Close up Quality

A friend of mine is very unhappy with the quality of his Nikon zoom lenses, so we are doing a test on them a little later today to check them out. I've already run the test on various Pentax and other items, and the results are rather interesting.

This is only applicable to the close up still life set I have assembled, so it makes no comment on longer distance work, and it is also only applicable to one aperture of f8 for SLR lenses, and f6.3 approx for the smaller formats of CCD. (I was trying to offset slightly differences in DOF).

So I had the set and took the pictures with the following results:

The Winner: Pentax Optio 750Z. Staggeringly sharp, although slightly noisier but not obtrusively so.

Runner Up: Fuji Finepix S7000 Almost as sharp, but a less pleasant noise structure, so not so good pictorially but very sharp in close ups.

3rd: Pentax *istDS with 18-55mm DA lens. Almost as sharp as the above two, but not quite, but a more pleasant rendering and no noise.

4th: Pentax *istDS with 24-90mm FA lens. Not as good as the digital lens by quite a margin. Edges relatively soft.

5th: Pentax *istDS with 35-70mm f4 A lens. Dreadful by comparison. This one does not have the same fine detail rendering and is quite doft at the edges. I shall leave this for film cameras, where it performs well, and for portraits on digital, where it also gives very nice rendering of skin tones.

A little later I shall tell you how the Nikon 28-105mm zoom fared on the Fuji S2 Pro. I might have to delay the report to watch "Doctor Who", but I'm sure you will understand!

Hi John
Me and my Optio 750z will pop in on Sunday to see how you got on

Right, the results are ready....

And the Nikon lens proved itself to be on a par with the *istDS with the DA 18-55mm lens.

But it is not quite as simple as that. The Nikon lens revealed marginally superior fine detail, but bear in mind that it could have been extra resolution from the Fuji CCD, which gives a 12MP output from its 6MP chip. It was only marginally better fine resolution as well.

However, the Pentax lens was clearly superior in its colour rendering and in the contrast and overall sharpness. This could also be a property of the CCD as well as the lens.

So, if you want vibrant colour, high sharpness, low noise and lack of artefacts the Pentax had the edge.

If you want a little more resolution then the Nikon/Fuji combination had the edge, just.

So it was joint third place for the *istDS + 18-55 and the Fuji S2 Pro + Nikon 28-105, with the same caveats on the conclusions as before.

An interesting exercise, and a surprise win for the 750Z! Don't forget that I was evaluating sharpness and detail reproduction, not any other lens properties....

Interesting test, John.
However, bear in mind that most compact digitals will sharpen much more aggressively than a DLSR anyway. I have some seriously good Photoshop sharpening techniques that will make the *ist-D & *ist-DS much sharper
I'll dig out the links later
Matt

Here are the links (quite a read I'm afraid, but worth it )
http://www.creativepro.com/story/feature/12189-1.html
http://www.creativepro.com/story/feature/20357-1.html
I'm thinking of putting some actions together... I've tried these techniques and they're very good
Have fun!
Matt

I have to agree with Matt that these results probably reflect differences in the degree of sharpening applied in-camera rather than differences in the resolving power of the lens/sensor combinations.

I would be flabbergasted if the 750Z with its tiny sensor and cheap lens could outperform a D or Ds with the 24-90 zoom.

I also think it is not realistic to compare just one element of a lens's performance. It's easy enough to improve one element if you don't care what happens to the others.

It would be interesting to see what Matt can do in Photoshop with these images.

G

What you are both saying is undoubtably right, but there was a purpose for the parameters of this test, and it found the information that my friend required.

But you should try it out, because the sharpness and resolution of the 750Z was quite remarkable. There was no real sign of over-sharpening either. Where it did slightly fall down was in the noise level (unsurprisingly) but even that, at ISO 100, was not too excessive.

Anyway, exercises like this are a bit of fun on long Saturday afternoons....

I just thought I'd add a little more info regarding the methodology of the test we did.

I set up a still life subject on a 20"x16" area, comprising a book, a street map atlas and a small weighing balance. This gave me colours, bold outlines and graded fine detail (on the street map and the wood grain of the balance). It was lit by large window light, the camera aperture was selected at f8 (f6.3 for the smaller chips) and the exposure regulated by aperture priority auto exposure. Obviously a very solid (Manfrotto) tripod was used.

The pictures were taken with each camera/lens combination in turn, framing identically, and the unmodified files were saved as JPEG maximum quality and examined. The files were then optimised in Photoshop and saved as a modified version before further examination.

Within these parameters, the results were surprising but pretty clear, and I won't be using the 24-90mm zoom for this sort of subject. Noe would I want a Fuji S2 Pro with Nikkor lens, at least not the one tested!

But the *istDS with 18-55 came out very well, more so when the tonal qualities are taken into account. The 750Z with its 9 element in 7 groups aspheric lens is superb and will be a constant companion when photography is not the main issue of the day.

However, it is very fair to say that when testing a lens all the perameters should be considered, including:

Focusing distance (some lenses are better close up, some better at infinity)
Aperture (optimum apertures vary)
Zoom setting (quality varies throughout the zoom range)
CCD (how much is due to CCD quality, how much to lens quality, how much to the match between the two?)
Design parameters of lens (high resolution may mean low contrast, there is a mass of balancing to do in the design parameters of any lens)
Subject matter (some lenses are better designed for some subjects than others)

And so on. When I've got a spare week I might pursue it further!

Excellent work John, most interesting.
I wonder if it might have been an idea to have chosen the same resolution rather than the maximum for each camera ie. 3MP un-interpolated?

Hi Kim

That would be possible to do, but in this particular context we were looking for the maximum quality obtainable from each camera. I was generally interested in how they delivered quality, Paul was looking to check that his new Nikon zoom was critically sharp. One limitation was that the Nikon zoom ended up on the Fuji S2 Pro, and there are umpteen different parameters that we could have set for the camera. So we set every camera at the midpoint for sharpening, colour saturation and so on, but at the maximum JPEG quality.

So many possible choices....

Fair enough John, remarkable results for the 750Z - another Pentax triumph!