How much better will digital get?

[Jonathan Wienke]

OK Johnathan,

Take my #100 bet.  I'm betting that you are 90% wrong.  Of course  it's a long term bet as It will take three years to prove  out.

Did the price of the  new  1Ds II come down from the first 1Ds?  No it  did not.  Did not Phase One raise the MP of their backs over Kodak by 6M and more than double the price?  

These are the trends.  Not what you keep predicting.

First of all, the time frame for my prediction was ten years, not three. An it's a fact that less than 10 years ago Kodak was selling a DSLR that was less capable than the DRebel and it cost ~$30000. I'll leave the digital price/performance ratio change between then and now as an exercise for the reader. The 1Ds cost $8000 when it was introduced. They're going for around $4000 on eBay now, just a little over 2 years later. What are the 16MP Kodak backs going for now? Way less than $25,000. The biggest and best will always be expensive, but 10 years from now, the biggest and best will probably be full-frame 8x10 and 100+ megapixels. And it may well cost $30,000, but a 22MP MF back will be WAY cheaper.

You are drinking the  digital coolaide.  Go out and shoot  some film.

Film is too f---ing expensive. I shoot about 50,000 frames a year with my 1Ds. What's film & processing for that many frames of 6x7? Or even 645? I figure I've saved more than the cost of my 1Ds, 1D-MkII, my lenses, my Epson 7600, and all my computer crap, with about $20K left over, just in film I didn't have to buy and develop. BTW, that's why pros have no problem justifying the "outrageous cost" of a P25. In a busy studio, it only takes a few months to offset the $25K just in eliminated film costs, and then after that it's saving money. And then there's the workflow advantages and the quality improvement--an additional bonus.

You're the one drinking the kool-aid. The sky is not falling, and MF backs are hardly a dead-end market segment.

[BJL]

So, how do medium and large format folks ever get pictures that are not seriously compromised by diffraction and extremely narrow DOF?  Why do medium and large format large prints look so much better than 35mm? ...
There's a serious conflict between real life and math here, sort of like bumblebees being poorly designed for flight.

To the first question, perhaps you misunderstand me as saying that the DOF/diffraction trade-offs are worse for larger formats: I am saying that they are the same.

To the second question, I have no expertise, but at a guess some factors come from the lower degree of enlargement needed and the fact that not everything is scaled up in proportion to film format size: the same emulsion resolution and grain give finer resolution and grain on prints, and the lesser degree of enlargement also does some "noise smoothing" that can increase the ability to distinguish small tonal gradations by reducing the S/N ratio.

So do not get me wrong, I am fairly sure that you will get somewhat better results out of the best available MF digital system than from the best available 35mm format system.


And I expected that sooner or later, someone would mention that annoying anti-scientific urban legend about physics proving that a bee cannot fly. It is perhaps apropriate that a mathematician/physicist debunk this to a biologist in a forum about photography:

What was actually proven was that a bee could not fly if its wings were rigid, so that it is important that they flex in flight. Later, high speed photography confirmed this prediction about bee wing flexing, and such flexing is important in the design of helicoptor rotor blades.

So, it was not a failure of theory at all, but an accurate and useful prediction which also reveals a wonderful adaption in the form of bee wings!

[BJL]

IIRC, the effects of diffraction are related to physical aperture size

Yes, that is the way I like to look at it, especially when comparing different formats. The angular smearing of light passing through a small aperture depends on the aperture diameter: light spreads with the bulk of it within an angle of very roughly (wavelength of the light)/(aperture diameter), measured in radians. (Multiply by about 3000 to get it in minutes [corrected!] of arc.) For visible light this is about (0.0005)/(aperture in mm).

When it comes to prints of a given size and angular FOV, this angular spread is what counts; in a sense the aperture diameter sets a rough upper limit on the number of "lines per picture height" at which you can get decent resolution on a given subject. So when limiting diffraction effects on prints of a given size and subjects of a given FOV to a given level, longer focal lengths can use proportionately higher aperture ratios.


