New generation of "digital" lenses: only a commercial strategy?

Started Jun 28, 2001 | Discussions
Pier Luigi Regular Member • Posts: 143
New generation of "digital" lenses: only a commercial strategy?

Recently manufacturers are all offering new lenses projected for digital SLRs.

Do you think that this Lenses really perfom better than the old good ones with Digital SLR? Could this be only a commercial move ?--Pier Luigi

Davidlw Senior Member • Posts: 1,827
Re: New generation of "digital" lenses: only a commercial strategy?

Well I dont fully understand the mechanics but I think that it has something with the way that light must fall on a chip as compared to film. So lenses designed for digital cameras in theory should be better for that situation.

Pier Luigi wrote:

Recently manufacturers are all offering new lenses projected for
digital SLRs.
Do you think that this Lenses really perfom better than the old
good ones with Digital SLR? Could this be only a commercial move ?
--
Pier Luigi

Bob Klein New Member • Posts: 5
Re: New generation of "digital" lenses: only a commercial strategy?

David wrote:

Well I dont fully understand the mechanics but I think that it has
something with the way that light must fall on a chip as compared
to film. So lenses designed for digital cameras in theory should be
better for that situation.

David is correct. The problem is with wide angle lenses. The light rays at the edges of the image strike the sensor (CCD or CMOS) at an extreme angle. The currently available 14mm lens for 35mm film cameras is a 21 for digital. Those of us who shoot film have been spoiled by zooms down to 17mm. Its not currently possible to design a non-fisheye lens wider than than the 14mm. But something like a 10mm lens could be designed if it only had to service the pro digital SLR.

OP Pier Luigi Regular Member • Posts: 143
Re: New generation of "digital" lenses: only a commercial strategy?

Well, so guys.. one more reason (I already was oriented in this direction).. to wait for CCDs or CMOS with 1 : 1 correspondence with 35mm format

In this case (1:1 coverage) should a good old generation lens perform like a new "digital" lens ?

Pier Luigi

Bob Klein wrote:

David wrote:

Well I dont fully understand the mechanics but I think that it has
something with the way that light must fall on a chip as compared
to film. So lenses designed for digital cameras in theory should be
better for that situation.

David is correct. The problem is with wide angle lenses. The light
rays at the edges of the image strike the sensor (CCD or CMOS) at
an extreme angle. The currently available 14mm lens for 35mm film
cameras is a 21 for digital. Those of us who shoot film have been
spoiled by zooms down to 17mm. Its not currently possible to design
a non-fisheye lens wider than than the 14mm. But something like a
10mm lens could be designed if it only had to service the pro
digital SLR.

Weegee
Weegee Veteran Member • Posts: 3,016
Re: New generation of "digital" lenses: only a commercial strategy?

Pier Luigi wrote:
Well, so guys.. one more reason (I already was oriented in this
direction).. to wait for CCDs or CMOS with 1 : 1 correspondence
with 35mm format

In this case (1:1 coverage) should a good old generation lens
perform like a new "digital" lens ?

Pier Luigi

Bob Klein wrote:

David wrote:

Well I dont fully understand the mechanics but I think that it has
something with the way that light must fall on a chip as compared
to film. So lenses designed for digital cameras in theory should be
better for that situation.

David is correct. The problem is with wide angle lenses. The light
rays at the edges of the image strike the sensor (CCD or CMOS) at
an extreme angle. The currently available 14mm lens for 35mm film
cameras is a 21 for digital. Those of us who shoot film have been
spoiled by zooms down to 17mm. Its not currently possible to design
a non-fisheye lens wider than than the 14mm. But something like a
10mm lens could be designed if it only had to service the pro
digital SLR.

Luigi,

The train is moving away from the station ever faster! Time to get on because the next train is a long time off. By the time it arrives you won't know how to make it run.

Ger Bee Forum Pro • Posts: 11,470
Re: New generation of "digital" lenses: only a commercial strategy?

In this case (1:1 coverage) should a good old generation lens
perform like a new "digital" lens ?

Pier Luigi

No it won't - or all things working properly - then the old lenses are still too soft for digital. The photosites work on agitation which responds to wavelengths that are "invisible" on film and thus the CCD records more information which has the effect of adding to the file’s non-image forming light (on film) and effectively softens the image.

New generation lenses need to be cut to a much finer tolerance to eliminate these wavelengths and not to add to the vibrations and thus not get recorded as something that gets interpolated into the image – as was known about with film but was not recorded.

Film needs direct light to cause a reaction in a grain of silver and an accelerator to enhance the effect. CCDs convert all light into signal data (image/stray light/noise). These need to be recorded at the perpendicular for the cleanest images, any deviation cause the “purple fringing” or otherwise incorrectly know as “CA”.

