FWIW - film, sensors and glass

[jcolwell]

This is free advice and it's worth exactly what it cost - nothing

(well, it cost me considerably more to start my journey into digital

photography, but that's a different story).

 

I believe that full-frame 35mm digital sensors (24x36mm) will become

common and affordable in the next 3 to 5 years, and so I will not be

buying any APS-C sized image circle lenses. Of course, I will

probably eat these words, as the EF-S 10-22/3.5-4.5 USM is developing

some kind of strange allure, I only hope that I can resist long

enough...

 

In the short term (2 to 3 years) I will continue to use 35mm colour

transparency (slide) film for (i) ultra-wide angles, (ii) situations

where the digital sensor can cause flare inside the lens, (iii)

images with both really dark and really bright parts, and (iv) when

I'm in my sea kayak on salt water. I expect that the full-frame EOS

dSLR which I plan to buy in about 3 years will solve the first three

of these problems. I will probably continue shooting B&W and IR 35mm

film for as long as I can lift my Pentax SuperProgram with SMC Pentax

35/3.5, but I have stopped using colour negative (print) film. I

will also continue shooting slides, B&W and IR on 120 roll film with

my 6x7 Fuji GW670ii (and maybe a GSW690ii/iii some day), for as long

as these films are available. BTW, I know that colour print film is

better than slides for handling problem (iii) above, but I just can't

scan any more colour negs - it's killing me.

 

I also think that the image sensors in your typical dSLR body will

not last for nearly as long as the glass in your lenses. Some day in

the not-so-distant future, current dSLR users will find that their

cameras need to be replaced because of aging effects on the image

sensor. I don't know when it will happen, but it will (I'm guessing

10 to 15 years); and when it does, I'm pretty sure that full-frame

35mm image sensors will still be here - I'm not so sure about APS-C.

Of course, many users will need new dSLR bodies before this time

because of cumulative damage to the image sensor from dirt and stuff,

and from trying to clean off the dirt and stuff.

 

We are still in the early years of digital photography. The

capability of current prosumer dSLRs is good enough that droves of

serious photographers are making the switch (I now use a 20D, Elan,

Elan 7 and a bunch of EF lenses). The capability of current pro

dSLRs is already amazing (see the EOS D1s MkII), and it will be

astonishing in a few years. Many current technical problems with

dSLR technology (such as sensor flare, no off-the-sensor flash

metering, and saturation) are just that: technical problems for which

engineering solutions are being developed. The science part was done

a long time ago.

 

For the next 5 to 10 years (at the most), digital photography

technology will continue to follow Moore's Law. Then it will take a

big jump when quantum processors and holographic memory hit the

shelves, right next to the computer games for which they are being

developed.

 

Out of breath for now, Jim.

[matt_mead]

I, too, believe that full frame digital sensors will become the norm. It's too confusing right now with the mix of full frame and 1.6x sensors and EF and EF-S lenses. I have to scratch my head when I try to think of reasons why manufacturers decided to go with APS-C sized sensors in DSLRs.

 

<p>After having shot P&S digital for years, and also viewing thousands of digital images taken with DSLRs, I think digital imaging is still in its infancy. It's just not quite there with lifelike image recording. My film looks much more natural, I just wish my scanner could capture the colors more accurately. Maybe that's why you have banned the use of color negs and stuck to slides.

[jon_austin]

I tend to agree with you, Jim. I'm digital only, though; wouldn't be doing any photography at all if it weren't for digital.

 

And Matt, I'm nobody's pro or highly qualified amatuer, but for the most part I'm very pleased with the quality of images that come out of my 20D. The less than excellent images are maybe 95% the inadequacy of my skills / 5% the immaturity of the dSLR technology and 5% just plain difficult photographic circumstances.

[aaron_lam]

Is there any benefit the sensor being bigger other than the lens performing like it did on a 35mm film camera?

[mark_chappell]

<I>(ii) situations where the digital sensor can cause flare inside the lens</i><P>

 

? Is this REALLY a problem? I've heard the potential discussed but in the thousands of

digital images I've made

I've never seen anything even hinting at such an issue -- and I do quite a lot of

photography in strong backlight and other situations where there are very bright spots in

the image.<P>

 

<I> (iii) images with both really dark and really bright parts</i><P>

 

Here, I think you're making a BIG mistake (if indeed you use slide film). I've never used a

slide film that had as much tolerance of tonal range as a DSLR. The proviso is that you

need to expose for the bright areas. I've had vastly greater success in resolving shadow

detail in digital images than with Kodachromes, Ektachromes, or various Fuji emulsions.

Color negative film may be a different story, but I've not used it for decades.

