View Full Version : CROP COMPARISON (1.33x VS. 1.5x VS. FULLFRAME)
Zen-shooter
08-28-2006, 23:01
I've been noticing the lively discussion about M8 sensor size, 1.33x crop, etc. so I want to learn more...
This is what I understand...
Given the inherent design property of rangefinder lenses, we know that the best Leica can do with current sensor technology is a sensor that yields a 1.33x crop from a full frame (36x24mm) film size. We accept the notion that anything larger than 1.33x, the incident angle at the corners become too acute and the image degrades with severe vignetting and loss of sharpness.
This is what I assume...
So based on that, I calculate the 1.33x sensor size to be roughly 27x18mm. To illustrate, I've attached a graphical comparison of a full frame (in green), 1.33x (in magenta) and 1.5x (in blue) sizes to see the differences.
I'm sure that there will be many shooters who insists that full frame is the only way to go and anything less than that is below standard...
This is my conclusion...
But seeing what I drawn, I for one can accept the 1.33x as the benchmark size for digital rangefinder. I know that I will be utilizing the so-called "sweet spot" of the lens. Using the previous 50mm Summilux as an example, this lens design dates back to the Summarit 1.5, and at wide-open, the corners are often said to be quite soft compared to the center (film images naturally). Now if we crop the image to 1.33x, I can see that we would trim the perimeter softness out and keep the best part. On the flip side the natural effect of the D.O.F. would be lost.
As a side bar, here's more examples of current and popular DSLR sensors...
Canon 5D = 13.3M sensor, full Frame (36x24mm) Max. Res. = 4368x2912
Canon 1D Mrk II = 8.5M sensor, 1.3x (28.7x19.1mm) Max. Res. = 3504x2336
Nikon D200 = 10.9M sensor, 1.5x (23.6x15.8mm) Max. Res. = 3872x2592
So taking the proportions of the 5D, I get,
Leica M8 = 10M sensor, 1.33x (27x18mm) Max. Res. = 3284x2189
The calculation shows that the maximum resolution is lower than the 1D and D200 because the pixels are larger.
In a few weeks, we'll know all the facts...
Yes. A correct summary. However, due to geometrical reasons the loss of DOF is in the order of half a stop, so not really relevant.
patrickjames
08-28-2006, 23:09
The size of the pixels is important in sensor design because larger pixels = better image quality (less noise). Modern lenses are designed so that the light is transmitted at less of an angle to the sensor, which increases the quality at the edges. I wonder if the new Zeiss designs are like this? It would mean that a full frame Zeiss sensor would be inherently better quality than a leica M8 if Zeiss lenses were used. Zeiss has said that they designed their new lenses to be forward compatible with digital. This is an interesting statement.
shutterflower
08-28-2006, 23:22
Given the inherent design property of rangefinder lenses, we know that the best Leica can do with current sensor technology is a sensor that yields a 1.33x crop from a full frame (36x24mm) film size. We accept the notion that anything larger than 1.33x, the incident angle at the corners become too acute and the image degrades with severe vignetting and loss of sharpness.
I didn't know about the sharpness issue. I figured that maybe vignetting would be trouble in terms of light intensity but not clarity. This is news to me. Good to know. Maybe that cropped sensor isn't so bad after all.
Zen-shooter
08-28-2006, 23:23
The size of the pixels is important in sensor design because larger pixels = better image quality (less noise). Modern lenses are designed so that the light is transmitted at less of an angle to the sensor, which increases the quality at the edges. I wonder if the new Zeiss designs are like this? It would mean that a full frame Zeiss sensor would be inherently better quality than a leica M8 if Zeiss lenses were used. Zeiss has said that they designed their new lenses to be forward compatible with digital. This is an interesting statement.
Agree...
Regarding Zeiss, I posted something on that line yesterday. Their website's FAQ offering that clue...
Q: Will there be a digital ZI camera?
A: This may be a future possibility. However, we presently cannot comment on any details.
