Purrfect............sorry

yeah, not quite sure what I was getting at either - hence the 'confused' emoticon.

You're quite right to point out the noise and processing issues of high density sensors but they'll no doubt eventually be overcome.

It has been generally recognised for some time that a compact with around 4MP delivers similar results (at least for standard 6x4 prints) as a 35mm film compact and your's and Mr. Mead's calculations would support that.

You can of course be just as creative with a digital compact as with a DSLR but the nature of SLR photography demands, at least from the real enthusiast, that the extremes of what can be achieved are sought.

What seems impossible today, may very well be commonplace tomorrow!

He's not saying that 20MP/square inch is the maximum that you can have, he's saying that there is no advantage to having more than that.
The respective characteristics of silicon based CCD/CMOS sensors and the wavelength of visible light, results in a maximum resolution at around 20MP/square inch - if you had more than that, you simply wouldn't be able to see the difference.

Quite how that relates to the sensors in compacts, I don't know but it does make you wonder why DSLRs don't have the same pixel density.
It surely can't be a cost issue because the sensors are so readily available for compacts - can't a DSLR simply have two or more compact sized sensors?

However, let's say you had four 10MP compact sensors arranged in a 2x2 grid. You'd have a 40MP camera but each quarter of the image would only consist of 10MP......so would there be an advantage

Discussions regarding the number of pixels are very interesting but it seems to me that unless you want to dramatically enlarge a small area of a scene, 10 or even 6MP, in APS-C format is perfectly adequate for most people.

We know the final image quality is not all about the number of pixels but that is nevertheless a huge selling point.

Mr. Carver Mead of Foveon (he knows a thing or two about silicon chips) believes that a sensor with around 20MP per square inch will deliver the maximum possible resolution. That equates roughly to 14MP for APS-C and a massive 30MP for full-frame (35mm).

The obvious advantages for full-frame are: 35mm lenses will deliver the AOV for which they are designed (bear in mind that only truly exceptional lenses have no apparent drop-off toward the edges) and less enlargement is required to display or print at any given size - just think of the file sizes though, data storage these days might be cheap but flippin' heck!
Full-frame cameras would also be slightly bigger and heavier than their APS-C counterparts (I say slightly because MZ series cameras aren't exactly unwieldy are they?).

So I conclude that while a 30MP full-frame DSLR sounds attractive and may one day be common place (never say never in this crazy techno world), I think a 14MP 'K1D' is much more likely for the foreseeable future but whether it'll perform significantly better than the K10D, remains to be seen!

For the moment, I think Pentax satisfy the target market adequately with the K100D/K110D and K10D and there's no need for any intermediate models - but knowing how they produced *ist DS/DL variants, I wouldn't put it past them!

Re. A series 50mm on digital - that's another good point John, even the 'baby' of the bunch (the F2), might not be the best but it's still a cracking little lens and a good manual alternative to the hard to find F or FA series.

Another interesting thread, though perhaps not what petekd was after!

Suffice it to say that whether you select the aperture directly from the lens or from within the camera, that aperture has to be available ie. a lens with a maximum aperture of F2 cannot be opened to F1.4.

John is quite right to point out that modern, complex lens designs means that lenses of the same focal length and maximum apertures can vary enormously in their physical dimensions. The quoted focal length is really more to do with the angle of view, rather than any actual physical measurement - 50mm lenses, irrespective of their design, have a 47 degree AOV.

For any given focal length, a faster lens (wider max. aperture) with have a larger diameter front element and a really fast one might be so large as to require a wider body and therefore a larger diameter filter.

Prime example (pardon the pun): SMC Pentax A 50mm F2, 50mm F1.7 and 50mm F1.4 have progressively larger front elements but all take a 49mm filter and only become slightly larger and heavier. The 50mm F1.2 however requires a 52mm filter and weighs almost three times as much as the F2 lens.
The F2 features 5 elements in 5 groups (seperate elements, in other words), whereas the F1.2 has 7 elements in 6 groups (two of the elements are placed or bonded together).
The respective physical dimensions (max. diameter and length) are: F2 lens is 63.0mm by 31.0mm and the F1.2 lens is 64.5mm by 47.5mm.

