Ad
Ad
Ad
Pages: [1] 2 3 ... 14 »   Bottom of Page
Print
Author Topic: How much better will digital get?  (Read 28203 times)
Stef_T
Sr. Member
****
Offline Offline

Posts: 266


« on: January 02, 2005, 11:45:24 AM »
ReplyReply

I agree with you pom, that it seems like digital has nowhere more to go. However, what i am saying is that we have made this mistake before with computers. 10 years ago, no one would image we need a 1ghz processor, now, people are talking about dual 2.5ghz. I simply believe that digital will continue evolving, and getting better for many more (more then 10) years to come. I agree that digital photography is at such a  point where, one is very happy with its results, however down the road, things will change.

Also, being a pro photographer, you have the problem of having to meet your clients needs. That, above all else, will keep the digital line moving and improving. Your clients, I am sure, want the best, which unfortunately means gettubg the latest technology. In 5 years, your clients will look and say: 'what, you're still using a 10D, we can get 1Ds preformance at the same price.' You will then have no choice, but either lower your price substantially, or upgrade your equipement to meet your cleint's standards. So even if in your mind, your photographs are perfect, it is your client's view that counts.

As the professional digital SLRs evolve, then they will eventually pull up the amateur, advanced and entry-level digicams as well. I don't believe that we shall see the end of this evolving cycle, for quite some time.

Stefan
Logged
didger
Sr. Member
****
Offline Offline

Posts: 2030



« Reply #1 on: January 04, 2005, 03:00:01 PM »
ReplyReply

Quote
for an aperture giving the same depth of field on prints in which the subject appears at the same size, those prints also have the same diffraction spot size
Yes, no one has contested this.  However, the more pixels you have in the affected area, the more noticeable any artifacts are and the more noticeable degraded resolution is.  The lens projects exactly the same thing on the image plane no matter what, but the more resolution the image plane has for capture, the more conspicuous any defects will be.  You're not just getting your image sharper, but you're also seeing your defects better.

I guess you'll have to base any of your camera buying decisions on your understanding and me on mine.  I won't buy a 1ds2 but I will get a medium format digital camera as soon as feasible because I'm convinced that 22 Mpixels of MF sensor will not only give much more effective resolution than 16 Mpixels of 35mm sensor, but also better DOF and diffraction performance, relative to the sensor real estate, not relative to pure lens optics considerations.

If you happen to be an optics and sensor design engineer I'll defer to the greater understanding you would almost certainly have.  If you're just another guy with a grab bag education for all this sort of stuff, well, maybe my grab bag is just as good.
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8854


« Reply #2 on: January 06, 2005, 08:54:26 PM »
ReplyReply

To get the thread back on track, here's something to consider.

There are broadly 3 sources of noise in digital cameras, (1) Photon noise (photonic shot noise), (2) pre-amplifier noise, (3) dark current  (thermally generated noise).

Of these 3 sources, photonic noise (random arrival of photons) we can do nothing about. It's a constraint of the laws of physics, a bit like diffraction is in lenses.

However, because photonic noise is equal to the square root of the signal, we can arrange for it to be less significant by making bigger pixels that receive larger signals, which is the general idea of large format.

Consider a really big pixel on a large format sensor that can absorb, say 1 million photons. Photonic noise will be 1,000 photons or 0.1% of the total signal.

Consider a really small pixel that can absorb only 1000 photons to maximum well capacity. Square root of a thousand is 32. Photonic noise is 3.2% of the signal. A big difference.
Logged
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #3 on: January 07, 2005, 01:00:29 PM »
ReplyReply

Quote
Can software improvments (ie post processing in the camera or raw converter), overcome flaws and errors,and subsequently remold the data, to provide a higher quality output.
Google "Neat Image" or "Noise Ninja" and you'll find well-established products than do exactly this, on converted images. ACR, C1, and most other RAW converters have some kind of noise reduction built in as well.
Logged

opgr
Sr. Member
****
Offline Offline

Posts: 1125


WWW
« Reply #4 on: January 08, 2005, 07:26:50 AM »
ReplyReply

And how exactly is reflecting that anger back at us going to help? And what does Nostradamus have to do with it? I don't recall him as a vehement example of someone who's only contribution to development is arguing something will never happen.

