« Time for a Poll | Main | Oly 45 Arrives at TWH »

Saturday, 15 October 2011

Comments

Yes the RD-1 the only realistic possibility for a B&W sensor camera.

Kobayashi-san are you reading this?

Hi,

First, thank you for maintaining such an interesting site. TOP, Luminous-landcape and Leica forum are my 3 sites visited daily.

As far as a b&w digital camera is considered: there is already an excellent option in the PhaeOne Achromatic+ back - 39 mpx.

I am not sure of the usefulness of a small sensor b&w camera. Pre-digital, I mainly photographed with Leica M's and a Hasselblad 500CM using 90% or so b&w film. Switching to a Canon 5d, now 5d2, and more recently a Leica M9, I find that neither is an impediment to my b&w visualization, but also allow me to develop my colourist visualization --very different from each other.

So I would probably not be inclined to buy a b&w only camera, but then again, maybe!

Sincerely,
Jean-Michel Komarnicki


While remote, it's worth noting that Leica has reversed course before. The M8 was not supposed to happen. Then the full frame option was deemed impossible. The R10 was supposed to be the 'child' of the S2, but was cancelled. And so on. One never can be sure with Leica.

Leica has been known, too, to offer special editions, or a la carte versions, even when the market is a very small. But, that market is often Japanese, and the price is typically stratospheric.

I'm still hoping for the equally remote possibility of a digital M with manual shutter re-cocking, to avoid that obnoxious motor sound. I won't hold my breath.

There is another possibility (small) that I don't think I have seen mentioned. An after-market mod to a camera. Just as we now can have an after=market mod to permit IR and/or UV imaging, if an appropriate sensor was available, it might work. So if we have an older camera, now used as a back-up, or not at all, it might be turned into a B&W only camera for less than the cost of a new one. Obviously, a lot of dominos would have to fall correctly, but maybe...?

I think the best chance for monochrome digital is some kind of aftermarket conversion. I've seen at least one person who says you can do a monochrome conversion by removing the Bayer filter (and the microlenses, unfortunately) with solvent, so a conversion is at least theoretically possible. You'd ideally want to replace the AA filter at the same time, of course, but that's already possible.

The IR conversion business shows that there's a real possibility of making it work as a business. There are apparently plenty of people willing to spend hundreds of dollars to get single purpose IR cameras, even though it voids their warranties, mucks with their AF, and doesn't work quite right with existing firmware. I would guess there are even more people who would be interested in monochrome cameras under the same general terms. It just takes somebody bold enough to try making a go of it.

About "pole" vs "poll"...I think it was Fred Pohl who said that he and fellow SF writer Poul Anderson should get together and write a book about election methods in Warsaw. They would call it "Pohl and Poul's Pole Poll Poll".

You can't top that so I'll stop taking now.

Fuji hasn't made SR sensors (with the different sized sensels for some time now (4 years?). The current Fuji SuperCCD EXR sensor uses the same size sensels across the sensor.

The Fuji X100 (as I'm sure you know) despite having EXR trademark on the case uses a "conventional" APS-C sensor. Fuji's current "EXRness" on the X100 is closer to providing a decent UI for "ISOless" sensors. The other way to get better DR from low read noise sensors.

The better route might be a B&W X100 with a black and whitified current (Sony?) sensor. Or perhaps a B&W Ricoh GRD (with a fixed prime lens and snap focus it seems a better match than a G12/P7100). But it would need to be able to take filters.

A real issue with trying to use different sensitivity sensels is getting a color-neutral absorber (that would be new research and so not cheap) for those cells. And the usual issues of high sensitivity sensels with full wells (which is when you need the other cells) affecting the adjacent sensels. Perhaps more of an issue for an extended DR sensor than a regular sensor (as highlights bunch together).

Another possibility is making Bayer sensors with different dyes optimized for tuning B&W film response rather than matching the color receptors response of our eyes. Perhaps using four dyes and regarding each 2x2 set of sensels as a "real" pixel as the "4K" movie camera folks do (they don't count sensels they count pixels). But that drifts away from the "purity" of your original suggestion.

As for layout modifying the current Bayer mask for alternate rows for sensitivities (or colors) might be best as the eye is more sensitive to horizontal loss of resolution than vertical loss of resolution alternate rows of sensors might be a more sensible issue.

Oh and the "poles" spelling thing might just be overaggressive spell checking. Apologies if I missed a senseless "autocorrection" for sensel.

I wrote a post that disappeared into the Captcha vortex on the second Mono Sensor post. In it I theorized that a mono-only camera could be made using sensors that didn't pass QA for color use. My thinking is that a sensor that is mono only would be more tolerant of dead/stuck/hot pixels, and thus a camera maker could then make use of what would otherwise be wasted production. Further, this would be more cost advantageous on larger sensors, as their failure rate (I mean manufacturing tolerance here) must statistically be higher than smaller sensors. So maybe FF isn't a dealbreaker after all. Not being an engineer I'd be happy to be corrected about my assumptions re: yields and suitability.

In the vanished post, I also supposed that a manufacturer could use three sensors via a beam splitter and run each sensor in monochrome for superior color reproduction (and my thinking is that a camera w/ 3 2/3" sensors would offer superior image quality to an APS-C/m43 sensor of similar technology generation, at a lower cost (the green sensor could run at an internally higher ISO to help color performance). The video world is rife with 3 chip products. This isn't splitting the atom. I'm sure Ctein or Thom Hogan can reliably affirm or refute this idea re: sensor cost. But the 3 chip camera would have advantages for mono conversion in addition to giving Foveon type benefits (color/perceived resolution). Additionally, having three discrete color channels for mono conversion would mean that spectral response could better be tailored to different response curves, just like choosing different monochrome film.

Oddly, my enthusiasm for a mono-only camera is low in that I don't see benefit *for my needs*. It seems like a tough marketing proposition. What is the message (and how does the camera functionally differ from a color camera, aside from the obvious mono-only output)? These seem like non-trivial problems to surmount.

Patrick

"Larger sensors are more expensive to produce..."

Doesn't this assume the larger sensor's yield is the same as that of a smaller sensor? I would expect a larger sensor with larger photosites and supporting circuits would result in fewer defective parts, thus reducing the unit production cost. Perhaps?

You might be right. But still the result is that none of the choices in the poll applies to me although I'd be very interested in a B&W-only sensor. (And it seems I'm not the only one.)

And, FWIW, I for one said that I wanted a rangefinder. I don't care whether it's from Leica or another company.

I voted for mirrorless because they are the most versatile cameras around. You can put almost any lens on them. They offer characteristics of SLRs, TLRs (waist view)and view cameras in a compact package. And my humble G1 has the best eye relief of any camera I have ever owned--very important to this glasses wearer.

But I too would like to see a B&W rangefinder. Because when push comes to shove, I'm an RF photographer, and B&W and RFs go together like tea and biscuits. The main barrier to a digital Leica would be price. If a usable and affordable version by Somebody Else came out, I'd probably jump.

I'm also influenced by the desire for a tool with the basic controls readily accessible. And manufacturers, PLEASE don't forget eye relief. So many photographers wear glasses, and eye relief is often the last thing considered in a camera. One of the reasons I didn't buy an RD-1 when it first came out was that with glasses, I couldn't see the edges of even the 35mm (1.5x crop normal-lens) view.

Why don't those of us who would want a B&W sensor RD camera just ask Kobayashi-san at Cosina very nicely if he would oblige. What can we lose?

Maybe someday- interchangeable sensors!

If a black and white sensor is ever to be made, I suspect it will be APS-C and by Nikon, Canon or Sony...and Sony could be the main suspect. The reason simply has to do with an extension of their brand -- a lot of pros like B&W (or say they do.) If any of the FF-Three were to make an APS-C B&W camera, that company's professional standing would be enhanced, AND the new camera could use the full range of both FF and APS-C lenses that the pros already have, along with all the accessories.

Plus, if a young photographer is making the seemingly head-scratching choice between the FF-Three, why wouldn't you go with one that offered this other interesting possibility? Of course, if one company does it, the other two would probably follow...but the one that moves first could take a serious competitive edge.

I don't see why one of them wouldn't do it. We talk about cost and everything, but I don't see why a B&W would cost much more, and they're already churning out 3100s, 5100's, D90's, D7000s, D3s, D3xs...I don't think one slightly altered sensor, and some software, would make much difference...

Another thought, a question, to which I don't know the answer. Fred Picker used to sell a viewing glass that was sort of tobacco colored. You'd look through it, and the color would kill the external color, and you'd get a good idea of your B&W values. Would it be possible to come up with a screw-on filter of some (unknown) color that would produce a monochrome image that could simply be shifted to B&W in software, instead of a variety of tobacco-colored values? I know you'd lose a couple of stops, but that might be acceptable if it gave you true B&W.

The only reason the X100 exists is the sensor of the D90 (12 mpix also in Ricoh, and Leica X1) can be had cheap as chips. The last time Fuji made a sensor S-R was 6 years ago, and that was a rebranded D200 which cost $1800. Que the laugh track please.

We will never see S-R again due to economies of scale.

The D7000 and NEX now have a sensor that has equivalent dynamic range... the only advantage Fuji had in sensors, is long gone.

If Sony is selling NEX for $549, that tells you Fuji will be picking up that chip for the X200, since again, it can be had cheap as chips next year sometime.

Sony has an unassailable manufacturing advantage when it comes to sensors, the odds of Fuji getting back into the mix are the same as Leica producing a cheap camera. Even if they made a super sensor, Sony's would still be 1/4 the price.

I'm perplexed, why do you keep writing about black and white sensors as if they are just theoretical and a pipe dream when the Phase One Achromatic has been mentioned in the comments in each of these articles?

I'm not personally interested in a 'B&W only' sensor. Having said that it would seem possible to take the existing project to put alternate firmwares on Canon non-DSLR cameras (CHDK project) and make this only produce B&W images - you could then switch between colour and monochrome firmware when shooting. A bit like loading different films really. If there is no colour information saved that would seem to fulfill most of Mike's requirements - even though the camera is capable of colour recording if it only records mono there's no colour information there to be tempted by.

So it should be possible to get a monochrome recording only S95 or G12 by booting it from a card, if that's your bag.
I know about bayer filters but apart from that I don't get the desire for a monochrome sensor (aren't they all anyway?) when you can change the camera software to omit colour data and still have a fully functioning camera.
Would anyone have ever bought a camera that was designed not to be able to use colour film?

Perhaps Ricoh might do a B&W sensor. They have the knack of producing unconventional cameras. Quite a few people (including me) use the GRD in B&W mode. If you shoot B&W JPEGs at a ISO high enough to get noise (which, for the GRD I, is anything above 200) you get a lovely grain effect. If you shoot colour you get chroma noise which doesn't turn into grain when converted to B&W in post production. The GRD II didn't do this and there was quite an outcry, it was restored in the GRD III. So Ricoh do know they have customers who value shooting B&W digital.

Perhaps a version of their GXR Leica mount sensor unit?

"...Fuji's unique SR II sensor technology, which uses two sensels per photosite..."

This young ignorant would like to ask: "sensel" is an actual technical term? (I was wondering if it meant 'sensor cell' or somesuch.)

On the other hand, it would be nice if it's "two sensei per photosite" instead. O ye wise sensor!

By making the sensor color blind, you forsake the invaluable ability to tweak the color to gray tone conversion in post processing. And that exactly is the kind of wizardry I always dreamed about when using b&w film. In those days you had to decide on the filtering before the exposure and if you got it wrong, there was no method to fix it whatsoever.
So I absolutely dont get the point of a b&w only digicam. IMHO this is a pseudo-puristic approach with no practical benefit. Better get yourself a 4x5 view camera with some tmax film. And even on its ground glass you will see a color image. There is no way around it - to become a decent black and white photographer you have to learn to see gray where your eyes perceive color.

Markus

Bob Petersen, the number of sensors (or whatever chip is being made) per die depends directly on size, so even if yield is constant, larger ones are more expensive to produce.

I don't know how flaws within the big pixel wells affect a sensor; it seems like you really want it to be all working, so I'm not sure the yield won't fall along the usual curve. On the other hand, a hot or cold pixel here or there isn't that big a deal, so maybe they do get high yields. Sensors are, I think, enough different from ordinary chips to make a difference there.

Bill Hughes, CHDK is a great project, and in general I think open firmware on a camera has much bigger potential benefits that something like a B&W sensor. Having said that -- I'm not interested in a B&W only camera using a normal sensor; for my needs, I can do exactly what it does with my current equipment. But I am somewhat interested in a B&W sensor, because it could give better low-light performance than an equivalent-tech color sensor (fewer and larger pixels for the same real resolution, less dye absorbing light in front of the pixels, no color balance problems).

"sensel" is an actual technical term?"

Zeeman,
Yes--

http://en.wiktionary.org/wiki/sensel

Mike

hmmm.. interesting conundrum
However if the B/W camera were a FF, then there would already be a 35mm lens available for you to use with it. No hunting for a 35 equivalent required! A win win in my books

Sensel is used (by hardware people) to distinguish between the lax distinction between pixel (an full color (RGB) picture element) and a individual sensor element (either monochrome or with a colored filter).

There is also a minor distinction between photosite and sensel. Most of the time they're synonymous but in the Fuji SR case one might consider the two differ sensitivity photosites as part of the same sensel.

In still cameras this distinction between pixels and sensels has been blurred by marketing.

Witness the "how many pixels does a Foveon camera have?" problem. They actually have real RGB pixels but perhaps a third of the number an equivalent Bayer sensor camera will quote. So they make up a number in the same way the marketers of Bayer sensors do.

Perhaps the question should be how many pixels (not sensels) does a Bayer sensor really have. But I don't see that changing now. :-)

In the digital cine camera world they try keep the distinction straight but they've been having problems to. There is currently a lot of discussion over the next generation of "4K" cameras compared to current"2K" (HD) cameras i.e. cameras that can capture 4K pixel wide image with 4096 pixels or more.

In this case a pixel (should) always be real RGB pixel and is usually a 2x2 RGGB Bayer pixel i.e. 4 sensels though recently some marketers (like RED) have tried to fudge the issue.

For more on this "4K" argument see these Creative COW articles

http://magazine.creativecow.net/article/the-truth-about-2k-4k-and-the-future-of-pixels

http://magazine.creativecow.net/article/the-truth-about-2k-4k-the-future-of-pixels

To come back on topic this would be an issue for marketing pure monochrome still cameras.

Even though a hypothetical camera would have "twice" the resolution i.e. four times the pixels of a identical Bayer sensor with the same number of sensels that wouldn't show up in the comparative number of "pixels".

In actuality the resolution change would bit a bit less due to demosaicing in the Bayer sensor) but you take my general point.

The comments to this entry are closed.