On the other hand, when comparig different focal lengths and FOV's used with the same format, it is relevant that the diffraction spot size on the film/sensor is proportional to both this angle and the focal length, so is proportional to aperture ratio. That is why photographic discussions of diffraction effects measure diffraction in terms of f-stops. And if one's goal is to get all the sharpness one can out of a given pixel size (as with Jonathan's approach to CoC) the upper f-stop limit is the same for all focal lengths.

Using a formula from this page at Norm Koren's site and the rough estimate that the resolution limit of a Bayer sensor is about one line pair per three pixels, we get that for the 9 micron pixel pitch of the original 1Ds and current MF backs (giving 22MP in current backs, about 30MP in full 645 format), diffraction has 50% MTF (one stop of contrast lost) at the sensor's resolution limit for about f/21. If one more optimistically believes that the sensor can resolve all the way to one line per pixel, this becomes f/14. So f/16 emerges as a pretty good rough limit if you wish to yield no significant resolution to diffraction.

Even then, most interesting detail is not right at the sensor's resolution limits, and this somewhat less fine detail is hardly affected by diffraction at all even at f/22 (or f/32). Remember that the sense of sharpness of an image usually has more to do with having very high MTF (50%, 80%?) for moderately fine details than with the finest detail that gets just barely resolved.

P. S. As soon as you back of from the edge of resolution limits and look for good holding of contrast (high MTF) in slightly less fine details, sensors and such might well give way to lens MTF performance as the main limiting factor. This is one respect in which MF lenses can have an advantage over 35mm format, often overlooked when we look at numerical measures of performance at the extremes.

[didger]

Wouldn't all of Didger's problems be solved if he just got in his truck and backed up a couple of miles?

Huh?  I assume this is a put down, but I'm afraid I don't really grasp the cause or intent, but thanks for your concern.

[Bobtrips]

smaller sensors; smaller, lighter, cheaper, and easier to build lenses.

'Fraid not. That would only be true if the resolution didn't need to be increased to match the smaller sensor sites.  Increasing the resolution of the sensor means to take decent advantage of that the lenses need to be correspondingly improved in resolution.  That means tighter design specifications, tighter manufacturing procedures, much less tolerance for error anywhere in the process and finally much tighter quality control and also much better build so things don't start to shift after the lens leaves the factory.  Since build and sample variation problems are already huge with present day lenses and most of them are inadequate to take anywhere near full advantage of a 1dsMKII, there's little hope for much better lenses than that right around the corner.

Let's go back to Michael's review of the 1D MkII (with L glass) and the also 8 meg Pro 1.  

That little cheap Pro 1 lens did a pretty good job of keeping up with the L.

And those 8 megs of Pro 1 photo sites sure are tighter packed....

(Now Digger, you ought to be clamoring for things to move in the direction of smaller sensors/smaller lenses.  Imagine sticking a 22 meg camera in your shirt pocket on a future trip high into the mountains.)  

[Ray]

I would say it the better manufactures are absolutely state of the art. Its the cost compromises one must make that determines final lens capability.

Who's going to buy a $20,000 wide zoom?

Judging by comments on such forum as this, there's a significant variation in quality amongst lenses of the same model and price. It becomes very expensive if a manufacturer sets up a quality control system of such a high standard that say 50% of all lenses are junked because they don't meet that minimum high standard.

Far better, I would think, to test each lens in the factory (instead of the occasional one or two from each batch) and grade and price according to quality.

Instead of an RRP of $1000 for each and every lens, a lens could come in 3 flavours; $2000 for the few that are spot on; $1000 for the bulk of them that are average, and $500 for those near the lower end of the QC range. Total revenue the same.

Mind you, such precision might take the fun out of buying a lens. With current arrangements, buying a lens is almost like buying a lottery ticket   .

[Ray]

The bottom line is, the is no free lunch.  

There's a lot, lot more than free lunches. There's lottery tickets.