3-5 years is perhaps a little short, but I wouldn't be surprised if we'll end up with any of the following:
1. a computer controlled plasma lens allowing perfect focussed, non-deformed image projections across the entire magnetic spectrum.
2. a sensor containing just microlenses on each pixel, not needing any lens at all.
3. Photon Alignment algorithms.
4. single Photon detectors, which simply resample the image as much as is required to calculate photon ratios per sample, in a timespan that's well below any current ISO-exposure setting. The latter by the way will for sure be something of the past.

Or a combination of the above. If we simply detect the wave/photon pattern of the scene we want to capture, we can later calculate the desired DOF, sharpness, blur etc in postprocessing from the data. Even stereoscopic images would be possible in post, as well as a limited amount of 3D rotation.

And in 3-5 years I can still see Didger shoot with a 1DsMK3 because the increase in resolution simply translates in better and more detailed colorrendition. The 16MPix were single color, 12 useful bits pixels. The new 32MPix sensor will deliver true 16bit color for the same 16MPix previously, and even at 32MPix will deliver more depth in dynamic range etc. well beyond the diminishing returns mentioned. The lens limit btw is not a binary limit, so the increase in resolution may still enable a progressively sharper image, regardless of using the best lenses or not.
Logged

Regards,
Oscar Rysdyk
theimagingfactory
Bobtrips
Sr. Member
****
Offline Offline

Posts: 679


« Reply #5 on: January 12, 2005, 12:06:37 AM »
ReplyReply

Quote
Quote
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.)
Actually, I believe that some of the raw materials, such as special low dispersion optical glass, are themselves very expensive, especially for telephoto lenses. There is also such a thing as a maximum potential market for an item like a heavy, wide angle, single focal length lens, which gives a minimum price needed to cover fixed costs like R&D and tooling for production, regardless of how low unit production costs can be driven.

As soon as there is genuine competition (for the lenses or for the optical glass, often bought from outside suppliers), it defies the working of market forces to suggest that any sucessful company could increase profits by substantially decreasing prices. If that were true, one or more competitors would already have done it.
(Blame me for the quote, not Jonathon.)

Points taken.  

If I wanted to be defensive/argumentative I could say that even that exotic glass started out as a shovelful of sand.  

But I won't be that way....    Smiley

It can well be that producing the needed glass requires some manufacturing steps that drive up the price.  But I'd guess not as much as is reflected in the price differential between Zeiss/'L' lenses and the 'affordables'.

To some extent those higher prices can be (at least) temporarily maintained if the manufacturer has locked in the patent on some particular process/exotic coating/....

And in economics the term 'excessive profits' is used when the the profits are sufficiently high to cause other suppliers to enter the market.  Maybe one of the old high-quality film companies will rise from the 'almost dead' in partnership with a deep pockets company....

I'm just thinking that we're going to have a lot more high quality lenses at better-than-today prices hit the market in the next 5-10 years.

Do you ever remember anyone complaining about some Canon Ls weren't that good back in the days of film?
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8854


« Reply #6 on: January 22, 2005, 08:55:34 AM »
ReplyReply

Quote
Its a 6 megapixel camera then. And not what I'm looking for.
The S3 is a 6 megapixel camera with an arrangement of hexagonal pixels that allow better interpolation. The equivalent resolution is somewhere in between 6 MP and 12MP. There's a thread somewhere on this forum in which the resolution of the S3 was compared to the 8MP 20D and found to be slightly less.

I'd expect the interpolated 12mp of the S3 to be roughly on a par with the 20D.

I don't think the presence of the 6m R pixels justifies a 12mp designation because they are operating at a different level. Where the S pixels leave off, the R pixels take over. There needs to be some overlap, but essentially from 0-255 the image is described by 6 megapixels, before interpolation.
Logged
Peter McLennan
Sr. Member
****
Offline Offline

Posts: 1662


« Reply #7 on: January 02, 2005, 11:55:12 AM »
ReplyReply

"In 5 years, your clients will look and say: 'what, you're still using a 10D, we can get 1Ds preformance at the same price."