Thom Hogan
Thom Hogan Forum Pro • Posts: 13,661
Re: New generation of "digital" lenses: only a commercial strategy?

No it won't - or all things working properly - then the old lenses
are still too soft for digital. The photosites work on agitation
which responds to wavelengths that are "invisible" on film and thus
the CCD records more information which has the effect of adding to
the file’s non-image forming light (on film) and effectively
softens the image.

Say what? While the response curves for CCD sensors and film (and for that matter, between any two types of film) aren't the same, They're much more alike than different.

Your assertion that a "CCD records more information" is a strange one. Depending upon the filter array used over the sensor, a digital camera may see slightly more infrared and UV energy than some film, but if so, it doesn't add information, it adds exposure. Is it your assertion that existing lenses don't properly focus those wavelengths on the correct photosite? If so, it would have to be a pretty large error, as slight color aberrations tend to fall between photosites, not on them. And frankly, I haven't seen this problem or know of anyone who has. Or is it your assertion that the extra light a photosite reacts to (compared to film) overexposes that site and overexposure looks softer than correct exposure? Again, not something I've seen in working with digital cameras.

New generation lenses need to be cut to a much finer tolerance to
eliminate these wavelengths and not to add to the vibrations and
thus not get recorded as something that gets interpolated into the
image – as was known about with film but was not recorded.

Nonsense. If a lens is made poor enough that it is focusing different wavelengths at different points (chromatic aberration), that will show up on film, too. And what the heck is "vibrations...get recorded...[as] interpolation"?

Film needs direct light to cause a reaction in a grain of silver
and an accelerator to enhance the effect. CCDs convert all light
into signal data (image/stray light/noise). These need to be
recorded at the perpendicular for the cleanest images, any
deviation cause the “purple fringing” or otherwise incorrectly know
as “CA”.

No. The purple fringing so prevalent on consumer and prosumer digital cameras in very high contrast areas is generally caused by spillover electrons from adjacent photosites and has only a little to do with the lens used. Many reviewers have automatically chalked this up to chromatic aberration, but it rarely is. Chromatic aberration has a distinct pattern to it that relates to the radius from the center of the lens, the purple spillover can occur virtually anywhere, on axis or off.

On D1, D30, S1, and other interchangeable lens digital cameras, the current range of lenses are plenty sharp enough (some built-in consumer zooms show some interesting design problems, though). Even without interpolation (in fact, especially without interpolation), digital images are slightly soft because of the inherent limitations of converting a continuous value (analog amount of light) into a highly limited value (8-bit or 12-bit digital value). Both the "pitch" of the photosites as well as the "depth" of information stored influence apparent sharpness far more than any lens flaw. As D1x users are seeing, the method of interpolation used on raw photosite data also has a tremendous visual impact (each new version of Bibble and QImage Pro seem to get slightly better at interpreting the data from the D1x, but Nikon's own interpolation scheme built into the camera for the JPEG drivers still seems to beat them).

Thom Hogan
author, The Nikon Field Guide
http://www.bythom.com

Ger Bee Forum Pro • Posts: 11,470
Re: New generation of "digital" lenses: only a commercial strategy?

Thom Hogan
author, The Nikon Field Guide
http://www.bythom.com

I'll have this digest this reply - factual it may be and may contain some corrections for my own post.

Bottom line you are not fully correct either -- the newer lenses are sharper and visibly so with the D1X camera.

This has been reported many, many times - If I understand you reply to me this should not be the case.

Oh and just why does the spillover fall of the adjacent pixels in the first place? The photosites get excited and vibrate and all the wavelength can excite the photosite from any angle but a part of this is outside the visible spectrum needed to cause a reaction (or fast reaction) in film.

Everything gets interpolated as noise is filtered and replaced with “image” information, as you know only a small amount the vast quantity of information captured by the CCD is actually used.

In the CCD capture the element of the “none image forming” light in film is captured, this is not focused as you say but it is interpreted by the digital processing systems as valid image data to be retrained and not discarded and leading to an image softness.

This softness can be seen with different lenses from the same manufacturer – by whatever means it exists. Now I am, obviously open to correction, as, as you have stated other agencies have printed incorrect information – I rely on these for my own education and as best as I can understand it this is why the older lenses are softer, simply because the film cannot see the wavelengths that produce a “soft” edge to digital images (because the CCD can see these wavelength and records them).

Thom Hogan
Thom Hogan Forum Pro • Posts: 13,661
Re: New generation of "digital" lenses: only a commercial strategy?

Bottom line you are not fully correct either -- the newer lenses
are sharper and visibly so with the D1X camera.

"The newer lenses" is vague. Be specific. Which lenses? Compared to which lenses? Having used Nikon for almost 40 years now, I've got a wide variety of older and newer lenses, and I just don't see what you claim. The lenses that were sharp on my film cameras are sharp on my digital cameras; the ones that were soft with film are soft with digital.

