Autofocus Comparison

paul_di_biase · December 6, 1998

The following test results from Chasseur d'Images recently appeared on the Nikon Digest. You may make of them as you wish.

F5 13 of 13 photos in focus with 300/4, closest photo @ 6 meters

13 of 13 photos in focus with 300/2.8, closest photo @ 6 meters

N90s 7 of 7 photos in focus with 300/4, closest photo @ 9 meters

7 of 7 photos in focus with 300/2.8, closest photo @ 9 meters

EOS3 8 of 9 photos in focus with 300/4, closest photo @ 15 meters 10 of 11 photos in focus with 300/2.8, closest photo @ 13 meters

EOS1n 7 of 9 photos in focus with 300/4, closest photo @ 20 meters 8 of 10 photos in focus with 300/2.8, closest photo @ 17 meters

alan_gibson · December 6, 1998

Umm, is Nikon Digest possibly, by any chance, connected with the well-

known camera manufacturer? And they find that their cameras are

better than the competition? Well, there's a surprise.

louis_philippe_masse · December 6, 1998

I don't know about the Nikon digest but I have read the mentioned

article in Chasseur d'Images. Chasseur d'image is one of the most

trustable magazine in terms of independence of their tests. They test

only production items buyed anonymously in ordinary photo shops

(contrary to most mags who tests items "lended" by the

manifacturers). Also, they are not afraid to write a bad review of a

product (even for Canon or Nikon). So I don't think the possible

conflict of interest mentionned by Alan is an issue.

That being said, I don't know about the truthfullness of their

results. Results like this doesn't worth much in real life in my

opinion. Personnaly, I am a Nikon user and I doubt that the new EOS 3

AF is slower than my old F90x's (although I haven't tried the EOS 3).

Keep shooting!

L.P.

brad_hutcheson · December 7, 1998

Maybe you could post a little more about the test, if you have any other information. I was confused by the different focusing distances at first, but I assume this was a moving subject test, and the differences represent the distance of complete failure to focus on the moving suject befor it is too close for the lens to focus anyway. As I have said before, tests are nice for reference, but don't always hold up in the real world. This one is certainly hard to judge with the limited information though.

alan_gibson · December 7, 1998

I have also been told, in an e-mail, that Nikon Digest is independant

from the manufacturer. Fair enough, I'll accept the independance of

the report. I dare say the tests would be relevant to some people. I

use Nikon manual focus, and very rarely go as far as 300mm.

willio · December 7, 1998

It doesn't make sense for me. First to do such a test, you need to

shoot at the same subject, same distance, same time, so you need at

least a partner to do this, you didn't mention the setting as well,

every camera doesn't fit the same to another photographer, maybe you

fit better to nikon, he fits better to minolta/ canon, etc. so you

need to be more specific about this.

john_kuraoka1 · December 7, 1998

First off, I am a second-generation Nikon user.

Second, am I the only one who sees a statistical flaw here? Throw out

the one out-of-focus photo in each category with the Canon EOS3 (for

example), and you'd have 8 of 8 and 10 of 10 vs. 7 of 7 and 7 of 7

with the N90s. The inconsistency in the number of total photos

taken with each body is troubling.

paul_wilson12 · December 9, 1998

First of all, the Nikon Digest IS independent of Nikon. It was

created by a guy in New Zealand and is lucky if Nikon actually

acknowledges its existance.

Secondly, in answer to John, the 13 out of 13 or 7 out of 9 figures

are saying that the camera took 9 shots in a given amount of time and

got 7 in focus. The point is that the Nikons didn't take any out of

focus shots in any of the tests whereas the Canons did. Another way

to say this is the the N90s is more accurate but slower than the

Canons. The F5 is both more accurate and faster.

Don't take the above as me saying Nikon is better than Canon. The

statements are based on the info provided by Paul above.

stanley_mcmanus3 · December 9, 1998

I believe that this test is done by having a car drive directly at the

cameras at a speed of about 100 kph or 62 mph. Each camera fires away

as many shots as its motor drive can handle and they see how many are

in focus. What does it prove? It is one piece of test information

that we can use to evaluate the performance of autofocus of various

bodies. For me it shows that the overall autofocus system is what

counts and not just one part no matter how fancy it may be. For

example, the F5 with its screwdriver driven 300mm f/4 beats out the

theoretically faster electronically controlled USM lens by Canon IN

THIS TEST.

However, there are other tests that could be conducted though they

would be much more difficult to do. I would love to see a test of how

autofocus systems respond to a bird that suddenly bursts forth from a

bush and takes to flight. The results might be much different since

the autofocus system would have to include other factors that this

test did not evaluate. Normally I don't photograph cars coming

straight at me. I want to live to be a grouchy old man.

sean_hester · December 9, 1998

about "Secondly, in answer to John, the 13 out of 13 or 7 out of 9

figures are saying that the camera took 9 shots in a given amount of

time and got 7 in focus."

is it really that simple? i also notice that there's closent distance

in there too. it seems that the nikon lenses focus closer then the

canon ones tested. did they do the test timed? or distnaced? the F5

took more pictures, but the canons stopped at a farther distance,

(2-3 times farther away) giving the nikons more time to take pictures.

right?

louis_philippe_masse · December 12, 1998

Here is a small translation of the above mentionned article. I hope

it helps. Please note that these are not my personal opinions (no

flame please!).

___________________________

- The subject is a car at 100 Km/h (60 mph) on a straight road.

- The camera is handheld by a photographer standing on the side of

the road.

- Each body was loaded with 400 ISO print film. They were set on

continuous AF mode and high speed motor drive. For multiple AF point

bodies, the central AF point was selected.

