Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: Digital Vs Medium Format Dilemma!  (Read 6476 times)
Ray
Sr. Member
****
Offline Offline

Posts: 8853


« on: October 05, 2002, 08:44:41 AM »
ReplyReply

Quote
[font color=\'#000000\']But based on what I've been reading about the EOS 1DS (not particularly the camera, but more the direction of digital) I wondering if it's possible that some day, image chips / sensors the size of a 35mm frame may exceed even formats as high as 4X5?[/font]
[font color=\'#000000\']Interesting question! I'd like to know the answer to that. I don't know if a good quality 150mm standard lens for 4x5 format has higher resolving power than an equivalent quality 150mm 35mm telephoto. I get the impression the sharper results from the larger formats is due largely to the film not degrading the image as it clearly does with 35mm. OTOH, some people are making comments that the 1Ds full frame sensor is already close to the limits of a good 35mm lens.[/font]
Logged
Marshal
Guest
« Reply #1 on: October 05, 2002, 08:50:04 PM »
ReplyReply

[font color=\'#000000\']I'd hold on to those L series lenses if I were you. 35mm digital may never reach the quality of 4X5 film, but I'd be happy to have 35mm gear that could go toe to toe with medium format. I think with these new digicams and future models in the next couple years that it definitely will. Hold on to your L glass.[/font]
Logged
Wim van Velzen
Sr. Member
****
Offline Offline

Posts: 354



WWW
« Reply #2 on: October 06, 2002, 01:54:30 AM »
ReplyReply

[font color=\'#000000\']I really don't understand all the 'digital will be better than mf'. That's mixing format and film vs digital.

Let's say digital 24x36mm is better than 24x36mm film - even as good as 42x55 film.
With digital 40x50mm and within a year or 2 55x70mm sensors: how good will they be? As good as LF film? And LF digital?

The larger the better is the same with digital as with film.

Of course one might use 35mm digital where MF film was needed untill now.

BTW, I would always keep a 35mm kit when going mf or lf. But a 6x12 is veeeerrrry tempting  Smiley[/font]
Logged

I don't have a signature.
David Lawson
Guest
« Reply #3 on: October 06, 2002, 09:25:15 AM »
ReplyReply

[font color=\'#000000\']Not sure if things are the same in USA but in UK it is not that easy to get this size neg printed, it is very specialist. I can see in the future, with pro's using more and more, digital this format may well become even more difficult to get printed as print based labs cut back, if they do keep the service for the few users there are it could also get expensive. Just food for thought really. david[/font]
Logged
Marshal
Guest
« Reply #4 on: October 06, 2002, 10:23:15 PM »
ReplyReply

[font color=\'#000000\']Someone mentioned Darwin Wiggett and it occurred to me I haven't heard much, if anything about him in a long time. He used to have a monthly column in some travel photography magazine, but I haven't seen that magazine or his work anywhere for several years it seems. Apparently, from what you say, he's still out there and shooting.

George Lepp is also using the Kodak DCS 645 back with his Mamiya I think. There was a good article a few months ago in one of the magazines about him going to Holland and shooting with the MF digital back as well as a D60 during the annual tulip festival. He took a group over there and taught some and used a laptop screen to show what he had just shot on site. This was his first trip that was strictly digital.[/font]
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8853


« Reply #5 on: October 06, 2002, 10:48:17 PM »
ReplyReply

[font color=\'#000000\']There are some interesting issues in this topic (for me, anyway!). I've always been a bit perplexed at the really small apertures that large formatters use. I understand the reasons why it's necessary to use such apertures - to achieve a decent depth of field because of the long focal lengths that are standard with large format cameras - but I can't quite reconcile this with the oft quoted adage that all lenses are equal at f8. Whilst this might not be strictly true, it seems to me that, as one increase the f/stop beyond f/8, the truer that statement becomes, so that at f/64 (the aperture that Ansel Adams favoured, I believe) the quality of the lens becomes almost immaterial.

In other words, distortion due to diffraction is a property of f stop irrespective of the design, quality and focal length of the lens. We're stuck with it and we can't compensate for it (except perhaps a bit by tinkering around the edges).

It seems to me, the consequences of this concept are far reaching. Large format photographers, in oder to achieve the required depth of field, end up chucking away the resolution potential of their lenses in large shovels full. I can't see that there will ever be a truly large format digital back for this reason alone. Clearly, there's no incentive to develop better quality lenses if 'real world' contraints require that existing large format lenses are used well below their diffraction limit.

Just in case some people reading this might think I have a screw loose (and indeed I might. At best I might merely be propagating the half-baked truths and misconceptions one often finds on the net), the reason large format is 'better' is because all these defects which are really serious on 35mm film, such as film grain, diffraction distortion and other aberrations, although larger in absolute terms on large format are smaller in relative terms.[/font]
Logged
Geoff Shiel
Guest
« Reply #6 on: October 07, 2002, 04:27:28 AM »
ReplyReply

Quote
[font color=\'#000000\']Well I would hold anyonewho thinks for one minute that 35mm digital equals MF with a pinch of salt. The article you refer to here while in some ways correct is also a little misleading. I for one have the top of the range Imacon scanner, not the bottom and the image and scan difference is huge to say the least. I still haven't found a digital 35mm that can match MF anywhere.

Mind you though the MF digital backs are very impressive but also very expensive. I know Darwin Wigget has the Kodax back for his Mamiya 6x4.5 and the quality does indeed exceed MF film at the same size, but saying 35mm does is a laugh.

So what to do.

You shoot landscape, why shoot it with a small format, a larger format will always outdo a smaller format. And don't go for a 6x12 go all out and get a 6x17 fantastic image, and I know of a Australian landscaper Ken Duncan

www.kenduncan.com.au

Who's images will blow your mind. I have seen his 6X17 images 1.5 meters long by .5 meter wide. Lets see digital do that[/font]
[font color=\'#000000\']What about the new canon eos1ds.
Compared to my scanned 6x4.5 fuji superior 100 negs, the couple of images from the canon have blown them away.
In a few words, no grain.
Scanner is a minolta dimage scan multi pro.
Bring on the canon.[/font]
Logged
Matt M
Guest
« Reply #7 on: October 15, 2002, 04:16:09 PM »
ReplyReply

[font color=\'#000000\']To answer the original post, a 35mm sized image sensor using current tech does have a limit in density of pixels. They can only etch chips so small. Moore's law has to break down eventually, like any other exponential growth. But, that said, it may not happen for a long time. Let's play with the numbers: 35mm is about 1.29 sq. in. 4x5 is 20 sq. in. 20/1.29~16 times the film area. 35mm is about equal to the latest digital cameras (see the report about digital iq). So, to equal LF, it will take 4 doubleings of today's resolution, or about 72 months; a paltry 6 years to wait. And in nine years, digital 35mm sized sensors should be equal to 8x10 in pure resolution.
   But lenses can't resolve that finely. At 60 lp/mm
you get 3600 resolvable pixels per mm^2. A 224 mp sensor (4x5 equivalent) 35mm sized, would be 31746 pixels per mm^2. That's an order of magnitude larger than what lenses can do. Put another way, 35mm sized sensors won't be able to get more resolution past 25.4 mp, because of lenses. Of course, this is all conjecture; the physical limits of the lenses could be different than what I estimated them to be.  Medium format, at four times the size of 35mm (6x7),   will enjoy 100mp sensors, assuming similar lens quality/design. That is still short of 4x5, and miles away from 8x10.[/font]
Logged
David Mantripp
Sr. Member
****
Offline Offline

Posts: 686


WWW
« Reply #8 on: October 16, 2002, 05:24:18 AM »
ReplyReply

[font color=\'#000000\']Well I'm here looking at a 4800dpi scan from Provia 100F on an Xpan, and a chance candid taken on my Olympus 4040 digital in "normal JPG" mode (i.e holiday snapshot mode).  Whilst the Olympus obviously yields a smaller print, what a print it is. The image is more or less useable up to 100%. On the other hand, any scan, however good I get it, cannot match that, and has to be down-sampled due to grain, noise, focussing imperfections, dust, whatever. And this takes _hours_.

Whilst I don't have the means to go digital SLR, I have to say I'm beginning to get the point of it... My "heartstopping"  Photokina announcement was the Hasselblad H1.  Now I'm not so sure.  Now I'm starting to look at EOS catalogues!

I don't really care a fig about all the blabla on lpm, real resolution, etc etc. That's engineer speak. I'm an engineer in the office. I'm a photographer at home, and to evalute photography I use my eyes.[/font]
Logged

--
David Mantripp
http://www.snowhenge.net
matt
Guest
« Reply #9 on: October 16, 2002, 11:39:33 AM »
ReplyReply

[font color=\'#000000\']as an addition to my earlier post, I did the figures for smallest pixel size for red light. 700nm is correct, no? in a 35mm sensor, the number is about 694 million pixels. That's assuming no space between pixels. So the wavelength wont be a limitation before lens characteristics. Anyway, I'll still be using film for a while; I like t-max (and have learned it too well) too much to give it up. maybe in ten years...[/font]
Logged
pedz
Jr. Member
**
Offline Offline

Posts: 77



WWW
« Reply #10 on: October 16, 2002, 03:45:26 PM »
ReplyReply

[font color=\'#000000\']I like the articles on this site that introduce the concept (or analogy) of noise and film grain.  In short, the differences we see between film and digital is due to a high level of noise that film has relative to good digital imaging chips.  Starting from here, things start to make sense.

The larger the film (going from 35mm to MF to LF) is an increase in signal.  This is because for a particular object (or piece of an object), there is more film dedicated to recording that object.  So there is more signal but the noise stays the same and ultimately we can get bigger prints before we begin to see the noise.  You can reason this out by saying that you do not have to multply the image as much OR you can reason this out by saying that you have more signal to start with.

The larger the pixel area of a digital sensor, the more signal to noise it will have.  But we already seem to be at a shockingly low level of noise.  Increasing the area should increase the dynamic range since there will be less noise (relative to the signal).  They can use 14 or 16 bit a-to-d's rather than 12  From what I can figure, if they really are giving a true 12 bit output, then moving to 14 bits of real and true output will give us two more stops of dynamic range.  Moving to 16 will be another two stops.

Michael wrote that the 1DS has about 6 stops of range but I'm wondering if that is true after I saw the detail that he could pull out of the man on the bridge.  I'm wondering if what Michael is seeing may not be due to output limitations rather than what the camera captured.

As far as the original posts asking if they should sell their EOS gear: you sound undecided so, give it time, and the answer for you will become clear.[/font]
Logged

Brian R
Newbie
*
Offline Offline

Posts: 10


« Reply #11 on: October 25, 2002, 03:43:50 PM »
ReplyReply

[font color=\'#000000\']The short answer to the original question is No! Larger chips are now out and I've tried them, and they still have a long way to go. Stick with your MF film and you will never regret it. I have been shooting professionally for 18 years, and for the last 5 years I have been shooting 85% digital. When I have something big to shoot i.e. landscape or archetecture, I use film. The quality is better! I have the advantage of being the director of photography for a fortune 500 company. That gives me no more photographic knowledge than any other photographer. What it does mean is that I have every digital back manufacturer comming to me to have me test their equipmint (In hopes I'll buy one). I've used most of them and my opinion stands firm.
 I should mention that I use a Scytex Leaf Volare (Best on the market for color), on a Fuji GX680 and I love it.
 A lot of people said they scaned their film and the digital looked as good or better. Ha! Now take that film to a service bureau and have it scaned on a laser drum scanner. The difference will astound you. there is no scanner on the market for under $30,000 that can hold up to even a medium range laser scanner.

Brian[/font]
Logged
Bruce Percy
Guest
« Reply #12 on: October 05, 2002, 04:10:34 AM »
ReplyReply

[font color=\'#000000\']Hi Fellow Photographers,

Thanks for the very interesting report on the EOS1DS Michael.

I have an interesting dilemma at the moment in that I have two camera kits - an EOS system with L series glass and a Mamiya 7II.

For the last year, I've noticed that I seem to leave the EOS in the bag and use the Mamiya II pretty much all of the time. I take landscape photographs as well as like to travel and the Mamiya 7II is great for both.

Anyway, I was in the process of trying to sell my EOS kit (28-70 F2.8, 70-200 F2.Cool so that I could fund a move up to 6X12 as I've been hankering after making larger prints for a while.

But based on what I've been reading about the EOS 1DS (not particularly the camera, but more the direction of digital) I wondering if it's possible that some day, image chips / sensors the size of a 35mm frame may exceed even formats as high as 4X5?

Or is there a physical limitation somewhere that I'm unaware of?

Perhaps I'm being sucked into the 'hold of till the next version before I buy' sort of thing, but I'm really now starting to wonder if I should hold onto my EOS kit and start to think about a move to digital within the next couple of years instead?[/font]
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8853


« Reply #13 on: October 05, 2002, 08:42:02 PM »
ReplyReply

[font color=\'#000000\']Just came across some comments from the Large Format Lens Test Review at http://www.hevanet.com/cperez/results.html . Here's an extract that seems relevant to your question, but I'm afraid there's a lot of confusion between 'real world' limits and theoretical limits. I have been of the view myself that, say, a 48 megapixel 36x24mm sensor would perhaps rival the quality of 4x5 and allow results that would reflect more closely the true aerial resolution of 35mm lenses. However, I'm beginning to have doubts because of the effects of real world constraints such as camera shake and imperfect focus.

..."Keep in mind, that due to the smaller negative size of the 35mm format,
you need to divide the 35mm on-film resolution numbers by 4 to get the
resolution required of a 4x5 lens for the same on-print resolution for
prints of equal size.  In other words, a large format lens capable of
delivering 25 lpmm on film, will be the equal to, in final print
sharpness, the best 35mm lens (using Brian Guyer's 100 lpmm maximum for
an ideal diffraction limited lens).  If you review the test results
Chris has posted, all but the very worst of the large format lenses we
have tested are capable of resolving 25 lpmm on film.  The ones that
fail, are usually older wide angles at the corners at wide apertures.
Many of them exceed 25 lpmm when stopped down to normal working
apertures.  A  couple of the really poor performers have obvious defects
(separation in one of the 90mm Angulons for example - keep in mind many
of these older lenses are 40 - 60 years old).  Of course, the best of
the modern 4x5 lenses (and several of the older ones, as well) far
exceed 25 lpmm, even in the corners, even at their "worst" aperture.
For example, look at the 110mm Super Symmar XL at f16, it is capable of
resolving, on film, 67 lpmm across the entire 4x5 field.  You'd need an
on-film resolution of 268 lpmm from a 35mm negative to get equivalent
sharpness in the final print.  Although the best of the current 35mm
lenses may be capable of such a high resolution at some apertures, they
will be not be capable of producing anywhere near such results on film
due to the limitations of currently available films"....[/font]
Logged
KAPhotography
Guest
« Reply #14 on: October 06, 2002, 03:23:30 AM »
ReplyReply

[font color=\'#000000\']Well I would hold anyonewho thinks for one minute that 35mm digital equals MF with a pinch of salt. The article you refer to here while in some ways correct is also a little misleading. I for one have the top of the range Imacon scanner, not the bottom and the image and scan difference is huge to say the least. I still haven't found a digital 35mm that can match MF anywhere.

Mind you though the MF digital backs are very impressive but also very expensive. I know Darwin Wigget has the Kodax back for his Mamiya 6x4.5 and the quality does indeed exceed MF film at the same size, but saying 35mm does is a laugh.

So what to do.

You shoot landscape, why shoot it with a small format, a larger format will always outdo a smaller format. And don't go for a 6x12 go all out and get a 6x17 fantastic image, and I know of a Australian landscaper Ken Duncan

www.kenduncan.com.au

Who's images will blow your mind. I have seen his 6X17 images 1.5 meters long by .5 meter wide. Lets see digital do that[/font]
Logged
Steve Kerman
Full Member
***
Offline Offline

Posts: 142


« Reply #15 on: October 06, 2002, 07:16:09 PM »
ReplyReply

[font color=\'#000000\']This discussion brings some interesting questions that I've been wondering about for a while.

A while ago I saw an analysis of lens resolution for different formats when you take depth of field into account. Certainly, when you compare 35mm with a 50mm lens at optimum aperture against 4x5 with a 150mm at optimum aperture, shooting a test target on a flat wall, the 4x5 will demolish the 35mm.

But, the story changes considerably when you set the lenses to give you the same depth of field. As we all know, to get much depth of field on a large-format lens, you have to stop way down. It turns out that when you graph the ultimate diffraction-limited resolution in lines over the entire image for a given depth of field, the curves for 35mm, medium format, and 4x5 are almost on top of each other.

If you think about this for a while, it actually makes intuitive sense. If you think about what size of an aperture you need to have to keep near and far objects in similar focus, you can visualize how you need to have about the same size hole for a short lens covering 35mm, or for a longer lens covering 4x5. So, you would therefore end up with a similar amount of diffraction in both cases.

The folks who wrote this article then went on to do tests of very high quality 35mm, medium format and 4x5 lenses using very high resolution B&W film (I recall that they used either Tech Pan or T-Max). And, indeed, you see on their photomicrographs of the negatives that they were getting similar resolutions in all three cases. (However, the grain in the 35mm case were quite noticable--the film still wasn't as good as the lens.)

My point here is that there may not really be a huge amount of difference in what digital sensors are capable of in the different formats. With film, the larger formats produce higher-resolution images to a substantial extent because film is the limiting factor. (I note that color film, with its multiple layers, has much lower resolution than the single-layer B&W film they were using in their tests.) As we get electronic sensors with finer and finer resultion, we may well get to the point that the larger "formats" don't really give you much benefit.