This has been reported many, many times - If I understand you reply
to me this should not be the case.

Again. Please give me specifics to respond to. "Reported many, many times" by whom? Which lenses? Under what conditions?

Oh and just why does the spillover fall of the adjacent pixels in
the first place?

Well, let me quote from The Manual of Photography, 9th Edition: "One element forms a 'well' for charge but its capacity is limited and excess electrons must be removed by a 'drain' circuit otherwise they will migrate to adjacent wells and cause degradation to the image seen as 'blooming' and noise." In practice, we're still a fair ways away from the perfect chip in this respect. Add to this the fact that CCDs are "read" by row shifting the data across the adjacent photosites to the edge, where the data is actually looked at one pixel at a time and converted from A to D, and there's plenty of opportunity for stray electrons to influence adjacent data points (and if your crystal isn't accurate, you'll get a regular pattern of noise--something we all saw in the early D1s).

The photosites get excited and vibrate and all the
wavelength can excite the photosite from any angle but a part of
this is outside the visible spectrum needed to cause a reaction (or
fast reaction) in film.

I don't understand that sentence at all. First, the photosites are physically stable (they don't vibrate). The electrons passed through the sites are active, and can move from well to well, but I'm not sure I'd call this vibration (the proper term, I believe, is excitation). Second, a filter array sits on top of the actual CCD, so every photosite is exposed to a different set of wavelengths. Finally, that same array filters out most of the light outside the visible spectrum. Again from the Manual of Photography: "The spectral sensitivity of a CCD array is that of silicon and similar to that of a silicon photodiode as used in a light meter, with response extending to some 1000nm, with a peak at about 750nm in the infrared...Some cameras use an interference filter behind or in front of the lens to reject IR..." In practice, virtually every manufacturer puts an infrared filter in place over the sensor (Kodak, I believe, is the only one that allows this to be removed).

Everything gets interpolated as noise is filtered and replaced with
“image” information, as you know only a small amount the vast
quantity of information captured by the CCD is actually used.

Huh? You've got it backwards. All current digital cameras ADD information (the "interpolation" process) to that which was captured by the camera. Again from the Manual of Photography: "The use of a Bayer filter array assigns a group of four pixels to one piece of colour information so an image 'point' is effectively 2 x 2 pixels [photosites], theoretically reducing resolution by a factor of four, but interpolation gives the best estimate of the missing values."

In the CCD capture the element of the “none image forming” light in
film is captured, this is not focused as you say but it is
interpreted by the digital processing systems as valid image data
to be retrained and not discarded and leading to an image softness.

I would like to see some attribution of source for this contention. In none of my references, nor in any of my experience in working with digital camera designers, have I ever heard such a claim.

This softness can be seen with different lenses from the same
manufacturer – by whatever means it exists. Now I am, obviously
open to correction, as, as you have stated other agencies have
printed incorrect information – I rely on these for my own
education and as best as I can understand it this is why the older
lenses are softer, simply because the film cannot see the
wavelengths that produce a “soft” edge to digital images (because
the CCD can see these wavelength and records them).

The light source would make a difference in your contention, as well. The spectral curves of incandescent, fluorescent, and sunlight are all quite different, and have energies that vary considerable. Mid-day sunlight, for example, has twice the spectral power at 400nm that incadescent light does, while incandescent has over twice spectral power at 700mm than sunlight.

You know, there's actually a pretty easy method to test your contention: simply put two filters in front of the "older unsharp" lens that cut off all wavelengths below 390nm and above 700nm (the last digital camera I worked on used a filter that started at 800nm, if I remember correctly, and most digital cameras are still slightly responsive to infrared, so they must be filtering in the same area). I doubt you'd see any difference in image sharpness. In fact, because infrared has less photon energy than ultraviolet, you probably only need to cut off the response below 390nm to verify your contention.

Finally, as Manual of Photography says "The sensitivity of photographic [film] materials to scattered UV radiation...causes a loss of contrast...increasing with increasing subject distance." That's right, film responds to energy outside the visible spectrum (though if you compare Ektachrome 100VS to the Ektachromes of 10 years ago, you'll see a vastly lower response to UV).

In short, until someone can explain the mechanics of how it is possible for light outside the visible range (and outside the range that the ubiquitous IR filter would let get to the array!), I'd say that if you see a difference in sharpness between a newer and an older lens, it is most likely because the newer one is sharper.

Thom Hogan
author, Nikon Field Guide
http://www.bythom.com

p.s. While I was re-reading relevent sections of some of my source material, it struck me that if you're seeing differences, they may be due to flare, not nonvisible spectrum energy.

Keyboard shortcuts:
FForum MMy threads