[Frank Fletcher]

OP took 2 crops (piece of chain) with different number of pixels in each and then scaled at least one of them so that both will be the same size on the screen in his posting... either the crop from K20D is scaled up or the crop from WPi is scaled down...

Ahh...thank you. That would completely invalidate the results.

[Marc Sabatella]

To me, it's like saying someone should invent an affordable lightweight 17-70/2.8 lens

Here, unfortunately, I *do* understand enough of the physics to know this isn't going to happen :-)

[dylansalt]

I have an Oly 5060 which has extremely good glass (a class leader in its day).

After having done many many side by side test shot comparisons, 100% crops - against the humble kit lens on the K10d.

I can say that my Oly simply cannot come close and that is at 80 Iso (Oly) to 100Iso (Pentax) in Jpeg.

The other day I printed a K10 1600 ISO pic off a cheap Hp and was simply blown away.

Today 1600 iso performance definitely has real world applications - p&s's cannot compete whatsoever.

Compared to 35mm film camera's, digital Full Frame's are simply mammoths.

In fact my Contax and Minolta 35mm are way smaller than my K10 (and significantly lighter)-so for me Digital FullFrame will only be attractive if they get the weight and camera size down to realistic dimensions and weight

[marlon]

serious unscientific mismatches happening here... tehehe.
Id rather have less MP and better iso performance, which isn't just "sometimes" used, as popular belief states.
Besides, pentax is dropping the p&s market isn't it?

[Douglas_of_Sweden]

You are right, but then (in the film era), no one had shake reduction. At least, Pentax claims a gain from 2 to 3 full f/stops with shake reduction. And we get that with every lens, not only special lenses from C/N.

No doubt the digital era is making us lazy. Shooting ISO 1600 is far easier than shooting ISO 400 with a consumer zoom... at f/5.6 as max aperture (long end). Of course, considering a two stop gain (stability), then shooting f/5.6 becomes more handy at slower speeds, as if shooting f/2.8 with the same ISO.

Robert B.

Yes, SR is changing the rules, but it is still quite new. Before this I think the drive to give higher ISO was much caused by the loss in speed when zooms definitely replaced primes as the first choice.

[wildman]

The K20D sensor works for me...having the K20D sensor compared to my DL is like having longer fl glass.

I don't need a bigger sensor just more birds.

ISO 800. Taken through a 560mm scope at a distance of about 25 feet

[marlon]

Stunning work there wildman, my compliments!
And yes I agree, the MP race is a blind and uneducated one. I wish people could focus on what really makes the difference.
That is a stunning bird, what is it?

[RiceHigh]

I think you counted only the pixel count, but forget about all other very important things on image quality such as noise levels, dynamic range, DoF, colour accuracy and so on.

So, the limitation on the image quality is not on the resolving power of the lenses, but on all of the above other things.

I suppose you are not suggesting P&S DC image quality in future Pentax DSLRs, are you?

[wildman]

Stunning work there wildman, my compliments!
And yes I agree, the MP race is a blind and uneducated one. I wish people could focus on what really makes the difference.
That is a stunning bird, what is it?

It's a young first year Baltimore Oriole (Icterus galbula). Next year, when he grows up, he will look like this....
Taken with my old 6mp Pentax DL by the way.

[kenyee]

ISO 800. Taken through a 560mm scope at a distance of about 25 feet

Yow. Now I wish I had upgraded my K10D this summer

[rburgoss]

I discovered some further information that can shed some light here.

After checking both files properties, I noticed that the reported size for both, do not exactly match their size relationship between real actual sensor sizes, according to specs. That means that most surely, the WPI has interpolated its resolution by more than twice in order to achieve the full 6 mp. That is, considering the K20D is not interpolating. If both cameras are interpolating, then all calculations are meaningless.

The bottom line is that whether real or interpolated pixels, the compact wpi seems to perform a lot better (in resolution) than the K20D.

Attachment 19830

[MoiVous]

There seems to be some problems with the the maths and the physics here.

If you go back to the basics, the WPI has a 2.04um pixel, whilst the K20D has a 5.03um pixel size.

Furthermore, if you compare the crop factor for the WPI vs the APS-C camera, then the WPI has a crop factor of 4.5 x 6 compared to the APS-C value of 1.54 x 1.54.

So to have a valid comparison in image size and then compare pixels the 18mm focal length on the K20D is 27mm vs 81mm on WPI.

Me thinks it is not a fair comparison. You should therefore have the WPI set to a focal length of 6mm to have the same field of view on both cameras.