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)

A few years agon I had a new client look askance at my new digital camcorder, one with which I was about to shoot a three month job for him.  I wound up doing over three years continuous work for that client, all with the same camera.
Logged
BJL
Sr. Member
****
Offline Offline

Posts: 5121


« Reply #8 on: January 04, 2005, 01:46:18 PM »
ReplyReply

Quote
Do you think that a 16.7 Mpixel 8x10 back would give you the same diffraction and DOF constraints as a 16.7 Mpixel 35mm sensor? If that were true, then 8x10 film would be similarly constrained and useless.
Didger,
   perhaps the problem is that there are advantages to larger film formats, in areas like resolution and finer tonal gradations, but just not in the DOF/diffraction trade-offs that you are talking about. Perhaps I should make peace with fans of larger formats by publishing my list of their legitimate inherent advantages, with physical reasons.

In answer to your quoted question, I not only think so, but can prove mathematically what I stated before: that for an aperture giving the same depth of field on prints in which the subject appears at the same size, those prints also have the same diffraction spot size; changing focal length or format or cropping has no effect.

But this is far from showing that 8x10 format film has no advantage over smaller film formats, because DOF and diffraction are not the only relevant factors: other obvious ones are the resolution of the lenses and the film, and finer tonal gradations, a cousin of greater dynamic range. I have already mentioned the likely lens advantage of larger formats, so I will just comment on film emulsions.

Since the same emulsions can be used in any format, a study at the PhotoDo website suggests that the spatial resolution of 35mm format is often constrained by emulsions, while in large format, lenses are usually the main limit. Indeed, whenever DOF needs force the 8x10 user to apertures smaller than f/22, diffraction is a far greater limit on resolution that modern film emulsions.

If 8x10 at small apertures like f/32 has an advantage over MF or 4x5 at aperture ratios giving the same DOF, it is probably largely from aspects like the finer tonal gradations achieved with the lower degree of enlargment needed. This is physically related to the lower signal to noise ratio of a larger format, which measures the number of distinguishable luminosity levels, though the S/N jargon tends not to be used when discussing emulsions.
Logged
didger
Sr. Member
****
Offline Offline

Posts: 2030



« Reply #9 on: January 06, 2005, 09:17:25 PM »
ReplyReply

Yeah, sensor chip size is inherently limited by this very thing you point out and I just don't believe in any future breakthrough miracle cure for ultradense noise resistant sensors.  Computer chip densities also have some ultimate ceiling due to things like cosmic radiation hits and quantum level randomness.  We can shield expensive enough computers with lead, but you can't very well shield camera sensor from stray radiations producing noise.  

Along with these inherent sensor noise problems with more and more component density, there's also the problem of needing to make better and better lenses.  Again, the frequently expressed optimism that these better lenses are right around the corner is in no way backed by anything real.  There's not even any ultrawides being manufactured right now for any Nikon or Canon camera that are as good as the old Zeiss and Leica designs.  Design and manufacturing tolerances need to get tighter and tighter to produce these miracle lenses for miracle sensors.  Sample variation is pretty well out of control now even relative to existing (1ds, 1ds2) sensors.  Where is there any grounds for optimism that there will be lenses to adequately match 1dsMKIII sensors or 1.6x sensors of ever higher sensor site densities?  It's all fantasies.  The only way upward is medium format, and hopefully some day 4x5 digital.  

I expect that present day 4x5 film will be nearly reached with MF digital sensors and vastly exceeded with 4x5 digital, but small sensor cameras aren't going anywhere much farther than they are right now.
Logged
BobMcCarthy
Full Member
***
Offline Offline

Posts: 201


« Reply #10 on: January 07, 2005, 01:03:47 PM »
ReplyReply

Quote
Would it be possible at one shutter-speed/f-stop setting to somehow sample the incoming light twice? Or even to take the shot twice, but so quickly that the photographer wouldn't notice?
It would appear to be possible. Shutter speed and aperature wouldn't have to change. Turn up the amp' for the second shot (ISO) and have the capture run back to back with the original exposure.

Would require a steady platform. I have no idea how long it takes the chip to dump the data for the second calculation to start.

Issues, exposure timed is doubled +, noise on higher iso exposure, and getting a manufacture to make it work.