[jcolwell]

An example of (ii) situations where the digital sensor can cause flare inside the lens - there was no flare apparent in the finder<div></div>

[jcolwell]

"...a BIG mistake (if indeed you use slide film)...". An interesting comment - I've been using slide film since the '70's. I'm inclined to say "kiss my butt", but that would be rude. BTW, how much time do you spend in front of the computer "resolving shadow detail in digital images" ? I'd rather be taking photos.

[aaron_lam]

Well, one could also say... how much time do you spend in darkroom or in front of a light table. Dynamic range is an issue with both film and digital and it probably just takes just as much time to "fix" it on a print as it does digitally. Personally, I think fixing it digitally is easier, cleaner and faster.

 

I have never had any problems with sensor flare either. I find it hard to believe that you took that shot and there was no indication of flare in the viewfinder... you are pointed directly at the sun!

[mark_chappell]

<I>- I've been using slide film since the '70's. I'm inclined to say "kiss my butt", but that

would be rude. BTW, how much time do you spend in front of the computer "resolving

shadow detail in digital images" ?</i><P>

 

... and I've used slide film since the '60s AND I've used digital for several years. How

much time do I spend in

front of the computer 'resolving shadow detail'? Zero, for most images. It's right there in

front of you -- unlike with slide films that are challenged with simultaneous bright areas

and deep shade. For digital images with really deep shadows, maybe 15 seconds to bring

out detail that would be lost in an utterly featureless black in a transparency.<P>

 

With respect to your previous post: absence of visible flare in the finder is not a

completely reliable indication that it isn't there. There have been plenty of occasions

when I saw no flare and found it later <B>on the transparency</b>. I'm not saying

reflections off the digital sensor don't occur, just that I've not seen this phenomenon in my

own digital images, even in some that contain a direct view of the sun.<P>

 

And why the snide reply? I just offered my opinion that DSLRs do a lot better than slide

films in dynamic range, based on considerable experience with both. You want to ignore

what I think is a definite advantage of digital over transparencies, feel free. No skin off my

teeth.

[madhu1]

"I'm inclined to say "kiss my butt", but that would be rude."

 

Yes, that sure would be. But so is what you've said.

 

"BTW, how much time do you spend in front of the computer "resolving shadow detail in digital images" ? I'd rather be taking photos."

 

And having read what you had to say, and having seen your photos, I'd rather you first learn how to.

 

Jeeez!

[jcolwell]

My snide reply was prompted by your comment "(if indeed you use slide film)" in response to my statement that I use slide film. If you can't figure out why your comment might be interpreted as an insult, then perhaps you should consider taking some training to help improve your language skills.

 

Your comments on the relative capabilities of slide film and digital sensors to deal with images having a large dynamic range are appreciated, but are not especially relevant to the main points that I have raised. If you don't believe that sensor flare ever occurs, then I'm quite certain that digital photos will never convince you.

[mark_chappell]

Well, it wasn't meant that way, sorry. And if you read my last comment you will note that I'm

not denying that sensors can reflect -- just that I haven't seen it in images were it should

occur.

[daniel_taylor]

"I believe that full-frame 35mm digital sensors (24x36mm) will become common and affordable in the next 3 to 5 years, and so I will not be buying any APS-C sized image circle lenses."

 

I won't say it can't happen, but I doubt it will happen. Full frame is expensive because yields drop rapidly as die size increases, and Moore's Law doesn't change that. The price of the 1Ds hasn't budged despite 3 years and a model revision. Kodak is abandoning the 14 series and nobody else is even attempting full frame. IC manufacturing would have to improve dramatically to make full frame common. It will happen, but probably further down the road.

 

"In the short term (2 to 3 years) I will continue to use 35mm colour transparency (slide) film for (i) ultra-wide angles, (ii) situations where the digital sensor can cause flare inside the lens, (iii) images with both really dark and really bright parts, and (iv) when I'm in my sea kayak on salt water."

 

Regarding ii - With all due respect, the image you posted proves nothing, regardless of whether you could see the flare in the finder. Take shots of an identical scene using an identical lens with the light source at an identical angle to the front element. See if there's any additional or different flare on the digital shot.

 

My guess is that you will not see anything significant. The theory is plausible, but I doubt sensor reflections are a significant contributing source of flare because digital shooters in general do not report unanticipated flare problems with good lenses. I shoot plenty of sunsets and I'm pretty sure I would have noticed a problem or difference in flare between my digital and film shots by now.

 

Regarding iii - have to agree with Mark. A DSLR has more dynamic range than slide film. At least slide film on a desktop scanner, probably most slide film even on better scanners.

