How much better will digital get?

[Ray]

I'm sure there are some people in NASA who are still rolling in laughter as people get excited about the new advances in 8mp 1.6x DSLR sensors, circa 2005...

Those of us who are exited about the new 8mp 1.6x DSLR sensors also have something to laugh about ....... the very low price   . I wonder how much NASA paid for their 8mp sensors 10 years ago.... half a million each?

[RobertJ]

John, I think the ONLY digital cameras that will produce finer prints than the 1Ds2 will be the upcoming Medium Format digital backs/DSLRs, and whatever Canon replaces the 1Ds2 with in the future.  I just don't see anyone catching up to the 1Ds2 in the 35mm world, but when/if they do, (Nikon, or someone else), at the same time, Canon will be releasing their 1Ds2 successor, (1Ds3/1D3 with 12mp+ and 8fps?).  

Think more about which "System" you want to own, not which camera.  The D2x could be great, but if you like the Canon system, lenses, and technology better, then maybe you should forget about the D2x.  Canon owners have been shooting with their own "D2x" for more than 2 years now.  It's called the 1Ds.  Just something to think about.

T-1000

[Ben Rubinstein]

IMHO and experience, clients don't usually care what gear you use, just as long as your results satisfy their needs.  (ad agency suits/wannabees notwithstanding)

Exactly, my wedding clients just about understand the difference between film and digital (so we get Cd's instead of negs, great!). That argument is totally false, plenty photographers shooting weddings with old Leica's, Hassie's and F3's, no one asks them why they arn't shooting with the 1Ds mark II, for your client it's the pictures that speak, they won't peer at your camera to find out the model number, even if they were to know the difference.
This is even more the case for landscape photography were the only, ONLY thing that sells is your pictures, nothing else.

[Jonathan Wienke]

Somebody has built a camera that is far better than anything else in its class. It has no real competition. But it's expensive. This is the death knell of photography.

Sorry Scott, that makes absolutely no sense. The reason Canon has no real competition in the 35mm and under DSLR format idecause their stuff is so good, not because it sucks. Spending "up to $5000" for a fixed-lens film camera with far less capability is like cutting off your nose to spite your face. You certainly aren't saving any money in the long run if you actually use the camera (you'll have to buy a scanner to go with it, and film), and you aren't going to gain much in image quality.

[Ray]

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.  

Possibly. But I still remember the days I first got interested in HiFi. I got a book on the basics from Radio Shack which explained that the ear had a range from 20Hz to 20kHz and could detect harmonic distortions as low as 1%. (In fact, I later discovered this varies between about 0.35% and 5% depending on musical content.)

Yet all the advertisements for amplifiers in those days boasted THD of much less than 1%, like... err... 0.0001%. What's going on here? I guess the pressure is on for salesman to sell amplifiers. Companies who produce amplifiers want to continue doing so. The fact that their amplifiers don't sound any better than (and in the view of some, not as good as) pre World War ll valve amplifiers, will not deter them.

The 20D is more useable than the D60 at high ISO's thanks to computer chip and programming developments, but at ISO 100 it's no better. Why is this? Could Canon have produced a 20D with 6 megapixels which had lower noise than the D60 and 10D at ISO 100 as well as at higher ISO's? Did the marketing guys at Canon agree that an 8mp camera was going to sell better? Or is there some real obstacle as a result of the laws of Physics which just can't be surmounted?

If the last point is true, then pixel count will continue to increase because there's no where else to go   .

[Ray]

Well, if I can weigh in here (I'm a professor of clear thinking... only kidding   ).

I've noticed Didger's made a number of comments expressing an opinion that the 1Ds and 1Ds ll sensor is already as good as most of Canon's lenses and therefore there's no point in upgrading to the Mkll expecting better resolution if current lenses are not up to the job.

There's a misleading implication in such comments that current sensors are actually outresolving the lens. The same mistake was made with regard to film as film emulsions became finer grained with greater resolving power. You'd hear comments that film was becoming so good it could outresolve most lenses. In a sense this could be considered true in terms of MTF response at a specific resolution. At 20 lp/mm, for example, many films can record detail with no loss of contrast, ie the film has an MTF response of 100% at 20 lp/mm. No lens is this good. However, the same film at 30 lp/mm might be down to 50% MTF and many good lenses can do better than this.

T-Max 100 B&W film can apparently record up to 50 lp/mm with no loss of contrast and at 100 lp/mm MTF is around 60%. Clearly no lens is this good so I suppose it's a fair statement that T-Max 100 can outresolve all lenses.

I don't know what it would take to design a 16mp full frame sensor capable of recording 50 lp/mm with no loss of contrast, but if it is ever possible, I think you'd find that many current Canon lenses would show a marked improvement with such a camera.

[didger]

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.

