Dedicated B&W camera anyone?

[dangoldman]

<p>If i want the look of true BW film... guess what i do... i shoot BW film! jeezes. Why spend hundreds of dollars for a one trick pony?</p>

[Brian1664876441]

<p>Just about every digital camera made is a one-trick pony. You're stuck with the CCD that it was built with. Of course, I can take the DCS200ir back off of my N8008s and load it with film if I want color.</p>

<p>For digital black and white- getting rid of the mosaic filter essentially doubles sensitivity of the CCD. Additionally, You can dump the AA filter as the mosaic filter is the cause of color-aliasing.</p>

[michael_oxford]

<blockquote>

<p>I'd certainly like to see a low cost B&W "digital rebel" type camera at a price under $1000, but I don't expect anyone will make one</p>

</blockquote>

<p>I think Bob has a great idea. Imagine if someone could make a DSLR B&W camera with say, a Nikon AIS mount for less than 1K, preferably around ~$600USD. It would be a <em>perfect</em> student camera for the "new millenium" ... although, in reality it would probably just end up as a cult classic thing.</p>

<p>Jeez, I'm just brainwashed by the results people get from cameras like the STL-11000 from Santa Barbra Instruments Group... honestley, full frame 16bit black and white is <strong><em>stunning</em></strong><strong><em>.</em></strong> Forget "versatility." If your style centers around black and white (mine does), nothing could be better. Insane dynamic range. Super high resolution. Super high ISOs.... whats not to like?<br>

oh yeah, it doesn't shoot color.<br>

All those rolls of Tri-X in my fridge dont either. <br>

wahoo.</p>

[michael_oxford]

<p>Joseph, I've been reaing through the DPreview thread and am wondering how you dissolve the bayer filter and microlens array... that is some hardcore stuff!</p>

[yog_sothoth]

<p>A film camera and a scanner are the best way to go for B&W quality beyond converting color digital pics. A good B&W digital camera needs more than 256 greyscale levels or an adjustable dynamic range to replace film. Being able to under expose and over develop or over expose and under develop to match the lighting and subject is not really doable with digital unless you take multiple exposures and combine the images. </p>

 

[oskar_ojala]

<p>I do think that the most likely way of getting am affordable BW DSLR to the market would be that some major manufacturer decided that they make a BW-version of some of their popular models. That way, R&D costs would be kept under control, but the camera would still need a long product life to be profitable, It's pretty hard to say how big the market for it would be, but small is a reasonable guess :-)</p>

<p>Joseph, have you been able to make more quantifiable measurements of the performance (sensitivity, dynamic range, resolution) of BW DSLRs? I'm just guessing in which ballpark the image quality gains would be.</p>

[aplumpton]

<p>A dedicated B&W digital would have great advantages. First, it would 3X the resolution of the same megapixel Bayer adapted sensor, as all sites are used for B&W alone, thus much better quality than colour to B&W conversion and probably better dynamic range. <br>

And if you wanted to duplicate film rsponse, Nik Efex Pro added software will do that.<br>

The Leica M8 or M8.2 SHOULD have that option, as should good DSLRs. </p>

[aplumpton]

<p>And if you wanted to duplicate film rsponse, Nik Efex Pro added software will do that.</p>

<p>Correction: And if you wanted to duplicate a variety of differing film responses and grain textures, the Nik "Efex Pro" added software will do that.</p>

<p>True, it won't replace medium format B&W film and scanning-printing or (better) traditional darkroom quality, but it would be an exciting option.</p>

<p>Surely, someone will do it, just a question of time. Just like many college photography courses are now regretting having dumped their darkrooms, as students often now want traditional processing to be available and profs are realizing that silver base photography is an excellent way to teach principles, to be sure that (how and) what is photographed is what is submitted. One magazine has reported that several multiple new orders have been placed for enlargers to retrofit some of these colleges. </p>

<p>I'm glad I purchased my Focomat IIc when a local university was trying to get rid of it. </p>

[joseph_wisniewski]

<p>OK, comments for Ross, Michael, Oskar, Stephen, and Brian, all in one large post (sorry, it's either that or post five times).</p>

<p>Ross, we have interesting views on the two sites. I thought this one was more negative, and the other more accepting. The other is certainly more polarized, although we've got a few in this thread that sure turned up the volume. ;)</p>

 

<blockquote>

<p>Michael - "Joseph, I've been reaing through the DPreview thread and am wondering how you dissolve the bayer filter and microlens array"</p>

</blockquote>

<p>This is dangerous stuff: I don't recommend you trying it. Heat, solvent (MEK) and wedging with a blade to breech the sensor (remove the cover glass). A bit of scraping to clean the casing where the new glass will go. I used more heat, MEK, and Toluol to dissolve the filter and lens array on the D100. Iliah Borg used Toluol and Carbon Tet at room temperature on the D2X. I'm all for room temperature methods, heated aliaphatics like MEK or Toluol are scary. Again, this is just for discussion, I recommend not trying such things yourself.</p>

 

<blockquote>

<p>Stephen - " A good B&W digital camera needs more than 256 greyscale levels or an adjustable dynamic range to replace film."</p>

</blockquote>

<p>This is true, but irrelivant, as all modern DSLRs use at least 4,096 greyscale levels (12 bits). Some (Nikon D300, one of my conversion candidates) uses 16,384 levels (14 bits). I've tested the Fuju S3 (an IR conversion using a deep 80B filter to produce a monochrome response) and a well done conversion can yield 128,000 levels (19 bits).</p>

 

<blockquote>

<p>Arthur - "Just like many college photography courses are now regretting having dumped their darkrooms, as students often now want traditional processing to be available"</p>

</blockquote>

<p>Do you have any evidence to support that? Most university folk I know cite lack of interest by the students as the reason they discontinue the darkrooms. We're a private institution at MPW, and we saw our darkroom enrollment decline to the point where we did not get enough students to run a darkroom class in two years, at which point we made the decision to discontinue. We have had very few inquiries on the subject. We've even discontinued our hybrid classes like making digital negatives.</p>

 

<blockquote>

<p>Brian - "If Leica is not considering one for a version of the M8, I doubt anyone else will do it. Would solve a number of technical issues with the RGB Mosaic Filter and limited sensor to flange distance."</p>

</blockquote>

<p>Actually, the sensor to rear element (not flange) problem is solved by the microlens design. The microlenses are necessary to increase sensitivity, and you'd keep them, even in a monochrome conversion. Even MF cameras have microlenses now, sacrificing performance with shift lenses and view camera bodies in order to increase sensitivity. Except for my own conversions (I can't dissolve the color filters without first dissolving the microlenses) any monochrome camera will have them.</p>

 

<blockquote>

<p>Oskar - "Joseph, have you been able to make more quantifiable measurements of the performance (sensitivity, dynamic range, resolution) of BW DSLRs? I'm just guessing in which ballpark the image quality gains would be."</p>

</blockquote>

<p>I have measurements for a generation 1 camera (Nikon D100) but they're essentially meaningless when we consider third generation cameras like D300 or D90. The D100 is so dated that I don't want to share the "hard numbers". I can give you the "executive summary".</p>

<ul>

<li>resolution exceeds twice that of a Bayer color pattern with monochrome conversion (monochrome D100 outresolves color D2X).</li>

<li>dynamic range is improved by about a stop between the monochrome D100 and a monochrome conversion from a color D100.</li>

<li>sensitivity remained nearly constant, the loss of the microlenses negated the loss of the filters.</li>

<li>but on a correlary to sensitivity, the camera is usable about a stop higher, an ISO 1600 from the monochrome camera looks better than an ISO 800 B&W conversion.</li>

</ul>

<p>Next steps, I absolutely hav to do a better camera.</p>

[Brian1664876441]

<p>Wide-angle lenses for RF's are certainly a problem. I use a J-12 on a Contax II.<br />What is the relative thickness of the Mosaic Filter in a modern CCD? Getting rid of it increases sensitivity by a factor of 2x. The light fall-off as you get off-center of the CCD for a given lens should be predictable, and fairly easy to correct. I guess it boils down to the 2x gain vs light-falloff if you give up the microlenses. So- dump the Mosaic filter and Microlens, and apply a non-uniformity correction. Now, if you could keep the microlenses, the fall-off is not as bad, and is still correctable for a given lens.</p>

<p>Should add that my camera was $12,400 in 1992. The IR version was an extra $4,000 for a custom run of the KAF-1600 CCD. $600 does not seem so much in comparison. Fortunately, my camera still works.</p>

[jpo3136b]

<p>Amateur astronomers seem to have access to lower cost monochrome digicams. I think the average CCD starcam I've seen advertised runs under $400. Many of the models seem to plug directly into a laptop for controls; so, it's not quite the camera most photographers would expect; but, I bet someone into astronomy would know about the performance differences. </p>

<p>Apparently a lot of the CCDs out there are already good enough to photograph nebulae. If that's not good enough performance, what is? </p>

<p>The color astrophotos are pretty; but it seems because of the way the matrices are set up in the sensor that astrophotographers get an advantage out of using monochrome. Idea here is that the light sources appear so small that they may not span as many pixels in a matrix. </p>

<p>Really into black and white? You know I'm going to tell you to use film. The operating costs should come in under $12,400. Good luck. J.</p>

[michael_oxford]

<p>There is quite a bit of difference between conventional and astro cameras, but I'm just going to hit on some main points that come to mind. 1st (and probably most importantly) astro cameras have no IR filter (thus they get optimal response across the whole spectrum) 2nd The best cameras are monochrome for a reason - to take color photographs, they shooting a tricolor method (luminance, red green, and blue exposures). You can also apply this method with narrow band filters such as as Ha, Hb, and OIII for example. This makes them invaluable for scientific research because your camera has twice the normal resolution (no bayer pattern) and you can specify what you exactly want to see.</p>

<p>Astro cameras can also be water cooled to reduce noise and the best generally have >75% Quantum efficency (next to films measly 6%!). Really, if you wonder why digital cameras are better, there is no better place to look than astophotography. (granted, you can do miracles with Kodak Tech Pan... but they don't make it anymore...)</p>

<p>Thats why i get so jacked up about digital black and white. Because I know what it's ultimatley capable of.</p>

[yog_sothoth]

<p>"This is true, but irrelivant, as all modern DSLRs use at least 4,096 greyscale levels (12 bits)"<br>

How do you use them? the RAW software displays the numbers from 0-255. When I use "48 bit" color in CS2, I sample areas and get 0-255 for all colors. If the A/D converter has more bits, we still don't have those bits available to us.</p>

<p> </p>

[pete_myers]

<p>

<p ><strong>Kodak DCS 760m (monochrome)</strong></p>

<p >http://www.luminous-landscape.com/reviews/cameras/kodak-760m.shtml</p>

</p>

<p > </p>

<p >

<p ><strong>Enough Already!</strong></p>

<p >http://www.luminous-landscape.com/essays/enough-already.shtml</p>

</p>

<p > </p>

<p >

<p ><strong>Making Images---Not Taking Images</strong></p>

<p >http://www.luminous-landscape.com/essays/making.shtml</p>

</p>

<p > </p>

<p >I purchased one of the 80 digital monochrome cameras made by Kodak in the DCS 760m series and shot it for 18 months. It was a great experience, but alas support for the project at Kodak ended and I went back to film.</p>

<p > </p>

<p >Since that time, I have been active with a number of camera back makers and camera makers to try to get monochrome versions of their cameras into the market. Unfortunately, it is very difficult to get past the engineering mentality that a black and white image can be obtained from a bayer based camera with the same fidelity as a true digital monochrome. I can say directly from my experience that is complete nonsense. Perhaps a decent B&W image can be made from a bayer based camera, but it will not even be close to the same as obtained from a true digital monochrome.</p>

<p > </p>

<p >Leica backed out of making an M8m version when the previous president came in and nixed the project. Quite sad really, as it would have propelled the M8 to super star status in my opinion.</p>

<p > </p>

<p >PhaseOne is currently building a 39mp monochrome back on a limited basis---but the price point is so great that I certainly cannot afford one.</p>

<p > </p>

<p >The D3 is the perfect camera to go monochrome. Will it happen? Not likely. But it would be an astonishing image-makers if it did.</p>

<p > </p>

<p >cheers,</p>

<p > </p>

<p >Pete Myers</p>

<p >Fine Arts Photographer</p>

<p >Santa Fe, NM</p>

<p >

<br />

</p>

[oskar_ojala]

<p>Thanks Joseph, definitely a clear improvement. Even a D60 would be much more potent camera than the D100 and from a practical standpoint live view would be great (I assume you intend to use the camera for more than test shots after conversion :-)<br>

I would expect that those camera makers that already have huge volumes would be the best candidates to introduce a BW camera, since they can expect the largest market. But OTOH it seems that it takes a certain mentality in a company to come out with new, risky products (take Cosina's Voigtländer-lenses as an example).</p>

[stephen_asprey2]

<p>I have a dedicated B&W camera, full frame, full B&W dynamic control, solid construction, idiot proof controls, user over ride on all settings, ISO to 6400, works without batteries, has multiple bright screens available and is small. The images come in any format and up to 4000 dpi.<br>

Its called a Nikon FM2n shooting TRX, and a Canon 810 all in one printer to scan the negs.</p>

[chris_werner]

<blockquote>

<p>"Amateur astronomers seem to have access to lower cost monochrome digicams."</p>

</blockquote>

<p>Most color astrophotography is done using RGBL filters on a monochrome sensor and stacking the results (the "L" is for Luminance). What is also common is to use special filters to isolate the wavelengths of certain elements common in nebulae and galaxies, such as Hydrogen Alpha, Hydrogen Beta, Sulfur, and Oxygen III. In either case you end up with a series of images that are monochrome, but which individually represent relatively or extremely narrow ranges of wavelengths (RGB filters are fairly broadband, the others are narrowband, typically with a bandpass of 6-10 nanometers). They are then stacked in the computer with colors applied based on the wavelengths they represent.</p>

<p>We are increasingly seeing astrophotos with DSLRs, but red is a problem color for them, and cooling of the chip is hard or impossible to accomplish. Having a cooled chip is really important in astrophotography as the signal/noise ratios are so low that the increased noise of chips at higher temperatures is a problem. By way of example, I shot most of my astrophotos with the chip cooled to -25 Celsius.</p>

<p>Black and white astrophotography is also very common.</p>

<p>The cheaper, sub $1000 models generally have less sophisticated cooling and smaller chips, smaller chips representing challenges capturing enough field of view.</p>

<p>That's probably too much information, but what the heck. :-)</p>

[aplumpton]

<p>Joseph: I don't have a lot of evidence on the renaissance of the darkroom, but here in Quebec City a few of the camera stores are receiving a reasonable influx of young customers for darkroom chemicals, paper and (of course), film. They no doubt make use of both silver-based and numerical photography. From what I've read, some enlarger manufacturers at Photokina (Kienzle) apparently are seing two types of customers, the under 30 crowd (admittedly not all students) and the over 60. One company, making computerized light sources for enlargers (Heiland), is seeing similar business.</p>

<p>There is apparently a renewed interest in large format photography. This makes sense for both the B&W and colour photographer bent on high quality at reasonable price. Compare the huge costs of a MF digital system and the quality that can be obtained from LF (or even MF) film, a relatively inexpensive scanning device and/or an even less expensive darkroom set-up for B&W.</p>

<p>I have read that some colleges are realizing that the only way to control work assignments is to use the traditional methodology of image capture and reproduction, as it often confines the student to "what you shoot is what you get", without post production manipulation, and manual film cameras also allow a good teaching of the basics of light and imaging.</p>

[bernardwest]