 

"I also think that the image sensors in your typical dSLR body will not last for nearly as long as the glass in your lenses. Some day in the not-so-distant future, current dSLR users will find that their cameras need to be replaced because of aging effects on the image sensor. I don't know when it will happen, but it will (I'm guessing 10 to 15 years);"

 

Again, with all due respect...why do you think this? IC's don't really age or break unless they overheat or are connected to a poor power source. DSLR's don't overheat and run off batteries. Why should the sensor age any faster than the IC's in 35mm bodies, which in some cases date back 30 years?

 

No offense, but this is pure FUD. There's no documentation of "age effects" and every indication that today's IC's will last at least as long as the IC's manufactured with 30 year inferior technology. I don't really understand why you would think otherwise, or wish to spread such unsupported beliefs.

 

I'm far more concerned with the backlighting on my 10D's display, as LCD's and backlights are known to fade and wear out.

 

"Of course, many users will need new dSLR bodies before this time because of cumulative damage to the image sensor from dirt and stuff, and from trying to clean off the dirt and stuff."

 

The sensor is covered with optical grade glass. Glass that gets cleaned less and more carefully than lens front elements. As long as you don't scratch the glass, there is *no* damage to cumulate.

 

Again, where do you get this stuff?

 

"Many current technical problems with dSLR technology (such as sensor flare, no off-the-sensor flash metering, and saturation) are just that: technical problems for which engineering solutions are being developed. The science part was done a long time ago."

 

No one is going to bother with OTF/S flash metering as other metering methods surpassed it in the film days. And saturation is a Photoshop issue, not a camera issue. Expect DSLR's to continue to be calibrated much like portrait films because scene accuracy is first. Creative interpretation (i.e. saturation) can be added later.

 

"For the next 5 to 10 years (at the most), digital photography technology will continue to follow Moore's Law."

 

They haven't really followed Moore's Law up until this point (they've been behind the curve), and will slow down even more. Sensors are limited by the physics of light, not Moore's Law. You can only make sensor sites so small before S/N issues become overwhelming, and lenses can only resolve so much. Digicams are hitting the limits now, and while DSLR's have room for growth, they won't grow at the pace of Moore's Law.

 

"Then it will take a big jump when quantum processors and holographic memory hit the shelves,..."

 

I'm nitpicking at this point, but quantum processing only speeds up certain tasks. For most common computing tasks they will be no faster than classical computing. I don't see any real advantage for digital imaging here.

 

The next major leap in digital imaging will occur with the advent of biologically grown sensors and IC's. Eliminate the clean room and 4x5 shouldn't be too much more expensive than APS-C. Sensors could achieve S/N similar to the human eye, though that may be possible with electronics as well. At any rate, bio-chips are still a long ways off.

[chiswick_john]

"such as sensor flare, no off-the-sensor flash metering, and saturation)" - these are non existent problems. Shooting RAW with a 20D will yeild a far greater dynamic range than any tranny film. Full frame sensors will never be as economical as APS sensors - Moore's law does not only apply to full frame sensors. It's easier and cheaper to design and manufacture lenses with smaller image circles - the bigger your glass the harder it is to get good performance - this is a fact that has been recognised for many decades by lens designers and is not subject to Moore's law.

[yakim_peled1]

FF sensors will always be more expensive than APS-C ones. As FF sensor prices will go down, APS-C ones will go down as well. Therefore it is logical to assume that the two formats will coexist. FF for high-end bodies and APS-C for mid and low-end bodies.

 

Happy shooting,

Yakim.

[chiswick_john]

Dynamic range is fairly easy to measure for all pratical purposes. Photograph a smooth surface at a range of exposures (use shutter variation) and measure the density changes in the colour picker tool in Photoshop and plot a graph. You can do this with film as well but it's much easier with digital. Discard the dark frames where noise becomes a problem and make sure you know what you are doing with RAW processors - many default settings will clip tonal detail to give a good look to the midtones. Don't test with jpeg either as the in camera processing clips tonal data for the same reason.The reason many testers don't test dynamic range is that unless they can process the RAW file they can only test the limited quality Jpeg's. When a camera is first released many RAW apps havent got the plugins to process the RAW data and camera manufacturers own RAW processors are likely to do all sorts of hidden tonal adjustments which you can't control. Additionaly getting the maximum tonal range out of a RAW file AND getting it to look good requires quite a lot of work in both the RAW processor and an image editor - squashing the entire tonal range available from a sensor into a file can result in a very flat looking image - you often need to put a film like response curve on it to compress the highlights and shadows and increase contrast in the midtones.

[andrew_blankertz]

Imatest can measure dynamic range.

[jon_austin]

"FF sensors will always be more expensive than APS-C ones. As FF sensor prices will go down, APS-C ones will go down as well."

 

True enough.

 

"Therefore it is logical to assume that the two formats will coexist. FF for high-end bodies and APS-C for mid and low-end bodies."