Degree Physics 1969. I was and am still thinking Newton, all this string theory stuff is beyond me. As is all this "light" stuff.
Logged
didger
Sr. Member
****
Offline Offline

Posts: 2030



« Reply #11 on: January 08, 2005, 05:37:37 AM »
ReplyReply

Quote
It'll be funny to look at this thread in about 3-5 years, and see how accurate our predictions for the future were.
Yeah, and you'll be marveling that Didger knew all along that there would be no ultra high quality little tiny lenses for ultradense little tiny sensors and that the future would be larger sensors for pro quality higher resolutions. D2X and 1dsMKII are close to efficient max (if not a bit over). Wedding, sports action, and news shooters have more than enough resolution now and price drops and quality improvements will be the main further progress. Landscape photographers will switch to larger formats in droves as more options and lower prices become available. Nikon and Canon will lose much of the pro market, probably all of the landscape pro market, if they don't wise up. D2X is a pointless offering because it won't be nearly as good as 1ds and 1ds is now available for less money. That prediction will be verifiable soon. 1dsMKII's will never sell a lot and they'll be cheap on ebay much faster than 1ds got cheap because it's not enough upgrade for the bux for news, wedding, and sports and a stop-gap measure and poor investment and upgrade dead end for landscape. I'll be switching to MF soonest while others are waiting for small format sci-fi miracles.

Didger also said 30 years ago (and even 40) that the new millenium would offer no believable artificial intellegence or simulations of living organisms. He got much irate argument from those AI geniuses at Stanford University (where I used to sell my electronic music albums on campus), but those marvels are still limited to sci-fi even though computer speed, memory, and storage have exceeded anyone's wildest imaginations and at dirt cheap prices. Artificial life and self awareness won't be here in 100 or 1000 years either. They probably laughed at Nostradamus too (and even Alfred E. Neuman).
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8854


« Reply #12 on: January 12, 2005, 05:18:23 AM »
ReplyReply

Quote
As soon as there is genuine competition (for the lenses or for the optical glass, often bought from outside suppliers), it defies the working of market forces to suggest that any sucessful company could increase profits by substantially decreasing prices. If that were true, one or more competitors would already have done it.
That's true, unless you also decrease costs. I'm often amazed at how inexpensive some of the goods are that are made in mainland China, just as I'm sometimes amazed at the high mark-up on other items.

I don't believe China makes any really high quality camera lenses at present. However, if the general trend in anything to go by, I see no reason why it will not happen.
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8854


« Reply #13 on: January 22, 2005, 08:40:29 AM »
ReplyReply

Quote
Ray,

  have you seen tests using best practice conversion of RAW output from the S3? I cannot see any way that the sensor can be indicted as a failure without doing that;
BJL,
The only tests of the S3 I've seen that used RAW conversions are Michael's test in his review. Didn't he use a beta version of C1 V3.6 for the S3, the 20D and the 1Ds ll?

Here's a quote from that review:-

Quote
There is no question that Fuji's dual sensor technology produces a slightly / somewhat (choose your preferred adjective) bit more dynamic range. But, when it comes to working with the files in Photoshop I find that even minor use of levels and curves, and especially exposure compensation in the Raw converter, tends to dilute the distinctions.

It seems that Fuji have used a different approach to achieve a very competitive DR for their camera, but have trumpeted their different design with a great deal of fanfare to imply that it's a major breakthrough and a revolutionary concept, when in practice it's a rather small incremental improvement.

I'm not trying to say the camera is a failure, but just searching for a reason why the improvement in DR is not as great in practice as it would appear to be on paper.
Logged
Stef_T
Sr. Member
****
Offline Offline

Posts: 266


« Reply #14 on: January 02, 2005, 09:11:14 AM »
ReplyReply

To be honest, I must say that I disagree that digital photography is going to reach a point where it stops getting better. Unfortunately, it is the camera manufacturers that control this and it is not in their best interest to speed technology ahead as fast as possible. Manufacturers first resonsability is to the shareholders, to make big returns on their investments, unfortunately, we, the customers are next in line. Canon, for instance, their main priority is to make a profit. They know perfectly well that if they make a camera that litteraly can’t be beat, even by themselves, then they will continue to product that camera for a long time, without replacing it. Like it has been previously mentioned, manufacturers make their money off of several products and a diverse market, which they would lose if they made a body that couldn’t be beat.

Now this is why, we have the concept of monopoly, and that no one company can control an entire market, but lets face it, Canon is running away from pretty much anyone. Chances are that they won’t be cought for a long time. Eventually, they will probably get cocky and make some bad products in a row, allowing for Nikon and such to catch up, but this, I am guessing, is at least 5-10 years down the road.