(In a similar vein, the recent discussion on this forum about the "four-thirds" format asserts that 35mm lenses, because of their back focus distance, are limitited to around 6 micron resolution, whereas the four-thirds format could get down to something like 2 microns.)[/font]
Logged
Marshal
Guest
« Reply #16 on: October 06, 2002, 10:26:57 PM »
ReplyReply

[font color=\'#000000\']Since George Lepp is a bigtime Canon shooter, I suspect he'll be getting a 1Ds and may have helped test it out for Canon already. I'm sure he'll get first dibs on one of the first production models.[/font]
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8853


« Reply #17 on: October 07, 2002, 10:05:46 AM »
ReplyReply

[font color=\'#000000\']Steve,
I seem to be repeating what you've already expressed. Didn't see your post at the time. There's also another advantage that large format has in relation to film. I believe film performs best at small apertures (f/22 and smaller) because of the dispersion characteristics of light hitting the film surface. I don't know if the 3 dimensional digital sensors are subject to the same influences in this regard.[/font]
Logged
Steve Kerman
Full Member
***
Offline Offline

Posts: 142


« Reply #18 on: October 15, 2002, 06:34:26 PM »
ReplyReply

[font color=\'#000000\']Actually, I don't think that silicon technology is limiting the pixel density. Our current production processes are running around 0.13 micron feature size--that's 130 nanometers. The wavelength of green light is around 500 nanometers. (We've long since stopped using visible light to image masks.) The pixel size of current full-frame 35mm sensors is around 8-9 microns. That's huge compared to most of the stuff we're currently building--we can build sensors that are smaller than the wavelength of light. In fact, I believe that some of the non-interchangeable-lens digicams do have substantially higher densities, although I don't happen know the specs on any of them.

I suspect that some of the following are factors that led to Canon's and Kodak's selection of pixel densities:

1. Lens resolution. As discussed by others, the current pixel size is approaching the point of dimishing returns as far as having useful information to record.

2. Already better than film. Similar to the preceding point, 8-9 micron pixels are already giving better results than the best film, so it's not clear that higher density would buy much additional market share.

3. Yields. While minimum feature size isn't a limiting factor, the overall size of the die is. 24x36mm is a huge piece of silicon. To keep your yield up, it helps to go with relatively large features.

4. Noise. This is been discussed in this forum. The bigger the pixel, the better your signal-to-noise ratio.

5. The rest of the system. Once you've captured umpteen millions of pixels, you have to do something with them. You have to read them off the sensor in a reasonable period of time, you have to do whatever processing you do to the image, and you have to store it someplace. There's clearly a big tradeoff between pixel count and other useability issues.

6. Risk. Jumping from 6 megapixels to full-frame 12-14 megapixels is a pretty big technological leap. It makes good business sense to limit the size of your jumps so that you don't get blindsided by factors that you didn't even know to look out for.

In other words, I suspect that it is feasible today to build an EOS-1Dxx with four times as many pixels. It would cost twice as much because of the lower yields, you'd get one frame per second, one-fourth as many frames in a burst and on a microdrive, the ISO-100 performance would look about like the ISO-400 performance of the EOS-1Ds, it would top at about ISO-400, and it's not clear that you'd be able to see a whole lot more detail than you get with the EOS-1Ds.

--Steve[/font]
Logged
Marshal
Guest
« Reply #19 on: October 16, 2002, 03:37:39 PM »
ReplyReply

[font color=\'#000000\']
Quote
I don't really care a fig about all the blabla on lpm, real resolution, etc etc. That's engineer speak. I'm an engineer in the office. I'm a photographer at home, and to evalute photography I use my eyes.
Couldn't have said it better myself DRM! Exactly!

BTW, with people like Pete Turner, Jay Maisel and Greg Gorman going digital as they have already, along with George Lepp and others too numerous to mention, that is the writing on the wall to me.

As for "maybe in ten years" this and that, it reminds me of something I said about four years ago. I was talking to someone during one of the Santa Fe Workshops one night and the subject of digital cameras came up. I said "maybe in 10 years digital cameras will be able to match film" in a skeptical, but polite tone. The guy, who I believe knew someone at Kodak in R&D said, "More like 5 years."

I was talking about digital vs film in general, but the unspoken or subconscious intent of saying at equal format sizes. 35mm digital vs 35mm film, 4X5 digital vs 4X5 film, etc. I believe he was on the same page with me. Anyway, I was skeptical then, but not now.[/font]
Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad