full size sensor from Canon?

[harpoon]

How long do we expect to wait before Canon comes out with a camer with a full size sensor, preferably at the price of a 10D? Any guesses?

[deztoys]

Isn't the 1Ds a full size sensor? But it certainly isn't the price of the 10D.

But if you want to be impressed (assuming # of pixels does it) check out the medium format digital backs. 20-40+ megapixel CCD/CMOS. Now that would be nice for $1500........ guess we'll have to dream on...

[openspace]

I would guess 2 years at the very least to get a "full size" sensor below $3000. Bragging rights aside, frankly I don't think producing a cheap "full size" sensor is a top priority for any of the major manufacturers.

Technologically speaking, it is generally much cheaper to increase sensor resolution than sensor size. And higher resolution is what the "people" demand - and therefore what the camera manufacturers like to shout about. Every digital camera at your local electronics store sports a bright red, yellow or blue sticker that shouts "I've got more megapixels than the next guy!" But very few talk about sensor size.

To be honest, the "cropped" sensor issue really isn't much of an issue for most photographers anyway - that is once you get used to it. You can easily compensate for the "crop" by buying a wide angle prime, or if need be, "stitching" photos. Beyond that, for most shooters, there really is little issue.

If you ask me, I'd rather have Canon spending their R&D dollars on increasing the "cropped" sensor resolution to 8 megapixels and beyond. That is a much more feasible goal than a cheap full size sensor.

[dbarthel]

See track at Rob Galbraith: http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Board=UBB8&Number=166045&page=0&view=collapsed&sb=5&o=&fpart=1

I continue to wonder what is the attraction of the full sized sensor. You will get more distortion, vignetting, etc at the edges, you will have to use much bigger and slower telephotos, you will have to use thin filters on wide angles, so in general you lose rather than gain with a full frame sensor. A few more megapixels, OK. A specialty wide angle with less that full frame coverage from Canon or Sigma, Ok. But, please don't be envious just because you can't afford a 1Ds. I can, and since I do nature and birds, I don't want one.

Dan

[DaveG]

Using the conventional photographic maxim, the larger the negative the better the image, of course I'd want a 24x36 mm CMOS!

I want to have a super wide angle again - my 16-35 can hardly wait! I want a 12 + megapixel camera so it replaces my medium format stuff.

And don't kid youself about the telephoto bonus you get right now with the 10D's conversion factor. You could do this with 35 mm simply by enlarging a part of the neg, and no one wanted to do that!

[dbarthel]

Ah, but the conventional maxim falls apart, as one can increase pixel density in the smaller area. Again, think about what you lose vs what the smaller sensor buys you. APS was a good concept, except the small negative couldn't deliver. The 10D and D60 as is deliver more than the most carefully processed film. So why sweat physicals size? Mine is bigger than yours?

[DaveG]

dbarthel wrote:
Ah, but the conventional maxim falls apart, as one can increase pixel density in the smaller area. Again, think about what you lose vs what the smaller sensor buys you. APS was a good concept, except the small negative couldn't deliver. The 10D and D60 as is deliver more than the most carefully processed film. So why sweat physicals size? Mine is bigger than yours?


Yes they can make the pixels smaller and get more on the 10D size chip. But they will get even more of these small pixels on the bigger chip which makes for a higher megapixel count and better images.

APS sucked by the way. The bigger 35 mm was much superior. The loading and the idiot proofing of APS was better but not the image quality.

[harpoon]

From the optical stand point (I happen to have optical engineer as major), dbarthel was correct on that the image quality near the edge is usually worse than that near the center. This is just how the lenses are designed. Lens designer constantly have to make compromises. They would rather sacrifice quality at the corner that at the center of the image.

However, dbarthel was not right on that so long as you can pack a high number of pixels in a small sensor it's all the same as a big sensor. For one, lens can focus only to a finite point. If the pixel is smaller than the size of the focal spot of the lens, the pixel size no longer matters. Some will say, you can design a good lens with very little aberrasion. But alas even for the ideal lens with no aberrasion of any kind at all, there exists a diffraction limit. The lens can still only focus into a finite spot size. The size of the spot depends on the diameter of the lens and the wavelength of the light. bigger diameter is better, short wavelength is better. This is also why blue laser DVD can carry multiple times the information as the red laser DVD we are using today. Blue light can be focused into a smaller spot (-> denser tracks) with the same lens. If your pixel is smaller than the diffraction limit, it no longer matters for any lens.