Different strokes for different folks.  I've seen too many David Muench and Michael Fatali exhibitions to take a casual attitude about what some folks would deem unnecessary extremes to try to get the best possible resolution for the sharpest possible large prints.  Family snapshots work fine with disposable film cameras.  Very large fine art photography prints require another standard altogether and then there's everything in between, with everybody seeming to think that whatever position on the scale their style or economic constraints dictate is the "smartest" position, with those settling for less being undiscriminating and those insisting on more being unnecessarily fussy.  

For now I'm quite happy with my 1ds, Zeiss lenses, and CF tripods.  As soon as better stuff becomes available and I can afford it (even only marginally) I'll go for it, until (if ever) portable digital gets as good as LF film.  I don't give a rat's ass what anyone else may deem good enough for themselves.  I know what my ultimate aim is for myself, thank you.

"Bizaarr" would probably rate at least an honorable mention in any bizarre spelling contest.

Since larger formats use longer lenses to cover the same FOV, you can get away with f/64 instead of f/8 because you're using a 300mm lens instead of a 24mm to get the same FOV. But the physical aperture diameter where diffraction starts becoming an issue is roughly the same in either case.

OK, THIS makes sense and it means mathematical abstractions notwithstanding that a given composition with MF will give you a larger sharp print and no disporportional DOF and diffraction losses.  Everything increases proportionately as image size and lens size gets bigger.  Diffraction is what happens when light hits a sharp edge and the smaller the hole, the more the image will be affected, and yes indeed, a larger camera/lens will result in a bigger absolute aperture diameter at the same RELATIVE f stop.
And Howard, please don't be bothered if some of us want to do everything possible for improved technical quality.  That's just our personal preference and not something we're trying to push on you, just as your love affair with rules of composition is your personal preference and should not be pushed on us.  To each his own.  Jonathan and I are both very prolific and very happy with our photography efforts and in no way crippled or compromised by our technical quality obsession.  De Gustibus... and all that.

[DiaAzul]

I didn't read all the posts so far, but seem to have got to the point that the bigger the hole the less diffraction (OK, I know this - basic physics 101).

What happens though if we go to a lens which is purely based upon principles of diffraction rather than refraction and instead of adjusting the size of one hole we add or subtract holes to increase the amount of light entering or leaving the system. In this case we do not become diffraction limited and can use the diffraction effect to bend the light to the focal point. It all gets a little complex with the interference properties against frequency; however, Canon employ this principle in their DO lenses and perhaps it will enable us with the 35mm system to exceed the performance constraint of current lenses.

[BobMcCarthy]

For barely acceptable appearance in shadows a local S/N ratio at individual shadow pixels of about ten is needed according to Kodak, and as Ray indicates, this needs at least 100 electrons counted just to overcome photon shot noise. (Other noise sources push the practical lower limit up). If your well capacity is about 100,000 (as for current medium format back sensors), that optimistically gives a range of about 100 to 100,000, or 1,000:1, or about ten stops from highlights to darkest decently reproduced shadow areas.

I had no idea the numbers were so low. I had always assumed that captured photons were like electrons in current, so #### many that counting them was a challenge.

[Ray]

Use your histogram.

Expose to the left.

If you have the ability to greatly extend the camera's DR you won't blow out the highlights.

Don't you mean, expose to the right? I don't see this as a solution. This is just normal practice isn't it?

[Jonathan Wienke]

Use your histogram.

Expose to the left.

If you have the ability to greatly extend the camera's DR you won't blow out the highlights.

This is absolutely stupid advice. You're gaining highlight headroom by pushing shadow detail beneath the noise floor. In all cases, the best image quality and dynamic range is obtained when non-specular highlights approach, but do not reach the clipping point. Exposing to the left reduces DR by effectively raising the noise floor of the data.

Imagine a scene with 5 stops of dynamic range. If your camera has a 5-stop exposure interval between clipping and the noise floor, you can just barely capture the entire scene if you expose to the right and push the highlights right to the ragged edge of clipping. But if you underexpose by 1 stop, you've just reduced your captured DR from 5 stops to 4 stops because you've thrown away 1 stop of shadow detail by burying it in the noise. If you overexpose by 1 stop, you've also captured 4 stops of the scene because 1 stop worth of highlight detail is clipped away.

If the scene has only 4 stops of DR, a capture with the highlights close to clipping will have less grain and noise and more detail than a capture of the same scene with the shadows barely higher than the noise floor.

[Sfleming]

Price is determined by other factors than quality.  If I could make a lens as good as a Zeiss, it would likely be too expensive to sell, or I would have to lose money on it.

Well, we know that lenses as good as a Zeiss can be made.  

It occurs to me there may be a very good reason other than my 'paranoid' suspicions why  Canon hasn't built an 'Uber L' wide angle lens for the 1Ds II.

To get true WA for the Canon FF sensor one would have  to put the equivalent of the  Contax 645 AF 35mm (but in a 15mm or 20mm version) lens on a camera containing the Canon FF sensor but  having a full throated 645 mount.

Obviously this  will never happen.  So it may be that folks wedded to WA photography should stick with  view  cameras (6 x 9cm would be enough) and use  lenses  with at least twice the image circle of the sensor.

Just guessing.

[BobMcCarthy]

Thinking about the sensor issue, it dawned on me that the Fuji 3 is an interesting solution.I've seen very few in depth reviews yet. The whole concept is interesting with two sensor fields overlayed with each other with a sensor sensitivity difference and the sensor points  of different size. The large sensor is significantly larger than the small (could it may be possible the wiring (?) is all located on the small site.  Tight sensor packing for sure. It seems that the two sensor differences in sensitivity should be understood and may account for the 2 stop claim in extended DR.

I wonder if the fact that the claims are not easily seen by early adopters are a result of  underdeveloped processing software. There are 12+ million data collection sites, so it should be a 12+ megapixel camera. The overlap of sensitivities, means to me that the camera is 12 megapixel in the middle zones and 6 megapixel in the extreme highlight and shadows where one array or the other is outside of its intended range. Maybe improvements in software will give us a high megapixel Uber camera if the processing can be sorted out.


bob

[shaan]

After following this thread for some time i think i am not alone in stating the obvious answer. It is, without question...

42

Shaan

[Ray]

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.

[Quentin]

Quite possible.  From what I hear from pro PJs, it isn't that the 1ds is so great, it is the instant image of digital that matters.

Howard,

Near my London office there is an exhibition of photos taken by leading photographers using mobile phone cameras.  As I walked around the exhibition, I had this uncomfortable feeling I was being transported to the future, where this kind of photography was not a novelty, but a daily reality.  For some reason it was oddly depressing.

Strange.

Quentin

[Jonathan Wienke]

Light (as with all electromagnetic radiation) has characteristics of both a particle and a wave, and the physical size of the particle-like aspects of light (photons) is related to the wavelength of the light. Photons are not infinitely small. If you try to make a photodetector smaller than the wavelength of the light is is intended to detect, the efficiency of the detector will decrease dramatically as the size of the detector falls below the size of the photons it is trying to detect. It's kind of like trying to stuff a rubber ball through a hole; when the hole becomes smaller than the ball, the difficulty increases rapidly as the hole gets smaller.

[Paulo Bizarro]

So here's the question: Are the top line affordable cameras -- say the Canon 1DsMII -- good enough in producing 13x19 prints, that it's going to be a long time before newer cameras can produce better 13x19 prints?

 I'm willing to make the switch to Canon, but I'd be less willing if I were convinced that the 1DsMII is just another one-year temporary waypost on the way to a real camera.

JC

The 1Ds MKII "affordable"? I know all is relative, but this is still a strange concept for me. If you are not happy with the big prints from this camera, then you will have to look outside of 35mm format.

What do you mean "a real camera"? This is as real as it gets...

[Stef_T]

I see what you are saying Bob and Peter, and I can honestly say that I stand corrected. However, I can't see in the forceable future, a complete shutdown of digital photography evolution. There will always be something better, faster, lighter that can be built, and that there will be a market for.

[BJL]

Come on Quentinn and BJL...

Are we honestly discussing present day reality .... or just  tryin to win arguements?

It's not just the press pit ...

Primarily, my response was about refuting a blatantly invalid argument (a bit akin to those who trumpet ISO1600+ noise levels as the dominant measure of a camera's worth), since you offered the press pit alone as proof of a very broad claim, including it seems dominance over digital medium format options, as if the press pit has anything at all to do with that. The EOS-1D models currently dominate some important sectors of the market, but with total production levels of only a few thousand per month (I have read 2,000/mo 1Ds, about twice as many of the 1D, while total DSLR production is around 200,000/mo), it is clearly far from the whole picture.

There are plenty of areas in photography where Canon does not stand alone above all others, such as

- resolution and dynamic range, which are more important than speed to many landscape photographers.
Not only do MF backs outdo the EOS-1D series, but the much maligned Kodak DCS 14/n had an advantage in both resolution and dynamic range (as wel as price!) over the 1Ds, and quite likely the DCS SLR/* models still have that advantage in dynamic range (the unusual FillFactory sensors have an extremely large electron well capacity, giving them an excellent dynamic range despite their higher noise levels).

- High quality wide angle to normal lenses for the 99% of DSLRs with "sub APS-C sized" sensors. I think it is fairly clear that for those who cannot afford $4,000 and up for any of the EOS-1D series bodies, the Canon EF-S lenses (all zooms, mostly f/5.6 at the long end) do not match the overall speed, quality and selection of wide to normal lenses that Nikon and Olympus offer at shorter focal lengths with their DX and E system lenses. Olympus at least has indicated that more top of the line prime and zoom E system lenses are coming this year, while there is no hint of any EF-S prime lenses for example.

Given that sales and revenues in this price sector are vastly greater than those with 35mm format sensors, this can hardly be ignored.