Why some old Nikon FF lenses no good for digital FF

[roger_s]

I have read the comment, several times, that some older Nikon and Nikkor lenses

are not as suitable for digital, either DX format, and now FX full-frame

format. Or least, they do not perform as well for digital as they do for film.

 

The question is somewhat more relevant now, with the news of the D3 FX full-

frame camera.

 

If a lens is sharp for film, why wouldn't it be equally sharp for digital?

 

Are there good physics reasons for this, or is it just subjective?

 

If we have old Nikon/Nikkor lenses, how do we test for their suitability for

digital? Eye-balling the results would be merely subjective, in my view.

 

I have a swagful of old Nikkor AI-S manual lenses, 28/2.8, 35/2, 50/1.4,

105/2.5 and a Nikon 24-50 zoom, and am wondering how these would be useful for

the FX full-frame format ... not that I'm going to rush out to get a D3 at its

price. Rather, I'd wait as long as it takes for a reasonably priced prosumer FX

model. I'm very happy with my D80 and 18-200 DX-VR and 12-24 DX.

[sammurphy]

without knowing more specific details on the fx sensor in the d3, it would be hard to say.

 

however, there are some general comments that can be made on the light gathering abilities of digital sensors with regard to the angle light hits the film/sensor. in very brief and crude terms film can be exposed by light hitting the film plane at very sharp angles, where as digital prefers a more perpendicular light source. what this means is that wide angle lenses tend not to perform well into the corners on digital, even cropped dx frame digital. there is of course also greater chance for internal reflections affecting contrast and increasing the incidence of other abberations to a greater extent due to the more reflective nature of digital sensors compared to film.

 

leica, in designing the m8, tried to counteract this corner falloff by angling micro lenses at the sensors edge into the center to help the sensor register that light. they also implemented a software based vignette control that applies a correction to the image based on which lens is being used. similar in what nikon has said they are doing with automatic correction of CA's, i think. but leica's, or more generally, rangefinders, have much shorter lens flange distances compared to slrs, and so will be more prone to these corner effects, i think.

 

there is no way of predicting if your lenses will work well or not. some old lenses work really well on digital, and others are dogs. however, even those old lenses that work well on say a d200, won't necessarily work well on the d3, and visa-versa, dogs may become gems.

[tonybeach]

How are your lenses working on the D80? My experience with the 28/2.8 and 35/2 was that they were okay, but not spectacular; my 17-35/2.8 beats every prime I've compared it to except for flare issues. The 105/2.5 was nice too, but I saw no compelling reason to keep it since I have a 70-200/2.8 VR. Regarding the new FX camera and old lenses, it would appear that even my excellent 17-35/2.8 may come up short compared to the latest 14-24/2.8, which is why I think Nikon is offering the new lens. I don't know if this is the case, but in coming months we'll know more.

 

I'm no expert on these matters. What I do know is that photosites that collect light at more acute angles collect less light than photosites that collect light that is coming directly towards them (a phenomenon referred to as telecentricity, or lack thereof); film detects light equally regardless of the direction it is coming from because it is a flat surface. The problem might be solved by changing the angle of the microlenses that cover the photosites at the edges, but I suppose that can only go so far before it becomes deleterious to the performance of telephoto lenses which are inherently telecentric.

 

Another issue is chromatic aberration (CA). I suspect that the Bayer Filter Array magnifies CA coming from the lens because any focusing differences in the various wavelengths of light become more pronounced when they fall on separate blue and red sensitive photosites. The solution is modern lens designs optimized to minimize the problem using apochromatic doublets. There are also algorithms utilized by some software programs that can further minimize CA, and those are going to be incorporated into the D3 and D300 ASIC for out of camera JPEGs.

[frank_skomial]

"I have read the comment" - nonsense! good lenses of film camera are also good on digital for current line of cameras and pixel count/ pixel density Nikon has. This could possibly change in the future with pixel count in range 20+ MP or higher pixel density.

 

another comment on "dogs may become gems" - keep dreaming...

[ilkka_nissila]

A lot of older (and even some current) wide angle lenses perform badly on digital, to the point of being of no use. My 20mm AF-D is an example where a purple fringe of about 10 pixels (!) wide appears at the edge of the frame, creating a softness at the peripheral areas of the image which is unacceptable. This is by no means an exception. Reserve these lenses for use on 35mm film !!

 

For DX the solution is getting either the 17-55 DX, the 17-35, or the 12-24. I don't know which wide angle would work well on the D3 but I would expect the 17-35 to be one, and also consider the new lenses just introduced by Nikon. I've found all of my prime lenses 35mm and up to perform excellently on digital.

[bjørn rørslett]

"good lenses of film camera are also good on digital for current line of cameras" - this kind of sweeping generalisation cannot be made, since examples to the contrary are common.

[heartyfisher]

Actually I think that there is some truth to it..

 

I think it relates to 2 issues. both of which is due to the highly reflective sensor

 

1) Some of the material at the film end of the lens may be more reflective in the older lenses causing the light reflected back from the sensor to bounce around in side the camera and cause a loss of contrast.

