Ideally (no such thing though) it should be something like this:

I would be interested in some sample shots, Sigma SD14 has removable UV/IR filters so it should be able to do IR.
I would prefer something with F-mount, Nikon or Fuji for the price of SD14...

EDIT: alpha they are not using Bayer CFA

Sorry I wasn't clear soap. The 4 stages I was trying to show were light going through the bayer filter, being picked up by the photosites, the information being processed the Expeed, and then finally an output pixel coming out.

soap said:
In raw 4 sensor photosites do not equal only one pixel, and the onboard camera processor is not the one which demosiaics multiple photosites into multiple pixels.

I got the demosaic part, but didn't quite get the rest? Can you expand on that? Thanks.

Interesting thread, but no one really got back to me-

does anyone know anything about the Sigma SD14 and how it performs in the IR range? Is it as good as the Fuji IS Pro at IR photography and all that stuff?

I sorta got interested because the S15 is supposed to be coming out soon. Soon. :D

I have no idea what you are attempting to show there, ANR. Well, at least not the part which we were talking about. ;)

But, if you are shooting raw your last two legs appear wrong (if I am reading them right).
In raw 4 sensor photosites do not equal only one pixel, and the onboard camera processor is not the one which demosiaics multiple photosites into multiple pixels.

OK soap, let's try this again. I tried to draw myself a visual representation, is this "technically" correct?

I know it's a different sensor, I was referring to just the removal of the internal filter as can be done with almost any DSLR such as the D200. Yet for some reason the IS Pro cost much more than a regular S5 for having just no internal filter. That's why I was asking.

Totally different sensor than the D200, though, in that it has 12 megapixels worth of photosites, but really only 6 MP of effective output.

It has one large light sensor and one small one paired up. (One under each segment of the Bayer filter? (mb!)). This was their method to increase dynamic range.

What I don't understand, though, (mb!) is why they claim UV sensitivity. All the published specs go no higher than 380nm, the accepted limit of the classical "visible spectrum". Do typical digital cameras lowpass significantly below 380?

That's what I thought. Seems like an overpriced converted D200 than.

As for firmware demosaicing routines would be not only useless but would have to be disabled and you will either need changed firmware to produce proper image or you will have to shoot RAW and use specialized CaptureNX software for example.
AF will not work for IR or UV and that is why Fuji used Zeiss manual lens with distance scale markings for IR shooting.

All they did was to remove IR and Bayer CFA (Bayer CFA works well like UV by itself so I think D90 does not have one at all) and you do not need low pass AA filter in this case.
You therefore need 4 add on filters with such a camera, UV pass, UV block, IR pass and IR block to be able to shoot at all wavelengths and to differentiate between them. I hope this is clear enough and I also think we are getting to serous and technical here.
And yes Niko you are right, it is just a conversion of a regular camera such as D90, the only problem is that if you try to do it yourself you will most likely break the sensor because built in filters are fixed and glued to it.

NikoDoby said:
So why not just convert a regular D90 in the same way?

Assuming I am right (mb?) about it being simply a filter shift - how are the filters attached, and how does one lie to the firmware and the assumptions it is making about the light it is receiving?

EDIT: Inserted quote for context.

So why not just convert a regular D90 in the same way?

EDIT - scratch my earlier statement - I confused the physical (frickin' dual diode system) resolution.

And mb can make the definitive statement, but I'm >90% sure that, yes, all they did was move the (near) IR and (near) UV filters off-sensor.

soap said:
Whatever else the Fuji IS Pro has going for it, it actually has less resolving power, significantly less, than a stock D90.

Well yes obviously it has less resolution but I meant it can shoot visible,UV and IR. My question is how does it achieve this? I would hope there's more to it then just not having any filter over the sensor and then using various filters over the lens.

You can convert just about any digital camera to that.

NikoDoby said:
So mb isn't the Fuji IS Pro such a camera? I never got a clear answer on that?

Fuji IS Pro is such a camera to some extend, it is based on D200 body and Fuji own (today inferior) sensor, was very specialized (comes with with Zeiss lens and filters needed UV, IR ...) but was too expensive IMHO. Anyways it proves that there is a market for such a camera.

NikoDoby said:
So mb isn't the Fuji IS Pro such a camera? I never got a clear answer on that?

Whatever else the Fuji IS Pro has going for it, it actually has less resolving power, significantly less, than a stock D90.
~1700 lines per picture height compared to ~2200 for a stock D90 and ~2800 for our theoretical filterless model.
(Mind you the D40 with its 6MP sensor achieves ~1550)

So mb isn't the Fuji IS Pro such a camera? I never got a clear answer on that?

B&W P&S maybe of some interest but not for me, I need SLR camera to play with, but if it is B&W D90 like camera with 12 Mpix panchromatic sensor (meaning IR, UV and visible light as it would be without CFA) I think it would have the market.

Nah, I'm not a fan of the "middle ground" cameras. Give me a P&S or give me a SLR. I'm not a fan of those cameras too large to put easily in my jeans, yet without the advantages of a SLR.

soap said:
At $500 (an unreasonable expectation on my part) I would do it no doubt.

I'm just asking that at that price would anyone buy a canon G-series like camera if it was dedicated B&W?

NikoDoby said:
$500?!? So then would you buy a dedicated B&W point & shoot if it was something like a canon G11?

I don't see where I suggested that.

$500?!? So then would you buy a dedicated B&W point & shoot if it was something like a canon G11?

NikoDoby said:
But anyway back to the black & white D90. How many of you would actually buy one if such a camera existed? I know I wouldn't.

IF it had the AA filter removed as well as the Bayer filter AND it hit the theoretical 2848 lines per picture height (which would be greater than the D3x) AND achieved a one stop ISO improvement as a result of the filter removal, while staying ~$800 for the body, I would be very very tempted.

At $500 (an unreasonable expectation on my part) I would do it no doubt. At $1000 I would wait for used while watching how the performance of the next-generation of Bayered sensors measures up.

My wifey is the best! She takes care of me and corrects me when I have made a mistake and I'm not just saying that because she's standing behind me... :^)

But anyway back to the black & white D90. How many of you would actually buy one if such a camera existed? I know I wouldn't.