 

Not necessarily. Seen any -386, -486 or even earlier generation Pentium computers for sale recently? Windows 2000 will run on any flavor Pentium, 133MHz or faster. Windows 98 (and there are still plenty of folks running various versions of this OS) will run on any 486DX, 66MHz or faster.

 

But these CPUs, although obviously less expensive to manufacture than current generation Pentiums and Celerons, are no longer sold, simply because the more powerful, current generations are cheap enough for the mass market, so the demand for the older CPUs no longer exists.

 

Back on the dSLR sensor front, you don't even have the OS compatibility / system responsiveness issues to contend with. Once full-frame sensors become cheap enough for the mass market, the demand for cameras with less-than-full-frame sensors will simply dry up, regardless how cheaply they can be made.

 

When that time arrives, the demand for EF-S lenses and the like will similarly erode. Canon will eventually drop the EF-S line, as the installed base of bodies with which they are compatible diminishes, rather than incur the extra cost of maintaining these additional SKUs in their product line.

[flaviu_bor]

I just remember when (in '93) i buy my first "profesional" hard disk drive (and it was expensive !) Whow, 540MB, hudge !

Wake up James, we live in 21 century and Leica 35 MM is history (or become one). The technological progres aloud continous miniaturization

and I don't see no reason to cary enourmous chuncs of glas. I think 1.6 x sensors, are the next future and will see "some" improovments very soon.

ps. I too used slides since 60's !

[daniel_taylor]

"Not necessarily. Seen any -386, -486 or even earlier generation Pentium computers for sale recently?"

 

You are confusing chip density with chip size. The die size of a modern Pentium isn't all that different from the die size of an old 486, but the transistor density is dramatically different. Which is why people buy Pentium 4's and not 486's.

 

Transistor density doesn't get you to full frame. You have to increase the die size, which dramatically decreases yield rate.

 

APS-C is going to be around for a while.

[jon_austin]

"You are confusing chip density with chip size."

 

No, sir, I am not confusing anything with anything. My point has nothing to do with technology factors at all.

 

The FF sensor has already been achieved, and in fact is coming down in price. And when the price of those FF sensors reaches a "reasonable" or "affordable" level (whatever the marketplace deems that to be), demand for dSLRs with APS-C-sized sensors may very well dry up and blow away, regardless how much cheaper they are to make at that point in time than those the FF sensors.

 

And the die size and transistor density of 486 and Pentium chips didn't amount to a tinker's damn in terms of marketplace demand. People didn't start buying Pentiums over 486's because of the these factors, any more than because they were more expensive. They bought Pentiums because they were (a) "affordable," and (b) provided a better cost/benefit solution in the mind of the consumer.

 

Don't get too mired in the technology, my friend. VHS defeated BetaMax.

[daniel_taylor]

"No, sir, I am not confusing anything with anything."

 

Yes, you are.

 

Your analogy is that FF will replace APS-C because Pentium replaced 486.

 

But the only reason Pentium was able to out perform 486 at an affordable price point was because of increases in transistor density.

 

Transistor density per surface area and/or unit cost increases according to Moore's law. Die size *does not*.

 

Transistor density does not get you from APS-C to FF. You can only do that by increasing the die size, which dramatically decreases yields and increases per unit costs.

 

Your analogy is false because the forward progression of CPU's is based on the ease with which transistor density can be increased. But this does not help one get to a full frame sized sensor. So your analogy doesn't actually have anything in common with the problem being discussed.

 

"My point has nothing to do with technology factors at all."

 

You cannot separate cost from technology, or market penetration from cost.

 

"The FF sensor has already been achieved, and in fact is coming down in price."

 

Really? When did Canon announce the 1Ds mkII price drop? Because while FF has been achieved, it has been stuck at the same price point for 3 years now.

 

"And when the price of those FF sensors reaches a "reasonable" or "affordable" level (whatever the marketplace deems that to be), demand for dSLRs with APS-C-sized sensors may very well dry up and blow away, regardless how much cheaper they are to make at that point in time than those the FF sensors."

 

This is not going to happen for a *long* time. If FF was $500 and APS-C was $250, there would still be considerable demand for APS-C. FF is a long, long way from $500.

 

In response to Les - this topic has come up before on Photo.net, and someone in the industry posted numbers that basically said FF shouldn't be profitable at $8k today. Now obviously Canon is making a profit on the 1Ds, but it just goes to show that no average manufactuer can pull this off. His size/yield numbers were industry average. Chips 24mm x 36mm in size are difficult to achieve, and not many companies can pull them off at all.

 

He also posted projections which assumed dramatically improved manufacturing, and prices were still pretty high for FF.

 

Someday FF will be much more affordable, and will probably negatively impact APS-C demand, though FF would have to be dirt cheap to erase APS-C demand. Both days are a long ways off.