The "kit" lens revisited, is it as bad as they say?

I have taken 2 pictures, one at f8 and the other at f16 with the following
lenses:

Canon EFs 18-55 f4,5-5,6 ("kit" lens modified to fit on a 10D)
Canon EF 50mm f1,8
Canon EF 24-85 f3,5-4,5
Canon EF 35-80 f4-5,6
Canon EF 24-70 f2,8 L

So we have 4 zooms, 3 plastic cheapie lenses, one middle of the road zoom
and a primo pro lens. The "kit" lens is one of the early ones which came
with my Drebel, the newer ones are said to be better. I tried to put all
lenses at 50mm and not move the tripod but there are still some slight
changes from one lens to the other.

The pictures are at http://www.pbase.com/jeandr/50mm_test

Now which is which?

Jean

PS These are all TIFF files and at about 4Mb each so a broadband connection
is better for download. There is no EXIF to make sure no one cheats (unless
I missed something embedded in the files which is quite possible)

 

Lens 1 Canon EFs 18-55 f4,5-5,6 ("kit" lens modified to fit on a 10D)
Lens 2 Canon EF 50mm f1,8
Lens 3 Canon EF 24-85 f3,5-4,5
Lens 4 Canon EF 35-80 f4-5,6
Lens 5 Canon EF 24-70 f2,8 L

Lucky guess?

 

Jean,

Is this a good test? .JPG being the great web equalizer. Print might show
stuff worth looking at.

The quality of the Kit lens isn't tested (hardly) in this picture. Niether
are any of the other lens.

If you really want to test these lens, shoot my daughters basketball games.
Low light action no flash. The only lens that has a chance is the 50mm f1.8,
but its a prime so your composition choices will suck. The 24-70 f2.8 L
might have a chance if it has IS and its on a tripod.

I agree, the kit lens is sufficient for a lot of things, your test shows it.

But it tests *web .jpg's* and this isn't a very demanding environment. It's
like web jpg's are a static picture of a car, and you took a picture of 64
Bronco, 69 Mach One, 2005 Boss and a Ferrari (any year) and said "see, which
one is faster" You can't tell from the static picture. The Bronco would kick
on the off road, the Ferrari would kick on back roads, the Boss on... well
lots of things and the Mach One would kick at an old timers 1/4 mile retreat
and car show circuit scene.

DA.

 

It's also shot at f8 & f16. The former at an aperture where a lens made
of recycled coke bottles might be expected to perform well, the latter
at the level on a 10d where diffraction would be a great leveller of
results.

 

This only proves one thing: that if the comparator is dumbed-down
shots for web use, it doesn't matter which lens you use. They are all
bad enough.

;-)

 

20% You got one right!

Jean

 

If you don't see the faults in full size JPEG, you won't see it in
print. If you compare the print to the screen, you'd notice that the
screen has more vibrant color. This is _not_ a test of the printer, so
on screen JPEG in original size is fine for comparison.

And why do you want to test how fast a lens can capture your daughter's
basket game? This is a lens quality test, not a test of how wide or
expensive the lens is.

There is indeed one problem with the test, why choose f/8 and f/16? It
should be at the lens widest aperture and also one taken at the common
aperture of all the lenses, possibly f/5.6.

 

f8 was to equalize the slow lenses which would probably be at 5,6 nearing
full zoom (the kit lens). f16 I thought would help the bad lenses. I was
really surprised at the 35-80, for a plastic cheapie it's better than the
kit lens (another plastic cheapie).

Jean

 

It's easy enough to put the sharpest and softest images at each end.
The interior is borderline undecidable -- a machine can probably rank
them though. But who cares? f/8 or f/16, the above differences (with
perhaps the exception of image #5) can be ascribed to simple lack of
focus in the absence of other data.

 

it.

I guess so, for travelling by bicycle it would be just the thing, wide and
light.

Jean

 

? I should have shot raw? I can try it again with 4 lenses this time and
more colors in the picture.

Jean

 

Autofocus, or manually focused, with good eyesight? If autofocus,
forget comparisons.

Colin D.

 

You could also show a 100% crop of the same area in each image. Something
with some backlighting in places would be a good way to see vingetting/
color fringing.

 

This is the other problem with the test: since only you know the
answers, you can (if you wish) simply make it up as you go along. Who
can tell? You may wish to publish a hash of the results prior to
revealing the 'answer', though this can't remove all doubt.

 

Does that mean autofocus is erratic or never accurate?
-Rich


"Bittorrents are REFUNDS for all the BAD movie products Hollywood
never gave us refunds for in the past"

 

It won't matter what you do, if the results don't agree with the
Canon cranks preconceived opinions, they won't buy it.
It's just outragious to suggest that you have to shoot RAW to
make a comparision when dealing with full sized images.
Unless Canon's JPEG conversion is so terrible, and there is some
evidence it is.
The sad fact is, apart from maybe 2 or 3 lenses, Canon products do
not measure up to the best from Nikon or Olympus. Canon does not
produce lenses to be ultimate image makers, because the main
users of their high quality products are people like sports photogs
and they definitely do not need top quality. They need speed
(f-ratios and focusing) and a good range of focal lengths.
And THAT is who Canon produces equipment for.
-Rich


"Bittorrents are REFUNDS for all the BAD movie products Hollywood
never gave us refunds for in the past"

 

The slow lenses are wide open at f5,6, in order to give them half a chance,
I closed them down a bit and closed all the others to the same aperture to
have the same depth of field on all lenses so that would not be a factor.

Jean

 

And to me too, I like my Canon stuff. I don't have a Nikon so I can't say
which is better or if it is better.

Jean

 

You can doubt my honesty if you want, but why should I fudge the results? I
have nothing to gain or to lose. Since I don't want to drag this on for
months, I can tell you right off that the only person bold enough to venture
a guess had lens 4 right.

Jean

 

Noted.

Jean