Perhaps Nikon has the right idea in not make a full frame DSLR

[James G. Dainis]

Zoom lenses often deliver loss of contrast and sharpness at the corners of a

print. With the Nikon, that lens fault would fall out side of the 23.7x15.5

sensor captures. The print would have sharpness and good contrast throughout.

<P>

 

With a full frame, one could make an 11x14 enlargement and crop out an 8x10

print but that would ruin the composition, the thirds would no longer be on

the thirds. With the APS sensor the cropping is done in camera with the

correct composition through the viewfinder. <P>

 

"This lens is a poor performer on my film camera but it does much better on my

D50."

[mark u]

On that argument, Nikon should have dropped all 35mm SLRs in favour of Proneas using APS-C film long ago. I think that Nikon has not produced a full frame DSLR so far simply because they have no supply of sensors they consider adequate. I've seen several hints that Sony (Nikon's principal sensor supplier) is working towards producing full frame sensors - perhaps notably the fact that they have comissioned lenses with 35mm coverage from Zeiss in KM/Alpha mount - and I'm sure that Nikon will be in on the act if it happens, as seems likely.

[mars c]

I'm willing to bet $1,000- about Nikon making FF body within the next 5 years ( or maybe less).

 

I'm on the side who believes they will.

[fred_c1]

>Zoom lenses often deliver loss of contrast and sharpness at the corners of a print. With the Nikon, that lens fault would fall out side of the 23.7x15.5 sensor captures.

 

Never heard about the DX Zoom-Nikkors with the reduced image circles, huh?

[mtwhite]

(Names withheld to protect the clueless.) A gentleman was actually trying to convince me that the F-mount wouldn't allow a full-frame sensor, which seems a little odd, since the F-mount seems fairly capable of allowing full-frame 35mm film.

[ellis_vener_photography]

That was no gentleman. That was an ignorant idiot.

[squareframe]

James, your hero-status is in jeopardy with such twaddle about thirds and in-camera cropping. perhaps this too, was Kodak's logic with their APS format that won the hearts and minds of all serious photographers worried about contrast and edge-definition?

[dmanthree]

Interesting logic. I guess we should all use view camera lenses and apply them to the smallest

sensor we can find so we're only using the lens' sweet spot. Sharpness, corner to corner.

 

Seriously, bigger sensor=more light captured, more data points, more detail, etc. Just like

film. Larger sheets of film produced better images, didn't/don't they? Why are sensors

different? Nikon can't manufacturere FF sensors, that's all, and they don't have a source that

can. Unless Canon decides to sell them one.

 

Right.

[KenPapai]

I agree, the question as posed is nonsense UNLESS you accept that Nikon is giving up the pro market to Canon. Canon is 2nd to none when it comes to super low noise FF sensors. Sony is likely a couple years away from making larger sensors to fit in cameras under $3K in cost.

 

The basic premise of the OP is difficult to comprehend. I wonder if the OP ever shot 35mm film and wished for smaller framed film that consumers love and pros eschew? Digital sensors are the same as film -- the larger the capture format the better the print.

[leicaglow]

David, The electronics industry is moving to higher densities, not larger chips. I'd bet we won't see a FF Nikon, but instead DSLRs with more pixel density.

 

Your analogy that it's just like film doesn't carry over to digital sensors because current film technologies have a theoretical limitation in the amount of grain in a given image area. That limitation for digital is broken all the time, as smaller sensors are introduced with significantly higher resolutions.

[wgpinc]

I have read that Nikon is using a CMOS sensor instead of the CCD in their new D2XS high end DSLR. We're all educated guessing here but my take is with the camp that says they don't have access to full frame sensors that are at a price point they think they can sell and make money on. JMNHO

[dmanthree]

Michael,

 

Have you seen the results of the latest high-density chips? They stink. High noise, muddled

detail despite the increase in resolution. Sensors are very similar to film with respect to

physical limitations; larger photosites can capture more light and lower the noise. the laws of

physics cannot be violated here just as with film, smaller photosites will decrease image

quality. And just like with film, a larger image capture area will yield higher quality images.

Unlike integrated circuits, photo sensors are getting larger, not smaller, except for P&s

cameras, which are getting progressively worse due to the megapixel race.

[squareframe]

I suspect that making large-sensor arrays is not a huge hurdle. incorporating them into a camera and <b>making money</b> in a competitive market ... most likely is.

[James G. Dainis]

I don't think I made myself very clear in my original post. Let me see if I can explain better with a photo. The image below is a scan of an 8x10 photo taken with a Nikon F3 and Kiron lens that I was testing out. <P>

<center><img src=http://www.geocities.com/dainisjg/leaves2.jpg></center><P>

 

Since this is an 8x10 from a full frame negative about 1 inch has been cropped off each side. One can still see the loss of contrast and sharpness in the corners. Indeed, it is much more pronounced in the original 8x10, quite blurry while the center is sharp. The red rectangle represents approximately what would have fallen on a Nikon D50 23.7 x 15.1 sensor, no fall off in the corners. <P>

 

If I zoomed in with the F3 to the area shown in the red rectangle, I would again get the fall off on the corners now of the red rectangular area, perhaps even more pronounced with the greater zoom. If I were to back up or zoom out with D50 to encompass the area shown in the original photo, there would again be no fall off in the corners, as the corner fall off of the lens would not be falling on the sensor.

[fred_c1]

Don't want to use something better than a Kiron zoom? Try looking up the Olypmus four-thirds system with an even smaller sensor than Nikon's and see what we mean.

[noahfactor]

James, I think the best thing to do (in terms of image quality) would be to use lenses with larger image circles, i.e., longer lenses, which is essentially all that lowering the size of the sensor is doing.

[kelly_flanigan1]

The concept of lens performance peaking on optical axis and dropping off axis is many centurys old, many centurys before photography was born. A 50mm lens used on a 16mm movie camera uses less angluar coverage, and has a tighter DOF criteria than a 35mm still cameras 50mm lens. This is clearly described in pre ww2 Kodak books on photography. Sensor size in area ie x by y mm is governed by costs, something few seem to grasp. The peanut gallery loves to ignore cost and yields of giant sensors, to fit their dogma of dreaming. They probably also think taxes will drop next year, wars will end, yugos with 500 cid V8's will get 100 mpg, and full frame senors will magically drop in pricing. Dreamers ideas are interesting when one ignores costs and real world yields.

[fred_c1]

<i>He told me that, in his opinion, all the calculations that would ever be needed in this country could be done on the three digital computers which were then being built -- one in Cambridge, one in Teddington, and one in Manchester. No one else, he said, would ever need machines of their own, or would be able to afford to buy them.

</i><P>

 

-- Lord Bowden quoting Professor Douglas Hartree of Cambridge, 1951, in <i>American Scientist</i>, vol 58 (1970).<P>

 

Back to the present: the Canon 5D, costing $3000 when first introduced a tad over a year ago, is now $2800 before a $300 rebate (which is doubled to $600 if another selected Canon product is bought at the same time.) <P>

 

I guess Nikon should sue Canon for dumping, huh?

[James G. Dainis]

The thing with most people is they don't understand the 1.5X focal length multiplier, that it is a cropping factor.

 

<P>I got ahold of the first full frame Nikon prototype from my cousin Sushi and it takes lousey pictures. The corners are blurry with terrible contrast."<P>

"It's probably that cheap lens you bought in Mexico."

 

<P>"No, the lens is fine. The pictures I take with it on my D50 have equal sharpness and contrast across the entire image."

[pete_flanagan1]

According to an associate of mine who is an engineer who works in the optics field, the problem with 24x36 sensors in 135 format bodies is edge to edge sharpness.

 

He explained that edge to edge sharpness is not a problem with film because the emulsion is pretty much on the surface of the film base.

 

However, the design of digicam sensors places the photosites embedded slightly below the surface of the sensor.

 

Therefore, the angle of attack of light rays at the edges of the short sides of the frame of a 24x36 sensor would be too shallow to provide for sharpness at those edges.

 

Of course, a 24x36 sensor is technically feasible. However, to get acceptable edge to edge sharpness, the sensor would have to be in a body large enough to accept lenses having the barrel diameter of an RB67 or RZ67 lens.

 

Folks I know who have 5Ds have remarked about their disappointment in the edge sharpness, but overlook it in light of other positive characteristics of the camera.

 

If the day comes that edge to edge sharpness can be maintained with a 135 format lens, then you will probalby see more 24x36 sensors in 135 sized bodies. However, based on Nikon's heavy investment in the "DX" format lenses, that probably ain't going to happen for a while, if ever.

 

So, to address a comment earlier in this thread, the guy was absolutely correct when he said Nikon F mount lenses would not work well with a 24x36 sensor.

 

Someone owes him an apology.

[pete_flanagan1]

Also, there is no such thing as a "focal length multiplier" or a "cropping factor."

 

Taking a 50mm lens off your 135 film camera and placing it on your DSLR neither changes its focal length, nor crops the image. Due to the smaller size of the sensor, the angle of view narrows, that's all.

 

I switch back and forth all the time between DSLRs and 135s and 6x7 cameras. I don't sit there and ponder multipliers or cropping factors. All one has to keep in mind is the "normal" focal length for the particular format they are shooting.

[dave_cheng1]

CCD sensor elements are most sensitive to light striking perpendicularly. When light falls from an angle the sensitivity is reduced and the dynamic range suffers. This is physics and is the main reason larger sensor is so difficult to produce. This is when engineers have been successful in making the sensor higher in density but not in CCD size.

[roger krueger]

<i>Pete Flanagan , dec 21, 2006; 03:31 p.m.

According to an associate of mine who is an engineer who works in the optics field, the

problem with 24x36 sensors in 135 format bodies is edge to edge sharpness.</i>

<br><br>

What problem? Yeah, FF is less forgiving of crummy lenses, but there's no more problem

with corner sharpness than there is with film. My Leica 19, Canon 50, 135, 200 and 400

are all quite sharp right out to the corner on my 1dsII. Only my 24/1.4 never really gets

great corners, but that fast and wide there are compromises.

<br><br>

<i>He explained that edge to edge sharpness is not a problem with film because the

emulsion is pretty much on the surface of the film base.

<br><br>

However, the design of digicam sensors places the photosites embedded slightly below

the surface of the sensor.

<br><br>

Therefore, the angle of attack of light rays at the edges of the short sides of the frame of a

24x36 sensor would be too shallow to provide for sharpness at those edges.</i>

<br><br>

Nonsense. People keep repeating this urban legend, but no one ever has any tests that

verify it. You would need to have light ending up in the wrong pixel to lose sharpness, and

that isn't happening. You do lose some light to this effect, and to microlenses being less

efficient off their optimized angle, but this solely affects corner brightness, NOT sharpness

<br><br>

My favorite proof of this is where a guy tested a Voigtlander 15 vs. a Sigma 12-24 on a

5D. The Voigtlander, a non-retrofocus design that protrudes deep into the body way, too

far to let the mirror function, has a vastly greater angle of incidence than any normal SLR

lens that clears the mirror. Yet it beats the snot out of the Sigma in the corners. And the

Sigma is no slouch either, it stood up quite nicely to the Nikon 15 prime.

<br>

<a href="http://www.16-9.net/lens_tests/15mm_test1.html">

Voigtlander 15 vs. Sigma 12-24</a>

<br><br>

<i>Of course, a 24x36 sensor is technically feasible. However, to get acceptable edge to

edge sharpness, the sensor would have to be in a body large enough to accept lenses

having the barrel diameter of an RB67 or RZ67 lens.</i>

<br><br>

Hogwash on several levels. Plenty of lenses already have great edge sharpness on FF. But

you're never going to get a lens as sharp in the corners as it is in the center, no matter

how much it covers.

<br><br>

The extra coverage of an MF lens does absolutely nothing to address the angle-of-

incidence issues. The fact that shorter MF lenses are going to be a lot more retrofocus

than their 35mm cousins--because they have a bigger mirror to clear--does address the

angle-of-incidence problem, but there's nothing about 35mm mounts that precludes the

making of 35mm format wides that are more retrofocus than necessary to clear the mirror.