[didger]

Bob, on the whole I agree with your well thought out synopsis and I congratulate you for the fortitude to read the whole thread.  As for the goodies coming up, the one weakness remains lens technology.  True, DXO type software can help and may help more in the future, but lens sample variations (of neccesity even greater as physical dimensions become smaller) will limit this and the process is in any case severely limited.  Let's say that a small red spot in a subject is substantially too small for what the lens can resolve and no red gets "triggered" in the sensor.  No software in the world can give you that red information if it never gets into the data stream.  A tiny little blurry red blob can be more intellegently sharpened than what non lens specific software can do, but you can't create something out of nothing at all.  There's limits to what software can do and if all these sensor miracles can really be done, then the physical lens constraints will continue to be the bottleneck, just as they already are with 1ds to some degree and much more for 1dsMKII.

Your 95% figure is probably pretty close.  I assume that when it was all film cameras, MF and LF were no more than 5%, with the rest being 35mm of all levels and also polaroid and various other consumer formats.  No matter how good that 12-17 Mpixels in small formats gets, I'm more interested in what will happen with that 5%.  I don't expect huge price drops, but I do expect that fairly soon very serious MF and maybe even eventually 4x5 digital will get down to way under $25,000 and that there will be the same sort of vigorous development and competitition there once was for film MF.  5% is not nothing and there's the prestige of a company having camera that most people can't afford, but all drool over.  I'm not so willing to write off Mamiya as the company that will show the rest the way.

[Bobtrips]

You sure?

Folks who will pay $30k for a digital back shouldn't have a problem with $5k lenses.

Yeah! Dead sure. It's really a matter of how many people there are there throughout the world (we're in a global community here) who would buy a $5000 Zeiss lens or $30,000 digital back if the price were, say 1/5th.

If there are not many, then it's no go. If there are heaps, and I suspect there are, then it will eventually happen.

You guys are in love with the idea of owning a 22 meg back.

What if Canon hits the market with the 2Ds - 24 megs?  And they bring it to market at or below 1Ds prices.

Or if Kodak figures out their 14x problems and markets the 24x?  And for $4K.

How much demand do you think there is going to be for high quality lenses?  

(And let me change the "ifs" to "whens"....

Realistically there can't be $10 worth of material in a lens.  Some glass, some metal.  A bit of paint.  (OK, I don't really don't know about the $10 figure, but it can't be much.)

Tool up for extremely high quality they'll fly out the door.

[Paul Sumi]

The most amazing thing to me was the photo was taken with a MF camera on an airplane 65,000 feet above the golf course and traveling at 1200 mph.  That was 40 years ago and film.  How good can digital get?

Well, "they" are now reading license plates from earth orbit.  And that's just the stuff which is publicly acknowledged.

[Ben Rubinstein]

The biggest I have heard of is a bit more than 50MP, in a very large sensor from Fairchild I think

Were they shooting RAW?  :p  :p

[BobMcCarthy]

I may have used a bit of hyperbole' in my description. Then just what is light, a wave w/o mass or a photon which does (infinitely small but ?) I don't know, thats for sure.

I think many are confused on the effect of light on a sensor, relating to film which I see as an incorrect way to evaluate the data. I see the sensor not seeing light and reacting with an on or off result (film grain) but a "number" which is the  result of the accumulation of light during the period of shutter opening. It outputs that number (modified by a filter on top of the sensor to get the sense of what the color could be) and calculates a result. I like the Fuji approach of having two overlapping fields so that one sensor field is less sensitive than the other. They can calculate the difference or calibrate the difference.

It should provide additional headroom by the differential between the two fields. Now why Fuji used the approach of using a tiny sensor to do this, I don't quite comprehend. If I wanted to reduce the light hitting a sensor I could reduce the collection area (Fuji way) or add additional filtering (ND) to slow the accumulation of light. Either should reduce the light accumulation rate (a number) which could be factored up post exposure and would provide a valid result when the more sensitive field hits the ceiling and goes outside of its operating range.

If the tiny sensor only operates in the upper zones , then its a 6 mpxl camera. If it operates full range with a 2 zone offset, then its a 12 mpxl camera with 6mpxl extreme highlights..

I have been primarily a B&W photographer for 30+ years and since I have had a maleable dynamic range to work with, I find the expansion of DR very important. Had I been a color (transparency) photographer, I might look at this differently.

Can one go over 255 (or obviously offset the data).

[didger]

****Have you had a secret sneak preview of ZD, and found it lacking?****

No.  I've seen the 1Ds II and know Mamiya can't match it.

Well, your absolute knowledge in the absence of any actual facts (that I've been privy to) is simply remarkable.  I wouldn't dream of betting against such confidence, though I still don't share this confidence myself, since I've had no miraculous revelations either way.

[Jo Irps]

As soon as I have the money  I will spend up to $5k for a mint Plaubel Veriwide 100.  Or maybe  an Alpa.  Screw digital.

You can have a mint Veriwide for $1000 from Ralph Fuerbringer in NY. He did one for me, and I am a very satisfied user. I even have a second one with a Horseman 6x9 back and the newer Schneider SA 5.6/47. The Veriwide is so small that you can take it anywhere in your pocket for snapshots, just like a Leica. The other one is for more serious landscape work with a tripot. Digital is out of the question for me, period.

[jimk]

canonites are gonna hate me..(6) but i dont like cameras that cant autofocus straight (backfocus) and give me a plasticky creamy fake color

i would rathger stick to film than deal with digital if canon was the only digi company out there

[BernardLanguillier]

The number of megs will most likely continue to increase and prices will likely continue to fall.  But the number of megs, just like the clock speed of computers may well become less of an issue within the near future.  People are going to start hitting their "personal enough" levels and won't be tempted to upgrade just to get more pixels that they won't/will seldom use.

Hi there,

Another little discussed reason why more MP will stop to make sense very soon is hand shake...

I have never seen a serious 1Ds vs 1DsMKII for hand held shooting, but I am 90% sure that there is virtually no resolution advantage for the 1DSMKII when shooting hand held in most situations.

For all the people shooting handheld, the only solution will be to go for larger sensors and MF is then the only option. Here also, we are probably close to the limit though.

Regards,
Bernard

[Sfleming]

**** It was interesting to read Kev Raber of Phase 1 say that "Medium format systems and even present DSLR system will keep their price points.****

Yeah,  I bet everyone here went back and read that sentence twice.

****That is a trend I have seen with higher end products; certain markets (e.g. MF studio photographers) seem to have acceptable price points, and one strategy is to keep product lines at various such price points while improving quality.****

This is a marketing mistake.  They are ignoring the tens of thousands of amateures doing fine art for the thousands who can afford 30k for a back now.  They cannot stay alive at the present price point.  95% of their market will be satisfied and thus saturated with 22MP and not upgrade.  They will keep shooting with the back they have. When I say 95% I've already  lopped off the huge segment  that will say "screw this" I can do everything I need to do with the 1Ds Mk whatever.

The Mk II is a Phase One killer.  If they don't get the price point down ... way down.  They will cease to exist as MF back manufacturers.

I suspect Kevins statement (maybe he actually believes it) is mostly an attempt to squeeze out a few more sales at the  absurd prices now listed.

[Jonathan Wienke]

P. S. About my qualifications: I am a professor of applied mathematics, and publish in mathematics and physics journals on topics related to wave propagation, so vaguely related to the topic at hand. The questions  at hand only depend on the lens optics and geometry that I learnt as an undergraduate, not the internal physics of sensors.

Well your math regarding sensor size/format and its effect on DOF is wrong; larger formats do have innately less DOF when all else is equal. See this post for the proof.

[howard smith]

"Perhaps the evil effects of diffraction are being over-estimated here."

In my opinion, yes indeed.  Sometimes when reading this forum, I get the idea that even attempting to take a decent photograph without a Canon 1ds, a Zeiss prime lens at f/8, a carbon fiber tripod, etc. is bizzaar.

[John Camp]

Wouldn't all of Didger's problems be solved if he just got in his truck and backed up a couple of miles?

Huh?  I assume this is a put down, but I'm afraid I don't really grasp the cause or intent, but thanks for your concern.

Not a put-down. It's just that you seem especially concerned about performance in your wide and ultra-wide lenses, and see, if you just got in your truck and backed up...never mind.

JC