Q: Are the Carl Zeiss T* ZM-mount lenses ready for use with the future digital camera?
A: Yes.
Regarding the sensor, that was the reason I picked the 5D, the superior available-light high ISO capture ability; something us rangefinder shooter are quite sensitive to.
I still have to see a really convincing reason for a 35 mm sensor on a rangefinder. The only really relevant one, that wider lenses can be used is made far less relevant by the same argument Patrickjames uses. There is a whole crop of new extremely wide lens designs on the market. The pixel qualitty argument is overtaken each time by developments in technology, as the question starts to arise if any advancement in that field has any application in practical photography. The historical argument is obviously irrelevant. There remains the argument of getting used to the change on FOV. One would think that this would apply to switching formats between, for example between 645 and 6x9 as well and there has been no controversy over this since 1960.
An interesting aspect is that the smaller sensor allows the maker to keep the camera reasonably compact.
Zen-shooter
08-28-2006, 23:30
I didn't know about the sharpness issue. I figured that maybe vignetting would be trouble in terms of light intensity but not clarity. This is news to me. Good to know. Maybe that cropped sensor isn't so bad after all.
I added sharpness because light would have to travel further to reach the sensor (imagine the hypotneus of a triangle). The more it travels, the sharpness gets less; agree?
The best focus point / plane is the center of the sensor / film.
In order to have the entire picture be sharp and vignet-free, the rays have to hit the sensor perpendicularly (equal travel distance), throughout the surface. That means the sensor would have to be in a spherical shape (ie. being parallel to the rear lens element surface). Not practical and possibly impossible to manufacture in today's world.
Zeiss may have designed their ZI lenses to be more like Nikon DX series, by trying the "push" the rays straighter and perpendicular to the sensor... I get fantastic results from the 25mm and R-D1S...
Zen-shooter
08-28-2006, 23:33
I still have to see a really convincing reason for a 35 mm sensor on a rangefinder. The only really relevant one, that wider lenses can be used is made far less relevant by the same argument Patrickjames uses. There is a whole crop of new extremely wide lens designs on the market. The pixel qualitty argument is overtaken each time by developments in technology, as the question starts to arise if any advancement in that field has any application in practical photography. The historical argument is obviously irrelevant. There remains the argument of getting used to the change on FOV. One would think that this would apply to switching formats between, for example between 645 and 6x9 as well and there has been no controversy over this since 1960.
Excellent reasoning!
You deserve to be on the payroll of all 4/3 camera manufacturers.
Hmmm...may be the upcoming Digilux 3 isn't bad after all...:D
All these sensor size threads seem to be echo's of the past, when the miniature 35 mm format was introduced. The arguments are identical. Btw, Leica is in the process of introducing the 4/3 size S system aimed at the top end of the amateur and professional market.
Zen-shooter
08-28-2006, 23:52
All these sensor size threads seem to be echo's of the past, when the miniature 35 mm format was introduced. The arguments are identical. Btw, Leica is in the process of introduing the 4/3 size S system aimed at the top end of the amateur and professional market.
Agree 100%.
I like rangefinders a lot for just it's functions and portability but also for the memories had had in the past - call it subconcious romantic nostalgia. As a teen, I had a Nikon FM but dreamed of one day having a Leica. Now as a crusty middle-aged adult, I can get one but it wasn't for it's technical marvel.
Using automobiles as an analogy, just like American babyboomers buying up '69 Cameros and other muscle cars of the 60's, babyboomer phototogs have memories or aspired to own of Leicas. Unfortunately, digital is a whole different beast to film and some are not able to disconnect that. Automatically, they think Leica = 35mm...
I'm hoping that the new M8 will change that attitude...
As for the new Digilux 3, yes I saw Jorge's post yearlier today and compared it to the Panasonic.
Nachkebia
08-29-2006, 00:25
I'm hoping that the new M8 will change that attitude...
And what is the reason you want them to change it? Just to prove that digital technology is emature and unready to meat even 35mm size demands? smaller sensor you have bigger lenses you need, smaller sensor you have more noise you get, smaller sensor you have less control you have on DOF.... so on and so forth.....
I can live with the 1.33x cropping factor, but assuming the same lens is used a cropped sensor will always stress lenses more.
Any lens has a definite resolving power, so using a smaller sensor needs a lens' resolving power higher by a factor equal of the cropping factor.
Say a lens designed for 24x36 had 100 lpmm.
The same subject framed in a 1.33x cropped sensor should be resolved with a resolution of 133 lpmm to equal the full frame level of detail.
This is the same reason why lenses for 35mm format always had higher resolution than those for the medium format, and more so compared to the large format lenses.
With cropped sensor the only 'advantage' is that outer zones aren't visible, so no soft corners (for not-so-good lenses), less vignetting and maybe less distortion.
Even if both full frame and cropped sensor have the same total Mp (hence, assuming the lens outresolves the cropped sensor, the total amount of informations would be the same), the full frame sensor will nedd 'less' resolving power from the lens and we all know that the lower the spacial fequency, the higher the MTF.
If you read the MTF for 40 lpmm of a Leica lens you'll see what the lens delivers on full frame. On a 1.33x cropped sensor you should read the MTF for 53 lpmm, which is definitely lower.
In the end with all these outstanding M-lenses we should be fine with the 1.33x crop, but we know that whenever a full frame M will be released we will be even more happy and not only for the increased fov :)
So if I read your post correctly you want a large sensor to be able to put sub-standard lenses on your camera? ;):D
Nope :)
It's not about the maximum resolution obtainable, it's about the MTF at any resolution.
You always get better response on a larger sensor.
Anyway this won't hold me off the M8 !! :D
Zen-shooter
08-29-2006, 00:46
And what is the reason you want them to change it? Just to prove that digital technology is emature and unready to meat even 35mm size demands? smaller sensor you have bigger lenses you need, smaller sensor you have more noise you get, smaller sensor you have less control you have on DOF.... so on and so forth.....
Allow me to clarify...
I think we all agree that it's terribly difficult to adapt 35mm rangefinders to digital (speaking in terms of full frame and current lens line-up). The challenges that Leica is facing is well documented and discuss here and other places in great detail.
As such, I'm suggesting that we change our attitude towards full-frame DRF's and accept the fact that the M8 is the best solution. That's why I wrote that I can accept 1.33x. I bet that the images it produces will be acceptable to many.
If you want a full-frame DRF, then you really need to start with the lens. Remember, refracted light ray hitting the film/sensor perpendicular means no vignetting and sharp images. Zeiss knows about this and thus designed their ZI accordingly.
Nope :)
It's not about the maximum resolution obtainable, it's about the MTF at any resolution.
You always get better response on a larger sensor.
Anyway this won't hold me off the M8 !! :D
Well, we all know MTF is a limited tool anyway and OTF is too complicated to use, so we wil have to eyeball it in the end. And that is just my point. The results from even relatively small sensors , including the next generation of 4/3 are so good that they will surpass the quality obtainable in print orthe limits of the eye. It is like audio by now. Better means ten times the price an is meant for perfectionist geeks only. (my problem is that I tend to fit into that category :o )
Nachkebia
08-29-2006, 01:09
Well, I own only (only own) Zeiss glass, anyhow that is not the point, what ever the reason whole medium is right in the middle of development and it is a fact :) so justifying why leica can not come up with FF sensor is just bit infantile...
Nachkebia
08-29-2006, 01:11
jaapv : 4/3 will die in same suffering as Olympus will die (pitty for oly) just watch how sony`s gona eat them up... 4/3? common it is even funny at some point, sensor in those systems are smaller then the one on my mobile :D
Maybe. I would not go for it anyway, so it does not really bother me. But waiting for the next door to open, metaphorically speaking, does not make sense either. Let's just accept that the current crop of camera's is as good as it can be now, the future may or will or whatever be better or not as the case may be. So what? I want to take my photo's now.
AndyPiper
08-29-2006, 01:16
"The more it travels, the sharpness gets less; agree?"
Nope - the light intensity will fall off (as the square of the distance) which will cause vignetting independent of the sensor issues,or the aperture (which is why the 21 Super-Angulon for the M always had some corner darkening even on film and even at f/11). Sharpness will not be affected for any theoretical reason if the lens is corrected for flat field (which most are in this day and age).
While the analysis otherwise is OK, I really think this whole crop thing gets a bit "over-thought", especially when the crop is 1.5x or less.
The laws of optics don't change. A 21 will have more DOF than a 28 will have more DOF than a 50, and so on. Regardless of whether they are cropped or not. Worrying about whether a 28 f/2.8 will exactly equal a 31mm lens at f/4.356 is brain-damaging. Long lenses will blur stuff - short lenses won't. What further ANALysis is needed?
The M8 may or may not be "the best solution" - but it and the R-D1 are the ONLY solutions if one wants to shoot a digital RF.
If you want full-frame and rangefinder shooting - use film. If you want digital and rangefinder shooting - shoot cropped. Squalling and kicking one's heels on the floor like a spoiled child (not you, Zen) won't change those choices.
My 21 pre-ASPH is a lovely lens - it is still a lovely lens on the R-D1. It will be even more of a lovely lens on an M8. Same for every lens I own.
The only real downside to not getting a 24 x 36 sensor is that my "21" slows to f/4.5 (Cosina 15mm) - partially made up for by the fact that ISO 100 on digital (DMR tests for want of a real M8 yet) looks as good as Velvia 50, and ISO 800 on the DMR looks better than any ISO 400 film.
Zeiss may do something - sometime - over the rainbow - maybe. Doesn't help me take pictures today, or next month. If it gets here - then I'll see if it's "better" in some practical way.
Nachkebia
08-29-2006, 01:27
looks as good as Velvia 50 No digital photo on earth can look like velvia (or any other slide) proof is here ( http://www7.nationalgeographic.com/ngm/0609/sights_n_sounds/index.html ) :D
So what? I want to take my photo's now.
No problem at all, take photos, just don`t defend the whole system which is full of faults, and don`t forget you can take pictures with film and don`t forget market demand is created by you and you want to take pictures NOW! and see them NOW! and print them NOW! :D (I just bought 45 Kodak e100 vs`s, and don`t care about raw file at all) :)
No digital photo on earth can look like velvia (or any other slide) proof is here ( http://www7.nationalgeographic.com/ngm/0609/sights_n_sounds/index.html ) :D
No problem at all, take photos, just don`t defend the whole system which is full of faults, and don`t forget you can take pictures with film and don`t forget market demand is created by you and you want to take pictures NOW! and see them NOW! and print them NOW! :D (I just bought 45 Kodak e100 vs`s, and don`t care about raw file at all) :)
:confused: Nor does Velvia look like a digital photo. So what again. If you want the look of Velvia take Velvia photographs. If not take someting else. This issue is a comparison of different results, not a quality statement.
I think that the system of film has advantages and disadvantages, so has digital, so has a colloid plate or an albumin print. Why do you try all the time to pitch one against the other?
Nachkebia
08-29-2006, 02:04
Because one is eating the other one? :)
I don't think so. Film will be around for a long time, albeit in reduced circumstances. The Americans have a charming myth about my country, about a little boy, Hans Brinkers I believe, that put his finger into a hole in the dike to stop the flood. You seem to believe that works. We Dutch don't. :p
Nachkebia
08-29-2006, 02:10
In our country we say " Suger to your mouth" :D
Hi Nachkebia,
In our country we say " Suger to your mouth" :D
are you from the Caucasus? I had a Georgian colleague once who used to say this.
Philipp
Nachkebia
08-29-2006, 03:04
rxmd : I am georgian :D
No digital photo on earth can look like velvia (or any other slide) proof is here ( http://www7.nationalgeographic.com/ngm/0609/sights_n_sounds/index.html ) :D
Aha, a digitized slide which has been color balanced on a computer and converted to a colorspace suitable to a computer, retouched etc. on a computer display and then sent to a digital offset press is better than the output from a digital sensor which has been treated in the same way.
Actualy, I haven't seen neither a monitor nor a graphics card capable of more than the 8bits per pixel we call truecolor in the windows world an millions of colors in the Mac world.
While I can work with more than 3 colors plus black on a printing press, I'm reduced to Red Green and Blue dots on a computer screen, there is no such thing as light magenta or light black as they have in modern inkjet printers to expand the printable color space.
So my conclusion:
A projected slide is unbeatable by current digital technology.
A scanned slide has the advantage that you have something showing you what it should look like in print
A scanned C41 film has bigger lattitude and thus offers more possibilities to recover shaddows or highlights
If I shoot for offset printing purposes, digital offers an easyer workflow. color seperations from a slide or film where a PITA in the old days without computers and screening the color seperated films was even mo(i)re so :), with computers its just more work, C41 may have a limited advantage due to it's better lattitude.
For projection this is the other way around, there are no digital projectors with the resolution even the cheapest slide provides, so you have to get your digital files onto a slide, which is doable.
Shoot digital for printing and save the scanning step, shoot slide for projection and save the output to slide step.
Shoot B/W film because it's much more fun to develope B/W film than it is to desaturate digital files :D
And to the sensor sizes, a 24x36 sensor in a digital SLR makes much more sense than in a RF. The viewfinders size and brightness depends on the sensor size, all the APS-C dSLRs I know have darker and smaller viewfinders than those with bigger sensors or film.
I've compared a Canon 30d with a 5d in a shop, they feel and handle the same but one look through the viewfinder an I knew I have to save another 12 month for a new dSLR :-)
On a RF the viewfinder can be as big and bright as the designers want to make it, only restriction here is a magnification big enough for an effective base lenght of the rangefinder that it becomes acurate enough to focus long and fast lenses.
Some of the most insightful comments I have ever read.
Agreed with ALL you said.
Nachkebia
08-29-2006, 03:34
Socke, Agreed all you said but you are talking from technical side of view, I am more comming from aesthetic side of view, there is specific sentimentalistic mood that slide gives, simmulate that on digital I will be more then happy to use it, same with BW film :)
Let`s not make this topic film VS digital and not go to crystal grain discussion :)
Oh no, never intended to make it digital vs. film, I have good use for both and would not part with either :)
But a jpeg on the internet is a jpeg on the internet, it shows more of the skills of the operator than that of the film.
You need a slide projector and a silver screen to show the merits of velvia.
Nachkebia
08-29-2006, 03:57
But a jpeg on the internet is a jpeg on the internet,
I don`t agree with you on this one :)
I don`t agree with you on this one :)
Try jpg2000! Much better colorspace.
Nachkebia
08-29-2006, 04:53
Even gif file from slide is better, not technical :)
Even gif file from slide is better, not technical :)
Depends :D
Nachkebia
08-29-2006, 05:04
:eek: :o :angel:
I bet you like this backside better than the M8. :D. Not the first time I see this one Volker. I still wonder, you must be some kind of Adonis to be so persuasive...
As a side bar, here's more examples of current and popular DSLR sensors...
Canon 5D = 13.3M sensor, full Frame (36x24mm) Max. Res. = 4368x2912
Canon 1D Mrk II = 8.5M sensor, 1.3x (28.7x19.1mm) Max. Res. = 3504x2336
Nikon D200 = 10.9M sensor, 1.5x (23.6x15.8mm) Max. Res. = 3872x2592
So taking the proportions of the 5D, I get,
Leica M8 = 10M sensor, 1.33x (27x18mm) Max. Res. = 3284x2189
The calculation shows that the maximum resolution is lower than the 1D and D200 because the pixels are larger.
Uhh I don't get it.
The Nikon D200, with it's 10.0M pixels produces 3872 x 2592pixels. The Leica M8 with it's 10MP-sensor will produce the same amount. The size of the sensor doesn't affect the pixel count, surely?
Your calculation of 3284x2189pixels comes out to 7,188,676 pixels, 7.1MP. THe M8 should have a smaller pixelsize than the 8.5MP, 1.33Xcrop Canon, because it has a higher resolution than the Canon.
But then again I seem to miss the point of that table altogether. Please, could you explain?
Uhh I don't get it.
The Nikon D200, with it's 10.0M pixels produces 3872 x 2592pixels. The Leica M8 with it's 10MP-sensor will produce the same amount. The size of the sensor doesn't affect the pixel count, surely?
Your calculation of 3284x2189pixels comes out to 7,188,676 pixels, 7.1MP. THe M8 should have a smaller pixelsize than the 8.5MP, 1.33Xcrop Canon, because it has a higher resolution than the Canon.
But then again I seem to miss the point of that table altogether. Please, could you explain?
You are right, but this comparison is apples and pears, as it compares CCD's and Cmos', different generations etc. It does not mean anything, but it is always fun to play with numbers...
You are right, but this comparison is apples and pears, as it compares CCD's and Cmos', different generations etc. It does not mean anything, but it is always fun to play with numbers...
But it doesn't matter one bit whether it's CCD or CMOS, because pixels are pixels, and square inches are square inches.
The numbers were wrong in the first place. Or then I just don't get it.
It does not matter for resolution, but it does matter for result. And yes, there is a mistake in the numbers. :)
Nachkebia
08-29-2006, 06:03
Oh here he comes again with his canonite complexes, yeah yeah CMOS is the best :D :D
:confused: :confused: Please explain: am I pro or contra Canon? I don't really know myself you see. I am on my third Canon DSLR, and I have two CCD camera's. So I think I can say there is a difference in rendering between the two systems...
Nachkebia
08-29-2006, 06:14
I am just kidding, no worries :)
I bet you like this backside better than the M8. :D. Not the first time I see this one Volker. I still wonder, you must be some kind of Adonis to be so persuasive...
Not realy, this was on a truck on a techno parade and she's a professional dancer, ok, more than that :).
Every year a local security company has a truck on this parade and I happen to know the owner of the company so I'm on that truck as well.
The picture was taken on Sensia 100 in a Contax G2 with the 28mm, so I was close! I had to hang out of the truck like a sailer in hard wind holding myself with the left and the camera with the right hoping not to drop it while Beata steped on my left hand. Highheels hurt!
:confused: :confused: Please explain: am I pro or contra Canon? I don't really know myself you see. I am on my third Canon DSLR, and I have two CCD camera's. So I think I can say there is a difference in rendering between the two systems...
Yes, I often see differences between Nikon and Canon shots when I sit in the press room at the local six days bicycle race and have to decide which to publish and which not. Next to next in the same lighting the differences are obvious.
I can pick out a jpeg from a Canon 1DMkII against a jpeg from a Nikon D2x even if I get them at the same size, given they where shot in comparable lighting.
Even Canon 1D against Canon 1DMkII is possible.
When Canon came out with the 30d, not D30, everybody knew that Nikon was on the right track and CMOS will never provide low noise as the Nikon D1 did.
Somehow Canon found out a way of noisereduction which makes a current Canon at ISO1600 run circles around a Nikon, especialy when time is a factor and you get jpegs out of the PJs camera to publish in tomorrows paper or nearly realtime on a website.
From all I heard from engineers, CCD should still be better than CMOS.
I still have to see a really convincing reason for a 35 mm sensor on a rangefinder. ..........
An interesting aspect is that the smaller sensor allows the maker to keep the camera reasonably compact.
Thats really the key thing. There is not a lot of room in a M sized body to cram a FF sensor. More importantly is the continued use of the existing line of M mount lenses. The 1.3X sensor makes it a lot easier to get the kind of performance we expect out the wide angle range of M lenses.
The only down side, the crop factor, is ecceptable. It does mean that some new lenses need to be introduced on the wide end, but that isn't a bad thing for use geatheads!
Rex
Yes, I often see differences between Nikon and Canon shots when I sit in the press room at the local six days bicycle race and have to decide which to publish and which not. Next to next in the same lighting the differences are obvious.
I can pick out a jpeg from a Canon 1DMkII against a jpeg from a Nikon D2x even if I get them at the same size, given they where shot in comparable lighting.
Even Canon 1D against Canon 1DMkII is possible.
When Canon came out with the 30d, not D30, everybody knew that Nikon was on the right track and CMOS will never provide low noise as the Nikon D1 did.
Somehow Canon found out a way of noisereduction which makes a current Canon at ISO1600 run circles around a Nikon, especialy when time is a factor and you get jpegs out of the PJs camera to publish in tomorrows paper or nearly realtime on a website.
From all I heard from engineers, CCD should still be better than CMOS.
Not realy, this was on a truck on a techno parade and she's a professional dancer, ok, more than that .
Every year a local security company has a truck on this parade and I happen to know the owner of the company so I'm on that truck as well.
The picture was taken on Sensia 100 in a Contax G2 with the 28mm, so I was close! I had to hang out of the truck like a sailer in hard wind holding myself with the left and the camera with the right hoping not to drop it while Beata steped on my left hand. Highheels hurt!
Well, judging from this you must at least be more athletic than I ;)
On the Canon thing, I tend to think that the Canon look, whilst impressively smooth, has the most of that "digital" look film adepts fulminate against. Despite being attacked on this before, I do feel Nikon, when I borrow one, and even the small Digilux2 chip are a lot closer to the old-fashioned film-look. On the other hand my Canon prints, though the old 6Mp sensor lags behind in resolution, have a cleanness that I associate with mid-format.
AndyPiper
08-29-2006, 12:02
Perhaps I should clarify my comment (which was mostly an after-thought) regarding Velvia/digital etc.
I am a photojournalist and travel photographer - the final audience for my pictures is 10,000 or 50,000 or 200,000 people seeing the image in print. How a picture looks to 50 people on a gallery wall or to 50 people in a darkened room projected is meaningless for my art (although I do sell prints of my editorial work as aftermarket).
I am not interested in pictures that look like "Velvia" or "Kodachrome" or "Tri-X". I AM interested in pictures that look like what I saw and reproduce what I saw as clearly and truthfully as possible.
Any picture that will be printed (as in a book, magazine, or newspaper) these days WILL be digitized somewhere along the line. It will be a digital original file, or scanned from film.
It is my experience, after 30 months of working with digital cameras (Digilux 2, Epson R-D1, Sony R1) and trying VERY HARD to match their imagery with my film Leicas, that the extra generational step of scanning film (basically making a digital "dupe" from my slides or negs via an additional optical system) simply introduces too many artifacts (lack of focus in the edges, color fringes, grain, color shifts, dust, opaque shadows, etc. etc.) Once digital hit double-digit megapixels with a comparable lens, the contest was over.
For me, the only useful 'attribute' of Velvia is that it is virtually grainless - and digital RAW is virtually grainless with about 2 stops more dynamic range. Color neg films have better dynamic range than slides. But not significantly more - when scanned - than my digital RAW files. And negs lose a lot of edge definition and color clarity to grain.
All I have to do is look at a sky - which is made up of blue, blue and blue pixels in a digital original, and cyan, magenta, green, blue, and red pixels in a film scan - and the debate is over for me.
I've spent 30 years doing chemical photography, and enjoying it, because that was the means available. I studied and learned the techniques of silver so that I could control them as well as possible. But those techniques were never more than a means to an end - the images I publish.
Now that digital is here, and reasonably mature (if not perfected) - it has proven to be a better means to my end. It will be better yet when I can shoot with an M-sized camera and nice fast Leica-M lenses.
vBulletin® v3.6.8, Copyright ©2000-2008, Jelsoft Enterprises Ltd.