Do I think that digital photography will reach a plateau: yes, it will, but it will be a very short one, followed by another quick steep rise. Technology is so diverse that manucacturers will always find something new and better to add to their cameras. The same thing was said about PCs. Bill Gates once said that, no person will even need more then 128 Kbytes of memory (not sure if that is the exact number). Now, with photography, if you have a PC will less then 1GB, then you are hopelessly lost. If computers are any indication, then digital cameras will continue to get better, and we, the consumers, will continue to be amazed by how better things have gotten and will will continue to dish out the cash for them.
Logged
didger
Sr. Member
****
Offline Offline

Posts: 2030



« Reply #15 on: January 04, 2005, 03:47:33 PM »
ReplyReply

Quote
P. S. About my qualifications
OK, you've got me outdone in qualifications. Mine are similar level, but in biology, not math/physics. At this point I can't understand why larger format large size prints look so much better than smaller formats, but they sure do and DOF is not noticeably reduced in David Muench prints or Michael Fatali (8x10). If 22 Mpixel MF is no better for DOF or diffraction than 22 Mpixel 35mm would be, then nothing would ever be sharp in a large print, except stuff shot at f8 and with no DOF challenge at all. Large format would be even worse and using equivalent films would be no different from using equivalent sensor densities. 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? What good is it to have all that extra resolution if almost everything is out of focus if you shoot at large apertures due to extremely limited DOF or small apertures due to diffraction? In the middle you'd have extremely limited DOF and diffraction.

There's a serious conflict between real life and math here, sort of like bumblebees being poorly designed for flight.
Logged
John Camp
Sr. Member
****
Offline Offline

Posts: 1258


« Reply #16 on: January 05, 2005, 06:12:22 PM »
ReplyReply

Wouldn't all of Didger's problems be solved if he just got in his truck and backed up a couple of miles?
Logged
didger
Sr. Member
****
Offline Offline

Posts: 2030



« Reply #17 on: January 06, 2005, 07:03:40 PM »
ReplyReply

Quote
I assume from this that you found the print satisfactory?
Did you take a look at the CD included with the print?

I gather you didn't check your regular email recently.  I sent you an email practically the minute I opened the package.  I didn't look at the CD, though.  No need for the time being.  I doubt that I'll be making more prints of any sort for quite some time.
Logged
BobMcCarthy
Full Member
***
Offline Offline

Posts: 201


« Reply #18 on: January 07, 2005, 01:18:54 PM »
ReplyReply

Quote
Google "Neat Image" or "Noise Ninja" and you'll find well-established products than do exactly this, on converted images. ACR, C1, and most other RAW converters have some kind of noise reduction built in as well.
Jonathan,

I was thinking more along the idea of adding predictable information between the data sites (pixels). I have a projector with a line 4 X multiplier. It takes each scan line and then adds data based upon fairly complex algorithims as to that is the most likely data is between each line (and data point in the scan.

 It produces a remarkable experience.

The software I see out there is a start, but with number crunching power possibly even more capable processing can be performed with the convenience of being inside the camera. Fewer data points can create higher quality output. Some of what I am talking about is probably happening now. With more cpu and creative invention, it could happen.
Logged
didger
Sr. Member
****
Offline Offline

Posts: 2030



« Reply #19 on: January 08, 2005, 08:11:37 AM »
ReplyReply

Quote
And how exactly is reflecting that anger back at us going to help?
Huh?  Are you operating in some alternate Universe?  What in the world (this one) are you talking about?  Anger?

Your technical speculations are utterly incomprehensible to me and beyond even Sci-Fi, but maybe I'm just too dumb to understand (in this Universe,  anyway).
Quote
And in 3-5 years I can still see Didger shoot with a 1DsMK3
I reckon not.  I reckon I'll be shooting 1ds, MF digital, and maybe 4x5 digital.
Quote
And what does Nostradamus have to do with it? I don't recall him as a vehement example of someone who's only contribution to development is arguing something will never happen.
Again, I can't imagine what world you live in if you took light hearted whimsy and humor at literal face value and if you saw anything "vehement" or angry in my message at all.  I'm surprised you didn't find some serious intent in my "Alfred E. Neuman" reference.  Did you forget to wake up this morning before starting your day's activities?

Only time will tell about all these predictions and I can't imagine any reason to take any of this so seriously; a little break from image processing while it's too rainy, snowy, and cold to be out shooting.
Logged
Pages: [1] 2 3 ... 14 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad