New DRF's via rumour mill

Roger Hicks said:
Dear Frankie,

And get a share of a market that the anti-M8 brigade maintain is too small to support ONE manufacturer?

Besides, look at the price of Nikon's high-end cameras -- and this would have to be a high-end camera. How much cheaper could it be than an M8?

A sub-$2000 Nikon DRF body has to be idle fantasy. In fact, sub-$4000 is probably pushing it.

Cheers,

R.
Click to expand...

What I wrote had probably touched a nerve in many. Some won't believe the market is big enough, some cannot fathom the product being cheap...

I didn't crave a cheap product.

However, I do know one thing from personal experience:
  1. Building a good product for uninformed consumers...and they will still complain, RTFM is too much work. [Making it intuitive takes real talent.]
  2. Building a good product and establish a competitive benchmark price for prosumers has risks...what if they didn't come? [Tell marketing to think.]
  3. Building a good product and charge what it takes, "when it has to be right" and all that...might only attract the religious. [Go preach.]
The Nikon way was: build it in a modular way and sell add-on pieces and a future following...the F, F2, F3, from eyepieces to lenses. You all know the history.

And, parts bin engineering is a quick (and cheap) way to get started...just study some of the carry-over bits from the SP to the F and F2...from shutter button to rewind crank. So take the D3 chip, EXCEED, whatever new electronic bits and the recently re-polished millennium rangefinder, and start adapting lenses.

I am not obsessed about the M-mount, I only have one such lens...the CV 40mm/1.4; nor the F-mount either, despite I have an arsenal of lenses...none of them would likely work in a DRF. I am looking at Nikon only because they can...but why didn't they?

Reality was, many well-known brands failed the transition, from Minolta to Contax. Nikon was well advised to be cautious, but that period was over. The D side had won and a new era had dawned.

Sometimes in business, you have to create your own market. The original M4 did. Picking on the M8 or lamenting the demise of the RD-1 is not the point. The real point is: does Nikon want to remake and own the future DRF market...AND do we want a decent DRF or not?
 
Dear Frankie,

Your points about 'parts bin' engineering are well taken, and it is not hard to redesign a camera with the enormously expensive, unnecessarily complicated and extremely restrictive Contax/Nikon mount (absurdly small throat for modern fast lenses) to take the far more useful M-mount (far simpler to build, much bigger throat, enormous lens choice).

If they went fot a different mount, sure, they could design a series of telecentric lenses in a new mount, and all they'd need in the way of focal lengths would be 12-15-16/18/21-18-21-24-25-28-35-35/50/90-40-50-75-90. At speeds from f/1 to f/4...

Even so, I have two reservations. One is why Nikon would bother, given the very high cost of and reliance on processing an FF image with a shallow flange/sensor distance.

The other is that I'd worry that Nikon would make the camera for a few years; possibly (though not probably) put Leica out of business, because RFs are Leica's bread and butter, not an ego trip where they can afford to lose money; and then quit, because the SD (or whatever they called it) wasn't profitable enough in the context of their much bigger DSLR business.

Cheers,

R.
 
Last edited:
Roger Hicks said:
Even so, I have two reservations. One is why Nikon would bother, given the very high cost of and reliance on processing an FF image with a shallow flange/sensor distance.
Click to expand...

Dear Roger,

help me walk through an example, and explain to me where I'm wrong:

Assume the “cosine fourth” illuminance falloff function holds in absence of micro-lenses, etc.

Take a 21mm lens in 35mm format. FOV is around 90 degrees, maximum off axis angle is 45 degress. power(cos(45 degrees), 4) is roughly 0.25. I.e., we need a gain of factor 4 to correct this (theoretical) vignetting. Reduce the sensor sensitivity by factor 4, correct the vignetting in dependence of the angle in (in-camera) software. Makes you loose 2 bits of the original sensor output.

For a focal length of 16mm, power(cos(53 degrees), 4) is roughly 0.13, correctible in software by sacrificing 3 bits (max. factor 8).

Given a new, sensitive enough sensor, this is just a software implementation problem, no ?

Roland.
 
Last edited:
Double Negative said:
The recent rumblings about the Micro 4/3rds format lenses is interesting for several reasons indeed. Personally, unless they move away from the current 2x crop factor of regular 4/3rds, I don't know that I'd really be interested. I've done the 1.6x thing, and it's not my thing (I lean towards wide angles) and am currently enjoying 1.3x as a reasonable compromise against FF.
Click to expand...

Others have commented about the "2x crop factor" not being accurate for 4/3s systems, so I won't add to that.

The exciting things about m4/3 to me is

  1. The promise of more compact kit, both body and lenses, is now more achievable
  2. WA and extreme WA lenses become more reasonable in both size and cost; I would surmise fast prime wides are now more achievable

If a mechanical/optical RF is not offered but a really good Live View implementation is all that's offered, I would be interested, but a true RF is really where I want it to be

I don't care if the sensor is smaller than "FF", I think we (most of us, at least) are just emotionally attached to that size. Only for very demanding pro application would we need 35MM sensor size. Hell, I have both 4x5 and 35mm, and you don't see me using 4x5 exclusively because of the comparitive microscopic size of 35mm.

If Tom can get a 35mm FL FOV on a 4/3s sensor, even he might go for it! Leica is in the 4/3s lens business, and I wouldn't be surprised to see them jump into this pool.

At this point in the sensor technology, only a radical jump in capability will make the 35mm size feasible for M mount. Personally, I think Leica should ditch M mount for new development. It would give them the opportunity to improve bottom line performance, because with all that "legacy" M glass out there, they are forced to make money on new digital bodies, and they really, really can't compete due to the size of their market share.
 
Trius said:
At this point in the sensor technology, only a radical jump in capability will make the 35mm size feasible for M mount.
Click to expand...

That radical jump already happened, Earl. Leica or not, from a customer perspective competition is always good. The M mount offers backward compatibility like only the M42 mount.

Roland.
 
Last edited:
ferider said:
Dear Roger,

help me walk through an example, and explain to me where I'm wrong:

Assume the “cosine fourth” illuminance falloff function holds in absence of micro-lenses, etc.

Take a 21mm lens in 35mm format. FOV is around 90 degrees, maximum off axis angle is 45 degress. power(cos(45 degrees), 4) is roughly 0.25. I.e., we need a gain of factor 4 to correct this (theoretical) vignetting. Reduce the sensor sensitivity by factor 4, correct the vignetting in dependence of the angle in (in-camera) software. Makes you loose 2 bits of the original sensor output.

For a focal length of 16mm, power(cos(53 degrees), 4) is roughly 0.13, correctible in software by sacrificing 3 bits (max. factor 8).

Given a new, sensitive enough sensor, this is just a software implementation problem, no ?

Roland.
Click to expand...

Dear Roland,

Whew! Given that it's after dinner (Ilagares '07 and Bordeaux '04) I'll need to check the cosines, but assuming you're right (I have no reason to assume otherwise, except that I'm not sure Snell's law is not overly generous to digital sensors), it seems do-able -- ignoring noise constraints and the cost of developing the software. There's also the colour fringing issue, but I'm not enough of an optical engineer to argue with you here.

Tell you what: I'll try to get a better answer (from Leitz, Zeiss, Cosina and Kodak) at photokina.

Cheers,

R.
 
ferider said:
They are in an efficient sensor architecture. For cost reasons.

PS:

[GEEK ALERT]

- in signal processing, given a black box, its dynamic range (also the "signal/noise ratio") is the ratio of largest to smallest signal that can be processed. In analog systems measured in "deciBel (dB)", in digital systems measured in bits.
- in signal processing and information theory, #bits is only a unit to measure information quantity.
- if you look at the entire sensor chip as black box (light in, digital bus out), the sensors DR is indeed the #bits, including possible control bits used to add/subtract sensor gain. To be more precise, the #bits needed to represent all possible sensor outputs with a uniform binary code.
- if you split the sensor into image input device (for example CCD or CMOS transistors), A/D converter and digital output processing, the system better be balanced, with the input device, A/D converter and output logic having similar DR. Otherwise components of the architecture are either over or under-designed.

[/GEEK ALERT] 🙂
Click to expand...

Thanks for the lesson! However, I rather suspect that the sensor technology is what limits DR, because adding bits (i.e. precision of the measurement of the signal) is relatively trivial. The system is balanced, I would imagine, when you have enough bits to cope with the DR the sensor can put out, plus a margin to allow for post-processing, maybe. I freely confess I'm not an expert on this, but I have studied a bit of information theory, quite a long while ago. On your other issues, I would agree with Roger and you that "de-vignetting" is indeed theoretically just a software issue, and not a difficult one, but the limiting factor (I would guess) is the sensor technology (i.e. noise issues etc as Roger mentioned). Whether or not the software should apply the cos^4 law or not is not really important, because if done in camera it would probably apply a data table based on actual fall-off rates and radius from image centre.
 
Roland, without the microlenses, the sensor doesn't even see the image from extreme angles. There is nothing to fix. It is blind to that part of the image because the receptors are effectively in a hole.
 
ferider said:
Dear Roger,

help me walk through an example, and explain to me where I'm wrong:

Assume the “cosine fourth” illuminance falloff function holds in absence of micro-lenses, etc.

Take a 21mm lens in 35mm format. FOV is around 90 degrees, maximum off axis angle is 45 degress. power(cos(45 degrees), 4) is roughly 0.25. I.e., we need a gain of factor 4 to correct this (theoretical) vignetting. Reduce the sensor sensitivity by factor 4, correct the vignetting in dependence of the angle in (in-camera) software. Makes you loose 2 bits of the original sensor output.

For a focal length of 16mm, power(cos(53 degrees), 4) is roughly 0.13, correctible in software by sacrificing 3 bits (max. factor 8).

Given a new, sensitive enough sensor, this is just a software implementation problem, no ?

Roland.
Click to expand...

Roland,

Your theory ignores the fact that the sensor has depth and its face is not its imaging surface. There is a secondary correction at the edges of the sensor where the micro lenses over the pixel wells correct the light paths.