So cropping the sensor does two things to the lens. 1. it foregoes the image quality requirement to the edge of the image; 2. it imposes a higher requirement to the image quality at the center.

Two other side effects, 1. you lose wide angle, obviously. 2. when the conversion factor is very big, such as that of the Minolta 7i 28-200 lens (conversion factor 4), the lens is actually a 7-50 mm in 35 mm world. When shooting a portrait, even if you open the F stop all the way to 3.5, and set zoom at 200mm, it still is impossible to blur off the background, because you are, in reality, using a 50 mm lens.

[BrettD]

I too eagerly await an affordable (to me 10D pricing) full frame sensor. One of the major benefits of a full frame sensor, is that it will have larger pixels or more of them, either way the end result is much less visible noise at high ISO settings. And I get my 17-40 to be a real wide-angle lense.

Brett D

[dbarthel]

Harpoon, the topic of a fractional extender (0.6) came up on another forum. The idea is to compress the full frame image to the size of the sensor. Is this possible? Upsides?, Downsides? Good to hear from someone who knows his subject.

Dan

[harpoon]

To be quite honest with you dbarthel, I don't know what a fractional extender is. Do you have the link for that discussion? Thanks.

[dbarthel]

The concept is that instead of an extender which increases the focal length (magnifies the image), build one which decreases the effective focal length (shrinks the image) so that the full image will fit compressed on a small sensor. Very similar effect to a wide angle adaptor on the front of a lens, except this guy would sit behind the lens like a regular extender. Link at Rob Galbraith is http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Board=UBB8&Number=165112&page=0&view=collapsed&sb=5&o=&fpart=1

Am I explaining myself better? In numbers a .6 multilpier instead of a 1.4 or 2.o multiplier.

Dan

[pwagner]

Yes, you read that right... ZERO point six converter. As the previous poster suggested, such a device would turn the 10D into a full-frame camera.

I asked about such a system a while back; never have I been met with such uniform opinions on an online forum! Unfortunately, the opinion was that the system would suffer ALL the problems of the that the extenders suffer, namely: degradation of optical quality, loss of light, loss of aperature. Added to that is the possibility that some lenses might have severe focus range problems (i.e.: won't focus at infinity or have a fairly far minimum focus distance) Apparently there was a camera a few years back that had a built in 0.6; that camera was a disaster. (Don't know who made it.)

I was surprised that loss of light and aperature would happen. I mean.... the 2.0x loses more light and more aperature than the 1.6x, so I would have thought that a 0.6 would GAIN light and aperature. Oh well.

[harpoon]

I got the idea dbarthel. In theory it should work however I am not quite so hot on add-on optical components because I think it will compromise the image quality.

I did a projection TV lens design for my BS thesis. Through this work I realized more than ever that designing lens is a balancing act and a process of compromise. Each element will have its abbrasionb and distoration, that's simply physics. A lens designer needs to carefully select the curvatures, the material of different index and dispersion, and the relative positions, etc. to have the abrasions and distortions of different elements cancel out each other (you may find many lens, such as the classic double Gaussian used for most standard lens to consist of two halves that are almost symmetric to each other so the abrrasions are of opposite signs and thus cancel out). So the lens is always optimized as a whole with every one of the elements considered.
http://www.cartage.org.lb/en/themes/Sciences/Physics/Optics/Optical/Lens/lenspr16.gif

Introducing other one-size-fits-all optical components in my opinion tips off this carefully achieved balance and compromise image quality simply because those add-on elements were not considered in the original design.

After reading back what I have written I feel perhaps I should put a little disclaimer here. After all it's been more than 10 years since I did my last lens. Many things have changed in lens design since then. So perhaps with the new material, more powerful computer aided design a fractional extender can work just as well as original. But I won't believe it until I see the test result.