It's just that no one (besides maybe Oly) has seen fit to actually do this. The real advantage

to MF lenses might be that no one makes crummy MF lenses, any modern MF lens is pretty

much "L" quality.

<br><br>

But even more blatently wrong, you don't need a bigger body or mount to use larger-

coverage lenses. You build the right adapter (including a helical and means of holding the

shutter open in the case of RB/RZ lenses) and the vast majority of MF lenses will function

just fine on any 35mm SLR mount. Any lens that can clear an MF SLR mirror is

exceptionally unlikely to get elements anywhere near the 35mm mount ring. The only

possible catch might be a long fast lens that has its principal point too far forward--that

is, it's "less telephoto"--might get vignetted by the mount. But long lenses are where you

see the least edge degradation anyway.

<br><br>

<i>Folks I know who have 5Ds have remarked about their disappointment in the edge

sharpness, but overlook it in light of other positive characteristics of the camera.</i>

<br><br>

Then they need to buy better glass, and be realistic about wide-open corner

performance.

<br><br>

<i>If the day comes that edge to edge sharpness can be maintained with a 135 format

lens, </i>

<br><br>

It came about 40 years ago for normals, 30 years ago for teles and non-retro wides, and

maybe 20 years ago for retrofocus wides. Although the only really edge-to-edge sharp

retrofocus ultrawides are German<br><br>

<i>then you will probalby see more 24x36 sensors in 135 sized bodies. </i>

<br><br>

You're not going to see many FF bodies anytime soon, because the large chunk of silicon

is going to be much more expensive, and have much worse yield for the forseeable

future. And for a lot of shooters, FF doesn't offer anything special. Anything over about

100mm equiv. the faster, shorter lens you can use on a crop camera exactly compensates

for the high-ISO-noise advantage of FF. But APS has no answer for the fast+wide of the

Canon 24/1.4 or the fast+dead sharp of the Contax 21 or the latest Leica 19. The latter

APS can have if someone wants to, but they'd need a 16/1.0 to equal FOV and light

gathering of the 24/1.4.

<br><br>

<i>However, based on Nikon's heavy investment in the "DX" format lenses, that probably

ain't going to happen for a while, if ever.</i>

<br><br>

While they've poured plenty of money into DX recently, they've still got a much larger

investment in FF designs.

<br><br>

<i>So, to address a comment earlier in this thread, the guy was absolutely correct when

he said Nikon F mount lenses would not work well with a 24x36 sensor.</i>

<br><br>

No, he was absolutely wrong. Zero reason for Nikon F-mount to have any more trouble

with FF than Canon EF. Remember, they made sucessful FF film cameras for decades?

<br><br>

<i>Someone owes him an apology.</i>

<br><br>

Someone--YOU in fact--are the one owing an apology here for spreading this nonsense

as if it were fact.

[dmanthree]

Very interesting observation regarding Nikon lenses, and their inability to cover a 24x36

image area. I guess my old F100, with it's 24x36 imaging area must have ben an anomoly.

 

And those Kodak FF cameras that used a FF sensor with the F mount must have been an

illusion. What a load of trash talk.

[James G. Dainis]

When I started this thread to post an interesting observation, I didn't realize I would touch off such a fire storm and misconceptions. <P>

 

<center><img src=http://www.geocities.com/dainisjg/tree3.jpg></center>

 

Lenses throw circles of light, they do not throw rectangular or square forms. The image on the left shows how much of that circle falls on a 35mm film or full frame sensor. The circle on the right shows how much falls on a digital camera sensor. The circular segments are cropped off in camera either by the shutter window in 35mm or by the amount that falls on the film in MF. How much is cropped off is the cropping factor. 1.5X more is cropped off on the D50 than on 35mm film or full frame sensor.One doesn't have to physically do anything to crop. I can raise my enlarger to display an 11x14 image on the easel and stick an 8x10 photo paper under it to crop out an 8x10 section. I don't have to make an 11x14 pint and use a paper cutter to crop out an 8x10.

 

<P>The angle of view of a lens is an optical characteristic. There is no electrical contact in the mount that rearranges the lens elements when the lens is put on a digital camera. The angle of view of the lens stays the same; what changes is how much of that view falls on the sensor.