This one is as unsharp and saturated as you can get.
menos
Veteran
I would.
I don't care so much about color lately and enjoy shooting my EPSON R-D1 and the film Leica bodies B&W.
When I feel the urge for color, I grab a Nikon DSLR.
I think, it has a lot to do with the subjects and situations, I shoot in with range finder cameras, which do interest me more in B&W than in color.
The discussed (elsewhere?) technical advantages of a true B&W sensor opposed to the now used Bayer sensor in digital cameras is also of interest to me.
In the ideal world, I could get a 12 − 16 MP 24x36 mm sensor with the low light capabilities of my Nikon D3 and a resolution way high enough to do A3+ prints.
I would win on every account.
If technology would move forward further and come closer to the wonderful transitions and dynamic range, Tri-X (even pushed) brings me in high lights (think softly rendered transitions and details of streetlights in night scenes ) - I would grab a credit card and get a new M9 monochrome in a heart beat.
I don't care so much about color lately and enjoy shooting my EPSON R-D1 and the film Leica bodies B&W.
When I feel the urge for color, I grab a Nikon DSLR.
I think, it has a lot to do with the subjects and situations, I shoot in with range finder cameras, which do interest me more in B&W than in color.
The discussed (elsewhere?) technical advantages of a true B&W sensor opposed to the now used Bayer sensor in digital cameras is also of interest to me.
In the ideal world, I could get a 12 − 16 MP 24x36 mm sensor with the low light capabilities of my Nikon D3 and a resolution way high enough to do A3+ prints.
I would win on every account.
If technology would move forward further and come closer to the wonderful transitions and dynamic range, Tri-X (even pushed) brings me in high lights (think softly rendered transitions and details of streetlights in night scenes ) - I would grab a credit card and get a new M9 monochrome in a heart beat.
Gabriel M.A.
My Red Dot Glows For You
So predictable. At least the word "Nikon" wasn't included in the bashing.
If Leica did produce a Monochrome M9, it would be the most distinctive digital camera on the market. It would also find buyers that want it for photography, not some dust-collector.
alecgold
Established
I don't think software is trivial. The cost of software is usual gigantic for these kind of applications. Not just because it is comex, but look at the M8 how many updates it took to become good, stable and functional.
While I'm at it, what is wrong with the m8/m9 BW pictures? If you want to blow them up to 17x22" you would be wise to use film. You can shoot a lot of film instead of a $10.000 body. And Much better quality.
But for 6x8" a m8 or m9 makes nice pictures.
And I hardly can believe a BW m9.1 would make a wise investment for a professional because the ROI wouldn't be there. There Are (not much) assignements where they pay you double or more to get 13 stops worth of BW pictures. So to me it would seem like a nostalgius camera, but way to limited compared to what a regular m10 would offer in the future. And probably at less than half the price. Or less.
aizan
Veteran
http://www.luminous-landscape.com/reviews/cameras/achromatic.shtml
it doesn't look like monochrome sensors really are much better.
it doesn't look like monochrome sensors really are much better.
No moire in the 100% crops. No interpretation errors that I get in the 100% crops of the M8. The looks a lot better to me.
How many times do photographers load color film when they intend to shoot all black and white, and that is without moire issues.
A monochrome sensor is simply better for producing monochrome images than a color sensor. It's that simple. Monochrome film is better at producing monochrome images than color film, again that simple. Arguing over that fact is futile. There are opinions, and facts. You can state any opinion that you want, but the facts remain the same.
Tell you what- you put up some side-by-side images made with your color digital camera, and then "black-and-whitized", compared with your images from a monochrome camera. I'll bet there will be less artifacts introduced from the color interpretation process.
How many times do photographers load color film when they intend to shoot all black and white, and that is without moire issues.
A monochrome sensor is simply better for producing monochrome images than a color sensor. It's that simple. Monochrome film is better at producing monochrome images than color film, again that simple. Arguing over that fact is futile. There are opinions, and facts. You can state any opinion that you want, but the facts remain the same.
Tell you what- you put up some side-by-side images made with your color digital camera, and then "black-and-whitized", compared with your images from a monochrome camera. I'll bet there will be less artifacts introduced from the color interpretation process.
Last edited:
The software was trivial for the Kodak monochrome camera that I own. The current selection of monochrome/IR backs runs ~$40,000. They are made, and sold. So a $10,000 camera would be a cheap alternative. My Kodak was $12,400 for monochrome/IR. Kodak made about 50 of them, sold to the scientific market.