<blockquote>

<p>"This is true, but irrelivant, as all modern DSLRs use at least 4,096 greyscale levels (12 bits)"<br /> How do you use them? the RAW software displays the numbers from 0-255. When I use "48 bit" color in CS2, I sample areas and get 0-255 for all colors. If the A/D converter has more bits, we still don't have those bits available to us.</p>

 

</blockquote>

<p>The 12 or 14 bit data is there, it is just displayed in an 8 bit histogram. There are other programs around which can show you a histogram of the full cache of data.</p>

[michael_oxford]

<p>It is really nice to see Mr Myers chiming in on this. I read the original article on the 760m and have browsed through the other articles... very interesting. It seems especially noteworthy that someone who has shot both says the bayer converted b&w is rubbish...</p>

<p>hmmm.. So Mr Myers, do you have any other examples other than the ones on the Luminous landscape pages? I would love to see more examples from anyone that has a dedicated B&W digi.</p>

[p_mui]

<p>The true bit depth can be found by the full well capacity divided by the noise. See page 8 of this document, lower right hand corner.<br>

http://ccd.com/pdf/FullProductLine.pdf<br>

According to the SBIG site, the full well depth for the 11000 chip is 50,000 electrons. RMS noise of 13 electrons. This will give it slightly under 12 bits of resolution. Remember that this is for the sensor itself, monochrome chip. Yes, it may be a 16 bit converter at the end of the chain, but the sensor itself is only capable of slightly less than 12 bits.<br>

http://sbig.com/sbwhtmls/online.htm<br>

Another interesting link from clarkvision.com - the monochrome Kodak sensors are usually only good for 12 bits. Scroll about halfway down the page.<br>

http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary/<br>

None of the Fuji sensors are listed, but it would be interesting to find out the well depth and the noise readout in electrons. This would give a true rendition of the amount of bits that are coming out of the sensor in one shot. The record holder for sensors from Kodak is the 1001, with a huge 300,000+ well depth. That sensor is only about a megapixel, but the bit depth is slightly over 15 TRUE bits from the sensor.</p>

[dan_llewellyn1]

<p>We have done some work on a monochrome camera conversion. You may find some of our tests interesting. We compared a stock Canon 30D to a 30D we modified to be strictly monochrome.</p>

<p>http://www.maxmax.com/b&w_conversion.htm</p>

[dan_llewellyn1]

<p>We have done some work on a monochrome camera conversion. You may find some of our tests interesting. We compared a stock Canon 30D to a 30D we modified to be strictly monochrome.</p>

<p>http://www.maxmax.com/b&w_conversion.htm</p>

[p_mui]

<p>Dan,<br>

Excellent stuff! How difficult was it to remove the bayer array from the sensor? I have a 20d sitting around that may benefit from the conversion...<br>

I once tested out a monochrome prototype of a Leaf back that had the Kodak 16801 chip. The camera itself was all aluminum, no black paint or anything. Images from it were VERY good. This was around 2003/2004.</p>

[dan_llewellyn1]

<p>P Mui,<br /> Thanks! I have been thinking about removing the CFA for a few years. There are lots of technical reasons why this is rather difficult to do well. I have been watching the efforts of Joseph W and Iliah B. They have worked on this for many years with limited success.</p>

<p>I have modified a number of personal cameras so far: Canon 300D, 350D, 450D, 500D, Nikon D5000, D70. I am not ready to modify anyone else's camera yet. Too much risk at this stage. I have killed enough sensors that I would rather buy a new camera, convert it and then try selling it than having to deal with explaining to someone that I toasted their sensor. I am waiting on some custom parts and more experience, but I am getting close. You can see that the 30D monochrome camera is taking some really sharp pictures. Compared to the stock 30D, the monochrome camera simply blows the color camera out of the water - as long as you want a B/W image. I suppose you could take 3 pictures with RGB filters with the monochrome camera and then get a really high res color picture as well. An experiment for the future.</p>

<p>We are trying to be as careful as possible to quantify the differences. You can download many of the test images from the site in their original RAW form for your perusal.</p>