Two issues face anyone designing a full frame sensor for a rangefinder. The first is the increase in correction by the micro lenses as the sensor gets wider. The second is the publics thirst for ever higher pixel counts. This condenses the pixels on the sensor and reduces the pixel well width, resulting in a need for even greater correction in the micro lenses at the outer edge of the sensor.

Some people seem to believe that because Canon and Sony mass produce full frame sensors, it will be resonably easy to make one fit a full frame digital rangefinder.

The two sensor designs will be very different as are the current Leica M8 sensor and the Canon ID Mk3 sensor (both APS-H format).

Regards

SR
 
Roger Hicks said:
Dear Frankie,

"The other is that I'd worry that Nikon would make the camera for a few years; possibly (though not probably) put Leica out of business, because RFs are Leica's bread and butter, not an ego trip where they can afford to lose money; and then quit, because the SD (or whatever they called it) wasn't profitable enough in the context of their much bigger DSLR business."



Dear Roger,

I was speaking "frankly" and not as a Leica worshipper.

Leica did a lot of good things...also many disappointing things. If Nikon can light a fire under their collective behind, and caused them to wake up...that can only be good.

Whether RF is Leica's bread and butter or not, if they cannot even face a belated challenger, then death is warranted. If Leica dies, then Nikon would have stayed to harvest the vacant market...until perhaps CV repents, if not already too late.

And, if DRF is indeed a good idea, then the consumers would make up the mind for Nikon...whether DSLR is bigger or not.

I am not new to Leica but new to RFF, and I have read a lot of religious preaching...as if Leica is god, and must forever be worshipped so long as rangefinder cameras are being used.

[You should look into what digital photogrammetry challengers had done to Leica Geosystems in the last 2 decades...I was there. At one time Wild-Heerbrugg (before it was remaned Leica by the owner family in 1991) was photogrammetry.]

And don't worry about computer geeks...despite they had helped in defeating Leica Geosystems. They are not the new master of the universe in every field despite their belief. CF, SD, Memory Stick or whatever is newer is not the issue here. They merely serve our needs.

If indeed no one is interested in DRF, then let it die a natural death...and CV would have been proven right.😱
Click to expand...
 
Last edited:
highway61 - when you have tiered product line, you have to differentiate between the products just for the sake of it. hence the d40x might have an updated sensor but no support of older lenses. also folks who have a stable of older gen lenses are likely to pick something from the midrange D80/90 or the semi-pro D200/300 rather than the d40x. nikon has targeted converters from the digicams with the D40x and these folks are likely to buy newer lenses as well.

the biggest problem i see with any nikon DRF is the mount. I just don't see it happening with the M-mount. Why would nikon go away from the S mount? Heck, even the F-mount is more likely. they should just make it LTM and call it a day.

also, why does it have to be full frame anyway? i would be happy with the d300 sensor. if full frame is doable in the future, so be it. heck, i would just like to see it just to watch all these leica apologists and how full-frame dRF naysayers eat crow!
 
infrequent said:
the biggest problem i see with any nikon DRF is the mount. I just don't see it happening with the M-mount. Why would nikon go away from the S mount? Heck, even the F-mount is more likely. they should just make it LTM and call it a day.

also, why does it have to be full frame anyway? i would be happy with the d300 sensor.
Click to expand...
I fully agree about those two points.

I'd even say that about 95% of people keeping saying "full-frame, full-frame, full-frame" don't have any accurate idea of what it means actually.

What is "full-frame" in digital ? If you don't count the megapixels range into it, as for the sensor resolution matter, it doesn't make any sense.

For instance - the D700 sensor (12MP on a 24x36mm surface) has a resolution which is inferior to the D300 one (12MP on a 17x24mm surface).

Hence the D700 sensor being more gentle with older lenses than what the D300 sensor is (otherwise outstanding lenses still having a resolution intrinsically inferior to the sensor resolution is one of the main problem digital photography has to deal with).

As for my D40x example - what I'd like to have as for a Nikon DLSR on which I could fit my AI-S lenses is a small one (the D200/D300 are too big and too heavy for my taste) with the smallest amount of those pesky useless gadgets : just two metering modes (spot and center-weighted), no built-in flash, the D300 sensor, no results programs (just the A and M modes), two IQ modes (RAW and superfine/large Jpeg) only, etc.

Clearly what the marketing experts don't want the ingeneers to design any longer.
 
But we are camera geeks. 99.99 percent of the world has no interest in using 30 year old lenses on modern cameras. We are not camera makers target demographic.
 
I think a Nikon APS compact is much more likely than a Nikon digital rangefinder. The company may be idiosyncratic at times, but from a layman's POV the economics don't make sense. I'm also not sure why a 60th anniversary would merit something like this (simply because I don't associate 60 with anything special; 50, or 100, yes).
 
Ray Nalley said:
The M mount is an albatross around Leica's neck as it moves into the digital age. I can't imagine that a company producing a new, from scratch, dRF would do anything but create a completely new, all electronic mount.
Click to expand...

Nahhh. I want to use my Leica lenses interchangeably on film and digital bodies.
 
You must log in or register to reply here.
Back
Top Bottom