If anything, this poll shows that a larger percentage of Leica lovers would buy Monochrome cameras than those that bought monochrome Kodak cameras when they were made. That's because more Leica and RF owners load their cameras with black and white film and like to produce black and white images.
The nay-sayers on this thread, I suspect it's because they put out $7K once for a Leica M9 and would hate the idea of someone else producing better monochrome images with an M9m. Then, they would have to buy one as well. Pretty much, that's what we did when Kodak offered color and monochrome versions of the camera, bought one of each. Ran a total of $20K. That was the first time that I spent $20K on Digital gear, but not the last. Of course, it was for work.
The nay-sayers on this thread, I suspect it's because they put out $7K once for a Leica M9 and would hate the idea of someone else producing better monochrome images with an M9m. Then, they would have to buy one as well. Pretty much, that's what we did when Kodak offered color and monochrome versions of the camera, bought one of each. Ran a total of $20K. That was the first time that I spent $20K on Digital gear, but not the last. Of course, it was for work.
alecgold
Established
Brian, monochromes are better at monochrome if they are having the same sensitivity, the same noise, the same pixel size. But what I know from the monochrome sensors is that they are less well developed than the color sensors. There is a huge market potential for a good color sensor, but only a marginal for BW.
And I wonder why there are no retro-fitters for bw sensors to place it in a m8 or m9?
The cost of replacing a sensor and rewriting the software on an excisting model must be much much lower than the development of a new camera.
And RF is the ideal platform for retrofitting as you have no moving parts other than the shutter.
And I wonder why there are no retro-fitters for bw sensors to place it in a m8 or m9?
The cost of replacing a sensor and rewriting the software on an excisting model must be much much lower than the development of a new camera.
And RF is the ideal platform for retrofitting as you have no moving parts other than the shutter.
Last edited:
percepts
Established
I don't get the arguments in this thread. All sensors are designed to respond light wavelengths. There is no wavelength for white or any shade of grey. They are derived from measuring the range and intensity of the primary wavelengths. There is just intensity of any wavelength. So that means any sensor has to interpret what is hitting it and then process that to output. Surely that therefore means the it is the processing behind the sensor that detrmines what the output will look like and not the sensor itself. And that means if the firmware and/or software in current cameras is written to suit B+W film curve output then no special camera is required. And if the firmware and software can do that then so can photoshop or other third party software.
I suspect that most of the R&D and software development goes into colour and B+W is just added as a nicety without trying to mimick any particular B+W film curve.
The suggestion that a special B+W sensor is requireed implies that the sensor would be sensitised to different wavelengths than current sensors. But which wavelengths? Just about anything related to this is capable of being modelled by software so I fail to see why a special sensor would be needed.
I suspect that most of the R&D and software development goes into colour and B+W is just added as a nicety without trying to mimick any particular B+W film curve.
The suggestion that a special B+W sensor is requireed implies that the sensor would be sensitised to different wavelengths than current sensors. But which wavelengths? Just about anything related to this is capable of being modelled by software so I fail to see why a special sensor would be needed.
Last edited:
jaapv
RFF Sponsoring Member.
No - you got that wrong - a B&W sensor would be without the Bayer filter that assigns colors to be interpreted. That would certainly make a difference. The question discussed here is whether that difference would be significant, and whether it would offset the loss of control over the image that the presence of color channels makes possible.
DCS200ir, Micro-Nikkor 55/2.8, R60 filter, hand-held.
This is a 100% crop with my monochrome/infrared camera.
When I get some free time this Spring, I'll drag it and the M8 out to get some comparison shots. On the M8, artifacts creep in on small changes as shown here.
Of course, in the Infrared, the M8 does not suffer moire as the RGB dye used pass IR about equally. Another reason to use the M8 as an IR camera.
and the software for this image was trivial. Stored in Raw, took about 100 lines of FORTRAN to produce the image out of the RAW file. I wrote it in a day.
This is a 100% crop with my monochrome/infrared camera.
When I get some free time this Spring, I'll drag it and the M8 out to get some comparison shots. On the M8, artifacts creep in on small changes as shown here.
Of course, in the Infrared, the M8 does not suffer moire as the RGB dye used pass IR about equally. Another reason to use the M8 as an IR camera.
and the software for this image was trivial. Stored in Raw, took about 100 lines of FORTRAN to produce the image out of the RAW file. I wrote it in a day.
