Ok I'll play along.

I'd say the middle shot.   At least that how my shots with the 18-55 look...
Soft.

--

Rob

Well, I'll stick my neck out. I have to say that on screen the
differences did not seem that great in the central area.

To me, picture 3 appears to be the sharpest, both in centre and at the
edges. The writing on the TV connector sockets, in particular, is way
better than on 2 (it is cropped off on 1, but from the cloth it would
likely be worse).

Pictures 1 and 2 are harder to call. This may in part be because you
have (I suspect) set the lens for 1 at a slightly longer focal length,
and thus have slightly less DoF (it has a little more bird but a lot
less TV). Thus picture 2 seems to me better at the edge (closer subject)
- but it seems slightly, very slightly, less sharp than picture 1.

Thus I would go for 2 as being the poorest lens.

Whether that is the 18-55, I don't know. While it has been slated by
many here, plenty of people of good credentials have praised it. For
example, David Kilpatrick (editor of UK magazine Freelance + Digital,
and himself an accomplished professional) says in the June/July issue of
that magazine: "We have a Canon 300D which has a new cheap 18-55mm
digital lens - great little performer."

To be honest, even if a lens was a little below the standard I would
normally choose to use, I would expect it to be hard to tell by f/8. It
may be more informative if you were to use f/4.5, or whatever is the
widest all 3 can achieve. Use manual focus, take great care to focus on
the same point in all 3 - and take more care to get the focal lengths
exactly equal!

Nice to see someone actually trying an experiment rather than just
shooting his mouth off, though - thanks for that.

David

PS - I just realised my daughter has one of these controversial lenses -
came on the 350 I bought her a couple of years ago. I have never thought
to give it a try myself. Going on holiday tomorrow though, so it will
have to wait.

#3 looks like the sharpest
#1 is second sharpest
#2 is the least sharp

I know which is which, the EXIF says it all but my answers were not done
according to what I knew, but according to what my eyes told me.

I will have to duplicate this test, I have many lenses which can all be set
at the same focal lenght.

Jean

You have a lot of dust bunnies showing for F8.

I vote middle for the 'kit'; image quality/sharpness at normal viewing
or Web sizes, probably not much to choose among them, but in this
presentation my eye likes #3, #1, #2, in that order.

I can spot where you took down some pictures.  Nail holes!

I'm not going to guess on which one is which.  1 and 3 seem good, with 3
being the best.  

2 doesn't seems to have its focus behind the objects on the desk,
because the picture frame is in better focus than the foreground
objects.

One of the psychological advantages to viewing this sort of test is
knowing the photographer is so engrossed in their willingness to
demonstrate a point, they forget to sweep the path behind them and
almost always make the first image, the one with conjecture. It helps
too to have the EXIF data intact!

All of those shots have substantial chromatic aberration and none were
shot at the same exposure value. Which means they are not APO lenses and
like the OP said, not a scientific test. Because all the lenses are
mediocre quality, it is impossible to visually tell the difference
although the first image is the one from the kit lens.

Make = Canon
Model = Canon EOS 350D DIGITAL
Orientation = Normal
XResolution = 72.00
YResolution = 72.00
Resolution Unit = Inch
Date Time = 2005:08:06 14:06:44
YCb Cr Positioning = Co-sited
Exif Offset = 196
Exposure Time = 1/6 sec
FNumber = F8.0
Exposure Program = Aperture priority
ISOSpeed Ratings = 100
Exif Version = "0221"
Date Time Original = 2005:08:06 14:06:44
Date Time Digitized = 2005:08:06 14:06:44
Components Configuration = YCbCr
Shutter Speed Value = 0.1667 sec (1/6)
Aperture Value = F6.0
Exposure Bias Value = 0/2
Metering Mode = MultiSegment
Flash = Off, surpressed
Focal Length = 27.00 mm
Maker Note = "."
User Comment =
Flash Pix Version = "0100"
Color Space = sRGB
Exif Image Width = 3456
Exif Image Length = 2304
Interoperability Offset = 9230
Interoperability Index = R98
Interoperability Version = "0100"
Focal Plane XResolution = 3456000/874
Focal Plane YResolution = 2304000/582
Focal Plane Resolution Unit = Inch
Custom Rendered = Normal process
Exposure Mode = Auto
White Balance = Auto
Scene Capture Type = Standard
Focal Lengthin35mm Film = 43.80

