 15 inches (if not less). Hardly a "normal" viewing distance for a print of that size. The true viewing distance is twice the diagonal or 40 inches. At that, 10D is not a P&S camera, and wasn't built to be one. Don't expect 100% great results from files right out of the camera - they are not processes in camera much to give you the flexibility afterwards. Even JPEG is excellent, provided you know what you're doing and not trying to salvage images underexposed by 2 stops. If you are, you should be punished by a mandatory 30 day sentence of Pentax K1000 and minimum of 5 rolls of T-Max 400.

If you're in low light, ANY AF camera needs light to focus. If you don't like "flash light" get a hot-shoe flash that will give you a red AF assist beam.

PS: All spelling errors due to line noise.

[Panza]

It's not the precision that bother me.
It's that the camera seems almost unable to find a focus point at all when there is little light.
I take it the 10D isn't greatly improved then..

[Pekka]

The focusing in low light is slightly improved. As far as disappointments, there was some hearsay about front/back focusing problems - more talk than problems. There were some cameras with a problem - about 10% of the "Mine has a problem" actually did. As was explained previously, the reason is that most "problem" people did not have experience with AF before and the ones that did did not know how to work it right.

Accuracy in 10D is just as good as that in EOS-3, Elan, Rebel, EOS-1 - virtually any EOS body of modern design.

Not true. 10D has 1 DOF AF and 1D series has 1/3 DOF accuracy. This is more significant the wider you go and if your lenses are slow.

"1 DOF" means that the focus lock point lays within DOF of fully open aperture of the lens used. "1/3 DOF" means it lays at least in 1/3 of the DOF of fully open aperture of the lens used. 1 DOF AF can get it exactly right (especially with good light) - but the worst case acceptable by specs is that it also can get it to the very edge of DOF.

Also, the 7 points in 10D are too widely spaced for practically keeping subject in focus, while 45-point systems will let you compose quite freely and always keep subject in focus.

10D is a great camera and I've got great results with it but it is still a mediocre product in AF area. This does not mean you can't get good and sharp photos with it, it just means you'll need take more care about focusing and use faster lenses.

[BobbyC]

I think it is improved over the D30, not by a lot, but it is. When I shot weddings with my ElanII, I had the lense hunt all the time in low light trying to lock on and my D30 did about the same. The 10D is slightly better and I don't do but a couple of weddings a year anymore (if that much) so it's not been an issue. I spent many years with a manual focus camera and as I have always done with AF cameras (none of them work all the time) is switch to that little letter M on the lense. The active focus point does light up. I typically only use the middle sensor unless I'm vertical in the studio, then I'll use the right one (top when vertical).

BTW, I've not seen the front/back focus problem at all.

[leony]

10D is a great camera and I've got great results with it but it is still a mediocre product in AF area. This does not mean you can't get good and sharp photos with it, it just means you'll need take more care about focusing and use faster lenses

That 1 DOF, 1/3 DOF doesn't make much sence to me. Not to say that it's false, but I don't know how it works... What I know for a fact is this:

AF system focuses until it acheives max. contrast between different parts in the "view" of an AF sensor (those square boxes in the viewfinder). Max contrast is Max contrast - no matter how you slice it. I don't exactly know how the camera knows when the max contrast is within 1/3 DOF or 1 DOF, but then I don't work for Canon R&D either.

About fast lenses, you're not quite right with 10D though you heard stuff sort of correctly. Pro bodies (EOS-1, EOS-3 and their D products) have special "improved" sensors that work in low light. THis is acheived by having 2 modes for the sensor. Fast lens mode - when you use 2.8 or faster lens. SLow mode when you use 4.0 or slower lens. THe camera actualy uses the lens info to figure out focusing somewhere in the algorithm, but it's not DOF. 10D uses "cheaper" slow-lens sensors, so in adequate light, AF will work the same on 10D whether you use fast or slow lenses.

I've seen 100Mb files from drum-scanned chromes shot with EOS-1N, and they don't quite look as sharp as my 10D files - more pixels, yes, but not sharper. And I know it's not the Imacon scanner's fault.

2: Limit of AF sensors:
The f/8 limit makes the EOS 3 and 1V considerably more useful with long telephoto lenses and teleconverters than the EOS 1 and 1N. Long telephotos with TCs (slower than f/5.6) wouldn’t autofocus with the older models. The sensor limit of the cross sensors is the f-stop at which the high-precision autofocus sensors switch from cross sensor mode to linear sensor mode. (ie: the cross sensors in these cameras require fast lenses to work in high-precision mode)

The 3 and 1V have 45 autofocus sensors, all of which can detect horizontal and diagonal lines at normal precision when used with a lens of f/8 or faster. However they also have 7 cross sensors capable of high-precision focussing. Of this group of 7, the central sensor retains high precision down to f/4 or faster and the remaining 6 cross sensors require f/2.8 or faster for high precision.

The 1N’s single cross sensor required f/2.8 or faster lenses for high precision and reverts to normal precision linear sensing with slower lenses. The camera’s remaining four linear sensors require f/5.6 or faster.

The EOS 1’s single sensor was a cross sensor which provided high precision with f/2.8 or faster lenses and normal precision linear sensing with f/5.6 or faster lenses.
From: http://photonotes.org/reviews/1-1N-3-1V/#2

[Pekka]

None of the 10D 7 AF points are high-precision (HP) at any aperture. This is confirmed by Chuck Westfall.

[Sendide]

for a fair comparison,
does anyone have any idea how big should a digital picture appear in a monitor to be compared to a film picture print in term of quality?
so many times we read people comparing film camera's and digital's quality. As mentioned above, in digital world, we get the reflex to enlarge and crop, till we're ... well, not satisfied :x , but do we really compare with the right enlargement of film camera's prints?
Or...how limiting are the scanner dpi's if we wnat to scane prints and compare them to digital ones on the screen?
regards
khalid