Last edited:
jaapv
RFF Sponsoring Member.
Sorry, you lost me there, Brian. Why should there not be moire? It is basically an interference pattern caused by the regular structure of the pixels.I can see that there would not be color artifacts, but the resonance type in structures should be present. It might be less obtrusive, as a lens will usually render less sharp in IR ( a natural AA filtering) but for instance the Apo-Summicron 90 asph is sharp enough.
Arjay
Time Traveller
BW = higher resolution, less noise & more cost-efficiency
BW = higher resolution, less noise & more cost-efficiency
What makes an imaging sensor a "color" sensor is the Bayer filter pattern that is uperimposed on the sensor, essentially putting a blue filter on 1/4, a red filter on 1/4 and a green filter on 1/2 of all sensor pixels.
These filters
Since the physical layout of every sensor pixel is three-dimensional, with vertical walls between the photosensitive sensor sites, the pixel has a relatively narrow angle of light sensitivity. Any light hitting the sensor at a shallow, oblique angle will be shaded off by the walls between the photosites! This produces vignetting and can lead to color shift problems with a distribution similar to that of vignetting.
This is a problem with camera lenses like in the M system whose exit lens icome relatively close to the sensor, thus having a relatively large illumination angle from one edge of the sensor to the other. To compensate this in a color sensor, the component is not only equipped with the Bayer filter pattern layer, but with another layer of microscopic lenses on top that concentrate the light so that it hits the sensor's photosites at steeper incident angles. To make matters worse, this layer of microlenses cannot have an even distribution like the Bayer filter pattern, but the placement of the lenses must be slightly offset depending on their distance to the camera lense's optical axis in order to take into account the different light incidence angles, and to avoid light "spill-over" into adjacent photosites. Microlens placement becomes increasingly difficult and demands ever-higher accuracy as the camera lenses exit lenses come closer to the imaging sensor.
All of this makes color image sensors for RF cameras much more complex than those used in today's DSLRs.
Of course, any color image can be converted to BW via software, and this process offers superb contrast control, but producing a color image in the first place requires to have a very sophisticated, expensive image sensor. Removing the Bayer filters (and possibly the microlenses) would obviously save a lot of costs, permitting the camera maker to offer a camera at a substantially reduced price tag. And since many RF camera users shoot BW anyway, all of this boils down to giving up some control in software-based contrast control during BW conversion in favor of a less costly camera.
The question remains, however, whether this loss in control isn't an excessively high price to pay for a less costly camera.
I personally feel that this price is too high, because the ease and power of software-based BW conversion is a very powerful advantage that speaks for color sensors.
Anyone who is moving to digital from a film-based camera may likely not be aware of the processing ease and control that he would give up by just shooting BW images instead of shooting in color and converting to BW on his computer. It's like using multigrade positive paper in a conventional wet lab without the complexity that used to be connected with it.
BW = higher resolution, less noise & more cost-efficiency
Any digital camera's sensor is essentially a BW sensor in that it is sensitive to all light wavelengths across the visible spectrum (and a little more).
What makes an imaging sensor a "color" sensor is the Bayer filter pattern that is uperimposed on the sensor, essentially putting a blue filter on 1/4, a red filter on 1/4 and a green filter on 1/2 of all sensor pixels.
These filters
- convert the pixels into color-sensitive elements which selectively and exclusively respond either to red, blue or green light,
- reduce the sensor's overall resolution to roughly 1/3 of its initial value, and
- increase imager noise unevenly across the color spectrum due to the 1/1/2 distribution of the number of R, B and G pixels (green pixels offer both a little higher resolution and lower noise due to their higher number). This makes noise reduction via on-board software much harder.
Since the physical layout of every sensor pixel is three-dimensional, with vertical walls between the photosensitive sensor sites, the pixel has a relatively narrow angle of light sensitivity. Any light hitting the sensor at a shallow, oblique angle will be shaded off by the walls between the photosites! This produces vignetting and can lead to color shift problems with a distribution similar to that of vignetting.