- Each camera was loaded with the same batteries (Energizers lithium

AA's). For the Canon bodies, the boosters were used and for the F90x,

the grip was used.

- Each bodies was tested with both 300mm f/2.8 (AF-S for Nikons, L

USM for Canons) and 300mm f/4 (L USM for Canons).

- All cameras where buyed in ordinary stores while the EOS 3 was

supplied by Canon France since it wasn't available at the time.

- Note: In the article, all the images are shown.

- Note: I haven't read anything on eye control so I'm not sure that

it was turned off (note: eye control slows AF speed according to the

review of the EOS 3 in Chasseur d'Image #209).

Method:

The car was followed by the testers from far away. At a certain fixed

point the shutter button is depressed and the front of the car is

kept in the central area of the viewfinder until the car passed by.

The car takes aproximatly 2secs to travel the distance.

Evaluation:

- Number of shot taken: The number of shot is mostly determined by

the motordrive speed. It is not in any way a measure of the AF speed

For example, the F5 with it's 8fps rate take more picture in the 2sec

interval than a EOS 50 or F70 (also tested).

- Number of sharp images and Closet AF distance: At some point, the

cameras cannot follow the car anymore and the images becomes unsharp.

In this test, the distance from the camera at which AF stop to work

is noted. This distance is determined by the efficiency of the AF

meter, speed of the AF software, speed of the AF motor and

"smartness" of the predictive AF.

- Results:

<CENTER><TABLE WIDTH="450" BORDER="1" CELLSPACING="0"

CELLPADDING="0"

HEIGHT="219">

<TR>

<TR>

<TD WIDTH="14%">#image</TD>

<TD WIDTH="14%">#sharp</TD>

<TD WIDTH="14%">#image</TD>

<TD WIDTH="15%">#sharp</TD>

<TR>

<TR>

<TR>

<TR>

<TR>

<TR>

<TR>

</TABLE>

</CENTER>

The final standing in this test is:

#1: Nikon F5

#2: Canon EOS 3 and Nikon F90x (ex equo)

#3: Canon EOS 1n

_____________________________

Personnaly, i'm still stunned by the results. Why then so many pro

sport photographers use EOS 1n if it so slow? I guess this has

something to do with the fact that before the F90x, there was only

the F4 with it's notoriously slow AF. So in the late 80s and early

90s, Canon was the only option. Today, those pros have invested a lot

in white lenses and it would be stupid to change all over again.

Chasseur d'Image explain that the EOS 1n AF module is older than the

Nikons. For the EOS 3, they blame the software (and computer) speed.

It's speed was increased 8 times over the EOS 5 but since the number

of AF sensors has increased 9 times. Having such a large quantity of

information to process (even in single AF sensor mode apparently),

the effective speed is not simply 8 times the EOS 5's.

Anyways, I am curious to see if Chasseur d'Image will perform a

similar test on the F100.

If someone needs other details of the article, please e-mail me.

Regards,

L.P.

louis_philippe_masse · December 12, 1998

Darn! Sorry for the mess. I did something wrong when formating.

L.P.

russ_arcuri1 · December 16, 1998

A couple of points. The F5 was not using the screwdriver blade to

drive older lenses, as Stanley implies. It was using it's electronic

focus control to control the AF-S (electronic focusing, similar to

Canon's USM) lenses.

Honestly, I don't understand why anyone (other than possibly sports-car

racing photographers) would even care about this test. They tracked a

car moving at constant velocity, using the center focus point only --

even on cameras equipped with multiple focus points. How is this

supposed to relate to anything in the real world anyway?

And even if you do accept the results as being valid in the real world,

what is this supposed to determine? Think about it -- the EOS 1n

snapped 10 pictures in 2 seconds, 8 of which were in focus, and this is

considered SLOW??? Anyone who wants to get rid of their crappy old

EOS-1n and EF 300/2.8 USM or 300/4 USM is welcome to send them to me.

I'll be sure to only use them for non-moving subjects, like landscapes

;-)

danny_weber · December 16, 1998

Nikon doesn't have an AF-S 300/4. The only AF 300/4 is mechanically

driven, and it's currently the oldest lens in the Nikon lineup that

has not been updated since its inception. Most users consider it the

slowest-focusing lens Nikon makes.

stanley_mcmanus3 · December 16, 1998

Russ, the 300mm f/4 is not an AF-S lens, though I am sure many Nikon

shooters would not object to Nikon producing this lens and the 400mm

f/5.6 lens in a Silent Wave edition.

I do agree that the test is of limited use. Note my comments on

focusing on a bird that bursts out of a bush. Still it is better than

nothing. And at least it proves that Nikon's 'screwdriver' system is

not necessarily slower than Canon's all electical system. As others

have said it is the total system that counts.

sean_hester · December 17, 1998

is the nikon 300 2.8 and AF-S lens? if so, it's wierd that it's not

any faster then the non AF-S 300 f4 isn't it?

kinda makes you wonder why someone would spend the extra money for the

AF-S. aparently not for AF speed.

of course this is probably wrong. just one other reason to doubt the

usefulness of this test. since the car is coming toward the camera,

it has to move focus only slightly for each frame to keep up. making

the AF speed of the lenses a MOOT POINT. this test is testing the

capability of the bosies to predict, not the lenses to focus.

now... what would happen if the test was track the car for a while,

switch focus to the person standing way back at the starting line,

then pick up the car again. this would probably bear out that the USM

and AF-S lenses were better then screwdriver driven lenses.

Sign In

Autofocus Comparison

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Create an account or sign in to comment

Create an account

Sign in