Hi Bart,
> then I never meant it to be a serious test. The thing is that I was just
> worried after I had played a little bit with my 'new' 5 and my 'old' 300.
I wouldn't expect you or anyone on the list to undertake a rigorous equipment test;
I recall my own informal tests not so long ago regarding the effect of extension
tubes on the sensitivity of different focus sensors with lenses of different max
apertures (prompted by some questions from Julian L.) that were way
less-than-conclusive. It just seemed that you were comparing different settings
that weren't entirely . . . um, the word flew out of my head just as I was going to
type it . . . commensurate, and then drawing a somewhat misleading conclusion that
choosing a single focus point had a negative impact on AF performance.
That said, even though the EOS 5 was positioned further up the product line than is
the EOS 300, the 300 is a much newer design and almost certainly benefits from the
technical advances that first appeared in the EOS 3, so all else being equal, I
would expect the 300 to perform as well as or better than the 5 in terms of AF and
AE. Would you not agree? I didn't do a direct comparison between my 30 and my 5
during the brief time I owned both, but my subjective impression is that the AF of
the 30 is faster than that of the 5, and more accurate (less hunting). This is just
an impression, though.
> I found that the 5 with ECF would hunt a lot more
> for focus than the 300 with all focusing points selected. In
> some cases, the 5 was not able to focus at all, where the 300
> at no time during my little unscientific test failed in this respect.
Wasn't the 5 the first camera with ECF? I've had two bodies with ECF but I've never
really used it so I'm not familiar with its characteristics, but isn't its
performance highly dependent on how well it tracks an individual's eye, and how well
the individual is able to work within the parameters ECF requires for best
performance? In other words, is not ECF performance highly variable from user to
user? Assuming it is, these two factors in combination (age of the system and user
variability) could go a long way toward explaining why your 5 does less well when
set to ECF compared to your 300 without ECF. As I understand it, Canon markets ECF
as increasing overall working speed by reducing the effects of user reaction time
(move your eye instead of your head), not by increasing the speed of AF operation
itself. While (again) not by any means an observation grounded in measurements or
other research, I can't recall seeing any evidence that AF with ECF is faster or
more accurate than selecting one or all of the focus points using the camera's
controls.
> I used the same lens (28-135 IS) on both cameras
> and I tried to focus on the same objects. So what I meant
> by 'AF performance' was 'the ability to lock into focus'.
If you have the opportunity it might be interesting to try the same test with a lens
with a maximum aperture of f/2.8 or larger--the speed of a lens can have a
significant impact on AF performance.
> . . . the 300 will have more trouble locking into focus if I . . . manually select
> a focusing point on the 300.
This is the interesting result, because it seems to go against everything we
understand about Canon's AF system. So interesting, in fact, that I'm going to do a
little informal test myself. Be right back. . . .
LOL! I had to dig out the manual for my EOS 30 (Elan 7) to figure out how to make
all focus points active. . . . Anyway, with the 30, and my slowest lens (the
28-105/3.5~4.5) set to 50 mm (max aperture, f/4), in my living room on an overcast
day (no direct sunlight, I found that focusing was consistently faster with the
center point selected than with all points active. The targets were 8-10 feet away,
and included several low-contrast subjects: black audio equipment on a black rack, a
bookcase in the corner, the tv screen (off). The only subject on which the "all"
setting appeared to have an advantage was the black cloth grille of one of my
speakers, which offered insufficient texture/contrast for the center focus point to
get a lock; however, the advantage was illusory because although with all points
selected the camera did focus, it grabbed something off to the side and behind the
plane of the speaker. In other words, in neither mode could the camera focus on the
speaker grille; with one point selected the camera simply didn't focus, and with all
points selected it failed to focus on the object in the center of the frame. I
noticed this in other instances as well. Whereas the center point alone focused on
whatever part of a piece of black audio equipment it was aimed at, with all points
selected the camera chose whatever in the frame had the highest contrast (something
noted by another list member), not at what was in the center of the frame.
So. It seems to me what you should be looking for is whether, when you achieve focus
with all points selected, the camera focused on what you wanted it to, or just on
whatever in the frame it could get a lock on. The fact that with all points selected
both cameras will focus on *something* doesn't mean they've focused on the "right"
thing, and therefore isn't necessarily an indication of better AF performance even
when one or both cameras fail to achieve focus at all in other modes.
> Does anybody have any ideas how I could make this test a bit more scientific?
There's probably not a lot you can do without test equipment, but at the least you
can be careful to compare each camera in each mode, using the same targets. In my
test I took advantage of my lens's full-time manual focusing to defocus the lens
before each "shot." I suspect your actual test was more like this than was apparent
in your original post; what drew my initial response was the way you said (and I
exaggerate for effect) "the focusing of the 5 with all points selected is as good as
the 300 with all points selected but with just one point selected the 300 is only as
good as the 5 using ECF; why is focusing ability reduced if I manually choose a
focusing point?", thereby suggesting that AF performance is worse with a single
point selected compared to having all points active, which probably is not actually
the case.
Craig
*
****
*******
***********************************************************
* For list instructions, including unsubscribe, see:
* http://www.a1.nl/phomepag/markerink/eos_list.htm
***********************************************************