This is a problem with camera lenses like in the M system whose exit lens icome relatively close to the sensor, thus having a relatively large illumination angle from one edge of the sensor to the other. To compensate this in a color sensor, the component is not only equipped with the Bayer filter pattern layer, but with another layer of microscopic lenses on top that concentrate the light so that it hits the sensor's photosites at steeper incident angles. To make matters worse, this layer of microlenses cannot have an even distribution like the Bayer filter pattern, but the placement of the lenses must be slightly offset depending on their distance to the camera lense's optical axis in order to take into account the different light incidence angles, and to avoid light "spill-over" into adjacent photosites. Microlens placement becomes increasingly difficult and demands ever-higher accuracy as the camera lenses exit lenses come closer to the imaging sensor.
All of this makes color image sensors for RF cameras much more complex than those used in today's DSLRs.
Of course, any color image can be converted to BW via software, and this process offers superb contrast control, but producing a color image in the first place requires to have a very sophisticated, expensive image sensor. Removing the Bayer filters (and possibly the microlenses) would obviously save a lot of costs, permitting the camera maker to offer a camera at a substantially reduced price tag. And since many RF camera users shoot BW anyway, all of this boils down to giving up some control in software-based contrast control during BW conversion in favor of a less costly camera.
The question remains, however, whether this loss in control isn't an excessively high price to pay for a less costly camera.
I personally feel that this price is too high, because the ease and power of software-based BW conversion is a very powerful advantage that speaks for color sensors.
Anyone who is moving to digital from a film-based camera may likely not be aware of the processing ease and control that he would give up by just shooting BW images instead of shooting in color and converting to BW on his computer. It's like using multigrade positive paper in a conventional wet lab without the complexity that used to be connected with it.
retnull
Well-known
Not sure if this has been mentioned yet, but considering the question purely from the viewpoint of "branding" -- a B+W M9 makes sense. No other company manufactures a digital B+W camera, so Leica would gain press and attention. And it would reinforce Leica's pedigree: the fact that Leica is known for its history of producing the tools used for creating the greatest, most famous B+W images ever.
percepts
Established
But if you take out the bayer filter then you get a mix of all wavelengths hitting each pixel and what that suggests is that only a total intensity of light can be measured/accummulated from all wavelengths hitting each pixel.
B+W film doesn't work like that and you will have lost the ability to interpret the the light in the way B+W film does. Therefore you will need a filtering system anyway and since you already have one why would you remove it? i.e. you need to know what wavelengths you are getting and not just the intensity IMO.
p.s. and if there is an argument that you can post process it to make it look like B+W film then you can do that anyway from the current M9 or any other digital camera. This whole disucssion is about being able to replicate what B+W does currently so by default you will have to sotware process the file to achieve that. It seems to me that its a question of where that processing happens. Doing it in camera is a ridiculous concept because there are many films to replicate and more importantly everyone uses different developers for different times and then you print where more controls are applied with burning and dodging.on different papers with different developers. The camera can't do that for you so you will have to process in an image editor anyway so you might as well apply a B+W film filter at the same time.
B+W film doesn't work like that and you will have lost the ability to interpret the the light in the way B+W film does. Therefore you will need a filtering system anyway and since you already have one why would you remove it? i.e. you need to know what wavelengths you are getting and not just the intensity IMO.
p.s. and if there is an argument that you can post process it to make it look like B+W film then you can do that anyway from the current M9 or any other digital camera. This whole disucssion is about being able to replicate what B+W does currently so by default you will have to sotware process the file to achieve that. It seems to me that its a question of where that processing happens. Doing it in camera is a ridiculous concept because there are many films to replicate and more importantly everyone uses different developers for different times and then you print where more controls are applied with burning and dodging.on different papers with different developers. The camera can't do that for you so you will have to process in an image editor anyway so you might as well apply a B+W film filter at the same time.
Last edited:
jaapv
RFF Sponsoring Member.
No- why would you want to know the wavelength to have a B&W image. Just the luminosity suffices.
Last edited:
alecgold
Established
the 100% crop is nice, no doubt about, shot from the hand it looks perfectly sharp. But why do you 100% crop? Do you print huge formats?
If you print at A3, a M8 would do very nicely as would most (if not all) current dSLR's.
If you do print billboard size or larger, no body cares about 100% crops-details, because you don't stand within an inch of a billboard. You can't see what it is, if you're that close.
I understand the will to reach perfection, but when you have a Leica M9, a superb Nikon or a good Canon, what could you want more? If film is that much better (and it is still better than digital) and you need that resolution why not shoot your BW in medium format film?
jaapv
RFF Sponsoring Member.
Different - yes - better - no. The same for digital. I wish people would stop comparing the incomparable...