Now do your comparison.

The second issue to question in your argument is the ability of the pixels to resolve the photons striking them. Smaller pixels will have poorer signal to noise ratios. A smaller pixel will have less photons in it and the signal to noise (the square root of the number of pixels striking it) will therefore be signficantly higher.
Therefore, the WPI, with the smaller photoelement will have MORE noise as it has a physically smaller dimension to capture those photons.
In practice, this means that for smaller sensors you need far better lenses. Is the WPI lens better than your SLR lenses?
There will also be an effect on contrast and spatial resolution.

If you want a proper analysis of this check out the following links:
DSLR SENSOR SIZE AND PIXEL DENSITY
Clarkvision: Does Pixel Size Matter
http://isl.stanford.edu/~abbas/group/papers_and_pub/pixelsize_talk.pdf

So please, go back and take your images again - then we can validly compare them.

BTW, if you've got access to a K100D then the comparison there would be valuable.

[ManuH]

The second issue to question in your argument is the ability of the pixels to resolve the photons striking them. Smaller pixels will have poorer signal to noise ratios. A smaller pixel will have less photons in it and the signal to noise (the square root of the number of pixels striking it) will therefore be signficantly higher.
Therefore, the WPI, with the smaller photoelement will have MORE noise as it has a physically smaller dimension to capture those photons.
In practice, this means that for smaller sensors you need far better lenses. Is the WPI lens better than your SLR lenses?
There will also be an effect on contrast and spatial resolution.

It's true that a smaller pixel will have more noise but we have have to compare the effiency on a give area, not on each individual pixel. If a big pixel can be broke into four pixels, we will get more details, each small pixel will have more noise, but collectively they should have the same signal-to-noise ratio as the big pixel. This is the theory. In practice, it's important that all no photon hitting the sensor is wasted. That's why gapless microlenses designs are important.

Also the difference between the WPI and K20D can be due on the absence or very weak AA filter on the compact. They have higher pixel density so I wouldn't be surprised if they didn't need an AA filter.

On dpreview forums some people have guessed that more than 100MP was needed to capture all the details in a picture, even with *bad lenses*. So IMHO the D700/D3 12MP are wasting the FF potential of more details. They don't really have less noise, they just hide the noise. Noise is also part of the details. They hide noise, and thus they also hide details.

[Douglas_of_Sweden]

The K20D sensor works for me...having the K20D sensor compared to my DL is like having longer fl glass.

I don't need a bigger sensor just more birds.

ISO 800. Taken through a 560mm scope at a distance of about 25 feet

Great shot!
What scope was it? How did you connect the camera? T2?

[per]

Personally I really don't see the point in having anything better than iso 800, if people don't wan't to invest money in good glass then their images can just be noisy, it's their fault, lenses and equipment is cheaper now then it ever has been and things are only getting cheaper, but different from the old days, marketing has taken over. Where good quality equipment in the old days would sell itself based pureley on quality and word of mouth, this isn't so anymore. There is no prestige in producing good quality products anymore, it's all gone, as long as a product delivers up to a certain point price is all people care about. Money and profit. You get what you pay for, if you are not willing to pay $1000 for a zoom lens then you are missing the big picture. If you want good quality gear, you NEED to invest MONEY. A zoom lens is a much more complex construction than a prime, to get good quality images at all focal lengths from a zoom requires a lot of engineering, testing and failing. This costs money and this will reflect in the price of the lens. And profit is and you must not think otherwise, profit IS the only motivation behind any company producing any kind of equipment. So if they can make more money by selling us 95734956349576mpix cameras with slow cheap zoom lenses then they will do so. Pentax are a little better here than canon and nikon, i mean, EOS 450D kit lens is just a piece of crap, pentax kit lens is slightly better, but not because pentax want's to, but because this is one of their few weapons against the bigger companies, quality, if they can somehow get us to believe that pentax is better bang for the buck, then the marketing will work. Same with megapixels, we dont need quadzillion megapixels, id rather have 10 mpix sharp and with low noise, instead of crappy 84 million noisy shit pixels. FF or APS-C, pentax has chosen aps-c and the lens roadmap clearly illustrate this, If pentax were to go ff they would need to develop atleast 10 new lenses or use old designs, this would need facilities for production etc. etc. = cost money, I'm pretty amazed some of you ever thought pentax were going ff in the first place, where would the benefit be? They would need to maintain both aps-c and ff frame lens roadmaps, it would be to expensive for them.

Seems like I wen't a little outside the subject there, well well.