 --- Maker Specific Data ---
ExposureInfo1 = 92, 2, 100, 3, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1,
32767, 3, 2, 0, 3, 65535, 65535, 55, 18, 1, 65535, 65535, 0, 0, 0, 0, 0,
65535, 65535, 0, 0, 0, 0, 65535, 0, 32767, 0, 32767, 65535, 65535
Macro mode = Normal
Self Timer =  2
Compression Rate = Basic
Flash Mode = Not fired
Drive Mode = Single
Focus Mode = One-Shot
Image Size = Large
Easy Shoot = Manual
Contrast = High
Saturation = High
Sharpness = High
CCD ISO =
Metering Mode = Evaluative
AF Point =
Exposure Mode = Av-priority
Long focal =  55
Short focal =  18
Focal Units =  1
Flash Details =  0
Focus Mode = Single
ExposureInfo2 = 68, 0, 160, 120, 192, 83, 0, 0, 3, 0, 8, 8, 0, 0, 0, 0,
0, 0, 1, 0, 0, 192, 248, 119, 0, 0, 252, 0, 65535, 0, 0, 0, 0, 0
White balance = Auto
Sequence Number =  3
OpticalZoom Step =  8
AF point =  0
Flash bias = 0 EV
Distance =  1
ImageType = "Canon EOS 350D DIGITAL"
FirmwareVersion = "Firmware 1.0.2"
OwnerName = "unknown"
CameraSerialNumber = 730524401
CustomFunctions = 20, 0, 256, 512, 768, 1024, 1280, 1536, 1792, 2048
Noise Reduction =  Off
Shutter AE Lock Button = AF/AE lock
Mirror Lockup = Disable
Exposure Level = 1/2 stop
AF Assist = On (auto)
AV Shutter Speed = Automatic
AEB Sequence = 0

File: IMG_0095.JPG
Exposure Time: 1/4 sec
F-Stop: F8.0
ISO: 100
Lens: 24.00 mm
Date: 2005:08:06 14:07:26

File: IMG_0096.JPG
Exposure Time: 3/10 sec
F-Stop: F8.0
ISO: 100
Lens: 24.00 mm
Date: 2005:08:06 14:08:05

Looked at EXIF info.  Is the lens for #2 also a kit lens?  Wonder what
that says about standard vs wide angle zooms?  Also interesting to note
that the shutter speed was 1/8th for the 18-55 and 1/4th for the other
two:  wonder why.  Apertures were 8.0 for all...

Thanks,

Longfellow

Nice one Dylan...

I cropped a small section from each image and spread them three across in
photoshop and agree with the comments of everyone else as the the sharpness.

Whilst I know the point of this is a direct comparison, I did what I believe
is a more real world test on it.

I printed a full horizontal crop of the same portion of each image and
printed the three strips onto a A4 sheet of paper and then viewed the image
at the recommended 2.5 times the diagonal.

And you know what? At that distance, the normal viewing distance, to me it's
imperceptable. A very high quality image by all accounts, created on a very
good consumer camera with what I regard high quality lenses. (and I'm
currently using a Nikon D70)

At the end of the day, I guess it depends on where your photographs end up.

It's indicative of the kind of quality we expect and demand of a consumer
camera/lens these days. We seemed obsessed in splitting finer and finer
hairs and losing sight of the fact that although one lens will always be
sharper than another at certain focal lengths, pretty much all lenses
capture images with a sufficiently good sharpness for the end purpose. And
for most consumers, that's the ink jet printer or at best a consumer lab.

If you're needs are more critical, if that extra degree of sharpness is
critical to you earning your money, then it's likely you'll use a tool fit
for purpose. Maybe that means using a medium/large format camera or a more
expensive lens range.

I urge everyone to print the image as I have, and at the appropriate
distance judge for yourself if you would consider any of those photographs
as 'unacceptably soft'

Obviously those of you whoread the EXIF know the answer but is was :

1. 18-55 'kit' lens
2. 24-85 3.5-4.5
3. 20-35 3.5-4.5