Faster lenses are bigger, heavier and more expensive than their slower cousins but in some cases can be relatively poor performers - the more glass you have, the more critical the tolerances need to be!

I quite like all three of the original shots.
The blurred waterfall might be a little cliched but then again, no matter how original you might think your shot is, you can bet your life that someone, somewhere has taken one very similar.

This one works well for me, there's plenty going on, the colours are reasonably saturated and the whole thing has a certain fairytale quality.

I never place too much importance on the chosen frame or border because that can be changed and if I were to own and display the image, I might tweak it a bit here and there but I think it generally works fine.

All criticism can be viewed in a positive light and as we all see things differently, one person's opinion can be totally different to another's but both can be equally valid.
To say that certain rules under certain conditions must be strictly adhered to is the very philosophy that stifles artistic impression and creativity.

It's wise to seek criticism and to follow advice but ultimately, you should go with your own feelings and if you like the end result then that is all that matters.

Well, I've never been overly impressed with Sigma products, some of the lenses are superb and the cameras are OK but I wouldn't switch.

For my money, Pentax do and always have offered the most complete package, with genuinely useful innovations and there is very little if anything missing from the K10D.

Specification wise, the SD14 is a pretty good effort and I think the Foveon sensor certainly has it's benefits. I dare say that the design is patented but we'll eventually see similar systems in other cameras - perhaps the next generation DSLRs from Pentax!

Thanks for the comments folks

How do you rate the Sigma SD14?

I picked up a brochure at Focus on Wednesday and it's quite an impressive camera. It's about twice the price of a K10D and looks almost as though it could have been designed by Pentax.

There are a few omissions from the specifications - no SR (although it does have Mirror-Lockup), no redeye reduction for the built-in flash and there's no mention of extensive weather proofing.
Nevertheless, it looks like serious competition for the K10D/GX10 and the two main features that stand out for me are the high-eyepoint viewfinder (Pentaprism 98% x 98%) and the 14 mega-pixel Foveon x3 sensor.

The Foveon sensor has been around for a while and I'd heard the name before but didn't know anything about it and it sounds incredible!
It's a CMOS sensor that's cheaper to produce than CCD and has three colour (RGB) layers. According to the brochure, this sensor captures 100% of the red, green and blue light at every pixel location, unlike the standard Bayer-filter system that only captures 25% of the red and blue and 50% of the green light.

Apparently the Foveon sensor is unique in that it has all the colour sensitivity of film whereas the Bayer-filter requires interpolation.

Here's a quote from the brochure:

".......each pixel (Bayer-filter) only captures a third of the colour data, and the remaining two thirds have to be extrapolated using complex algorithms. Interpolation using a colour filter and artificial color synthesis results in a loss of detail, so in the conventional system, no matter how many extra pixels it has, a conventional camera can only record artificially-generated images by calculating colours that weren't even there in the first place."

At present, the Foveon is only available to Sigma and it's a pity that their cameras are not PK fit because I'd love to see some comparisons between the results from a K10D and an SD14, with perhaps a Pentax Ltd. prime lens fitted to both - I suppose a PK-fit Sigma lens on a K10D would serve that purpose though!

Also, I read an article by the developer of the Foveon system who said that the 16.8 mega-pixel sensor that they had developed was very close to the effective maximum that could be achieved. The limitation is due to the frequency range (or was it wavelength?) of visible light, so I guess we'll never see a benefit above 20 mega-pixels per square inch!

Hope you found that as interesting as I did

I think you've been really unlucky Peter and there certainly seems to be a fault somewhere..

It's worth remembering that rechargeable batteries require several complete charge/discharge cycles to reach their full capacity - whereupon they begin to die........but that's the nature of batteries!

My father's Olympus OM2SP recently died after many years of use and the mechanical engineering within the camera would make a traditional watchmaker proud. It's amazing that with so fewer moving parts, that modern cameras and electronic devices can be so unreliable but I guess it only takes one faulty diode or piece of circuitry, to render it useless.

Good luck with the next one