 

2) The back end of the lens is flat and not coated well so it reflects light back onto the sensor and back again and again resulting in weird Flare like artifacts. The new Nikkor lenses have a meniscus end so that any light reflected back from the sensor is deflected away from the sensor I haven't check out the sigma DG lenses (VS non DG) but I assume that they do the same.. maybe some one with sigma lenses can have a look at the camera end of their lens and report back here?

[sammurphy]

Frank, did you read my whole post? maybe it was ambiguous, but i was saying that lenses which were great on film, ie. 2.8/28 ais, but which are not so great on digital cameras currently available, may perform much better on the d3, but like i said, with so little information on the specifics of the d3 fx sensor, who can say? which is what the OP was asking, "If we have old Nikon/Nikkor lenses, how do we test for their suitability for digital?" Test when you get your hands on a d3, if they're good, use 'em, if they suck, sell 'em.

[frank_yang]

I think a lot of this is subjective unless someone use measruing equipment to justified any meaningfull conclution. I think that most people are buying into tech talk by manufactures that if they don't use digital only lenses, they'll produce crapy photo's.

 

If I have to zoom in 20x in a photo's corner on my screen to see the difference...even if there is a difference...to me it lost all photographic meaning, only technical masturbation purpose.

 

For me the fun part is mounting those old lenses and see how it project its personality onto the digital sensor and each lens will have its "signature". Remember the glow those old single coated lenses produce? There are so many softwares on the market to correct pretty much any conceivable variables.

 

The fast prime AIS Nikkors like 50mm f1.2, 24mm f2, 28mm f2, 35mm f1.4 and 85mm f1.4 will find its way back to many photog's bag once D3 has gain wide acceptance cause there is no comparables on Nikon's digital lens line up.

[ellis_vener_photography]

<I>If a lens is sharp for film, why wouldn't it be equally sharp for digital?</I>

 

<P>If by "sharp" you mean the lens' ability to resolve detail, that is only part of the story. The lens also have to be near apochromatic (meaning they focus all frequencies of light on the same plane) and some lenses have problems with color fringing. Film was more forgiving in this respect. Mostly these are problems with lenses that use older retro-telefocus designs, or in other words wide angles, and some of them were not so great on film either.

[frank_skomial]

"which are not so great on digital cameras currently available, may perform much better on the d3," - this is highly unlikely, while perhaps possible due to large pixel size on D3.

 

In general I would say that if current DSLR reveals a lens deficiences (e.g. D2X), the D3 would be even more obvious. However the D3 has larger pixel size/pitch than e.g. D2X, so perhaps it is possible a lens can work better on D3 than D2X.

[benniehoff]

I understand the lack of telecentricity in wide angle lenses causing vignetting on digital sensors; but what exactly are the physical reasons for color fringing becoming more pronounced? Ellis, you say "film was more forgiving", but for what physical reasons?

[bjørn rørslett]

For film, you actually put any chromatic error (in particular of wide angles) to good use. Remember film had colour dyes layered deep within the film base. Thus a slight degree of CA might activate the coupling dyes better when light rays had an incidence angle away from the normal. For digital, "depth" of the sensor vanishes since the Bayer matrix essentially is on the surface. Now you rather want to have as little CA as possible, otherwise the neighbouring pixels could be triggered and lead to false colours and loss of contrast and fine detial.

[m-k]

As I understand it in crude terms, it's because film is flat, and digital sensors are not. Their three-dimensional honeycomb surface will exaggerate any chromatic aberration -- especially true with wide angle lenses, the light rays of which are hitting the sensor at an oblique angle (esp. in corners).

[ilkka_nissila]

Frank, high pixel density is what brings out defects in a lens or the unideal combination of the optics in the sensor and the main lens. The D3 has over twice as big pixels (in terms of area) than the D2X. Thus the D3 is likely to be far less demanding of the lenses (in the center) than the D2X. At edges, it depends on the lens design. In addition, as time goes on, Nikon learns more about how to design its sensors so that a wider range of lenses work well with them. Ideally that is.

 

For example I would expect the D3 to give far superior performance for the 50mm f/1.4 wide open than the D2X gives on either the 50mm or the 35mm (50mm equivalent) lenses at that aperture.

[ShunCheung]

To Michael Kackman, no, it is the other way around. Except for the Faveon sensor that is only used on Sigma DSLRs, digital sensors are flat. Film has separate RGB layers. Therefore, as Bjorn says, if you have the right type of chromatic aberration so that RGB don't focus on exactly the same surface, but instead focus separately correspoinding to those film layers, it actually works better for film. But that doesn't work on the flat digital sensor.

[m-k]

Ah -- very interesting. Thanks!

[harvey_edelstein1]

Can't CA be corrected easily in post processing if its noticeable? If so then the fast AIS lenses are good value for the money and may even reverse the trend towards f2.8 zooms.