Sonys new 6 megapixel CCD !!!

I didn't see it because I wasn't looking for it, though I might
have read of it and passed it over as yet another cmos imager
attempt, not realizing some of the underlying technology.

At the same time, yourself, and I'd venture to guess, the
competition, weren't surprised. So I'll stick by my premise that
there is yet more to come, and sooner probably than later, from
Foveon's competitors to answer the challenge. Either way, it's
ultimately all beneficial to us.
It was a surprise to me. It's not my job to look out for these things and I didn't mean to imply that you unobservant.

I just wanted to underscore that this completely from out of the blue...

--Ron ParrFAQ: http://www.cs.duke.edu/~parr/photography/faq.htmlGallery: http://www.pbase.com/parr/
 
It's interesting to me that this imager is pretty close to half of the area of a 35mm full frame image. It's not that difficult to imagine a 4000 x 3000 pixel imager with the same cell size being produced for the pro market for use in SLR's.

Tom
I guess they are going to nickel and dime us all along the way.

But if I end up staying with this category of camera (and that's a
big IF), then I'll take a cell size of 2x my current one if the
cost is not prohibitive. This chip should not be that expensive
(the biggest argument against seeing it in a consumer camera).
Great News. However, probably 7.8 um is still small. Especially
even a good lens cannot quite utilize 10 um pixel, why don't they
make just bigger cell CCD at the expense of the number of pixels ?
--

Ulysses
 
It's interesting to me that this imager is pretty close to half of
the area of a 35mm full frame image. It's not that difficult to
imagine a 4000 x 3000 pixel imager with the same cell size being
produced for the pro market for use in SLR's.
I'm betting that the current product is intended for the Pro/SLR market given its area. Doubling the are of the chip would push the price into the stratosphere, if it could even be produced at all.

--Ron ParrFAQ: http://www.cs.duke.edu/~parr/photography/faq.htmlGallery: http://www.pbase.com/parr/
 
Yes, but we know that this is a large CCD, so the price of this CCD
won't come down much over time.

I thought we were talking about the new 6MP CCD and whether it
would be in a consumer product.
Sony has went from producing under 10 mill CCDs in '97 to over 20 mill in '00. So there is a study production increase. I think whatever cost savings derived from the earlier generations will get carried over to the new one as well.

jc
 
Sony has been heading the same direction with their CCD
development, getting smaller and smaller per photosite ... while
figuring out ways to improve the design so smaller photosites can
be as good as the larger photosites.
Technology does improve, but the gains are relatively small. There is a fundamental limit to the CCD pixel SNR that is set by the combination of temperature and photodiode size (assuming some illumination level). Modern CCD chips are not far from this fundamental limit. If you really want low noise, you need either large and expensive photodiodes, or you must cool the CCD chip. Unfortunately cooling the sensor isn't feasable with a battery powered camera.

---- Dave
 
Ya know, that's an interesting point about cooling the CCD.
How about if you had a fuel cell?

Homer
Sony has been heading the same direction with their CCD
development, getting smaller and smaller per photosite ... while
figuring out ways to improve the design so smaller photosites can
be as good as the larger photosites.
Technology does improve, but the gains are relatively small. There
is a fundamental limit to the CCD pixel SNR that is set by the
combination of temperature and photodiode size (assuming some
illumination level). Modern CCD chips are not far from this
fundamental limit. If you really want low noise, you need either
large and expensive photodiodes, or you must cool the CCD chip.
Unfortunately cooling the sensor isn't feasable with a battery
powered camera.

--
-- Dave
 
Ron,

No need to worry, and I certainly appreciate the url to the info. If anything, this makes the case for coming to this site: to learn and gain insight.

And I certainly value everyone's input.

I machine different types of widgets for various industries, and I believe that some product photography would enhance my business as well as my customers. Unfortunately, some of these widgets may be installed in less than ideal lighting condition, and my current forays into digital photography have produced someqhat less quality than I desire, especially at large (17 x 22) prints.

My immediate dilemma is that I believe that I require qualities (especially dynamic range) in a digital camera that are only slightly beyond the F707's, perhaps even in the range of the Foveon imager or a 6Mp Sony. Hence the desire for the next big thing.

The conservative alternative, for now, to a digital camera is a quasi-digital approach with conventional color negatives and a film scanner, and that's what I'm leaning towards, especially as I own a Yashica-Mat TLR. The problem is that film scanners aren't cheap, especially medium format, so I may end up purchasing something like an N80, with the knowledge that the lense(s) would be compatible with some (lower cost) digital body in the future.

While this instance above is somewhat unique, I'm pretty sure that everyone has to cross their own digital threshold. Mine gets easier with time, but I'd like to make a decision soon after seeing what is announced at the PMA.

tom
I didn't see it because I wasn't looking for it, though I might
have read of it and passed it over as yet another cmos imager
attempt, not realizing some of the underlying technology.

At the same time, yourself, and I'd venture to guess, the
competition, weren't surprised. So I'll stick by my premise that
there is yet more to come, and sooner probably than later, from
Foveon's competitors to answer the challenge. Either way, it's
ultimately all beneficial to us.
It was a surprise to me. It's not my job to look out for these
things and I didn't mean to imply that you unobservant.

I just wanted to underscore that this completely from out of the
blue...

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
The problem is that the ability to produce large chips affordably
has not been improving at a rapid rate. People look at things like
microprocessors and see how they drop in price, but what they
forget is that these drops in price are often the result of changes
to new process technologies that reduce the size of the die.
CCD's are expensiveto produce, but the X3 is Cmos so we may see some large sensors in consumer cameras.
A Pentium 4 today is significantly smaller than the first Pentium
released in 1993. It's more affordable mostly because it's
smaller, not because the cost per square millimeter of working chip
has gone down that much.
It's more affordable because of competition with AMD, secifically the Athlon. Did you know the P4 was/is much larger than the PIII (on the 18u process -- I'm not sure if Intel is currently down to 13u, but I believe it needs to be to be equivelent to the size of an 18u PIII)?
 
I'm not that certain, of course, but I have strong doubts about the F909 (follow on to F707) using this particular chip. Both the 505v and 707 used an aspect ratio of 3/4 while this chip is 2/3. It's also a far larger chip in physical dimensions so they'd probably have to step up to a considerably larger body/lens.

The specs on the chip make me think digital SLR all the way. The large site size is great but the chip figures to be priced way above the F707's chip for that reason. For a digital SLR the large physical size is a bonus, because it matches the existing lenses better and, of course, their price point covers the higher cost. Has there been any serious talk of a SONY SLR?

The new CCD dimensions 25.2X17.64 = 444.5 sqmm, F707 CCD dimensions 9.74X7.96 = 77.5 sqmm. Thats an area increase of 5.7X. The production costs approximately scale with area so this sensor figures to cost 5 times what the the F707 costs.

I'd expect the F909 to have a target price of under 1200USD and I'd expect them to follow the F707 formula as closely as possible. 3/4 aspect ratio with the same site size and 7.5MP. It could be that the F707 body and lens were diesigned with that much expansion possible but probably they'll have to increase the size a bit. Shrinking the site size is a possibility, too, or maybe keeping the active area the same and decreasing the wasted space in between... heck all those engineers need a challenge
Wow! Sonys new APS size CCD ICX413AQ !!!

Optical Format 1.8 inch, 28,4mm (APS size)
Aspect ratio: 3:2
Colour filters: RGB primary
Active pixels 3000x2000 (total 6 Million square pixels)
Chip size: 25.2x17.64
Unit cell size: 7.8 x 7.8 um
Test sample TS: End May 2001
Evaluation sample ES: End December 2001
Mass Production MP: End June 2002

(from my Kumamoto 2nd Techno-Park source! :)

Niklas
 
The new CCD dimensions 25.2X17.64 = 444.5 sqmm, F707 CCD dimensions
9.74X7.96 = 77.5 sqmm. Thats an area increase of 5.7X. The
production costs approximately scale with area so this sensor
figures to cost 5 times what the the F707 costs.
I meant to say the new sensor figures to cost 5 times what the F707's sensor costs.
 
The problem is that the ability to produce large chips affordably
has not been improving at a rapid rate. People look at things like
microprocessors and see how they drop in price, but what they
forget is that these drops in price are often the result of changes
to new process technologies that reduce the size of the die.
CCD's are expensiveto produce, but the X3 is Cmos so we may see
some large sensors in consumer cameras.
Large CMOS chips are expensive to produce too. The 3.5MP Foveon chip in on par with the Largest Pentiums ever produced.
A Pentium 4 today is significantly smaller than the first Pentium
released in 1993. It's more affordable mostly because it's
smaller, not because the cost per square millimeter of working chip
has gone down that much.
It's more affordable because of competition with AMD, secifically
Competition cuts into their profits and pushes them to make chips less expensively. This has been achieved mostly by making the chips smaller. Progress in reducing defect density on wafer, which would enable them to make larger chips more economically, has been very slow.
the Athlon. Did you know the P4 was/is much larger than the PIII
Every new generation is physically larger than the previous on the same process technology. Each new generation starts out very expensive too. The price then comes down from competition, economies of scale, the learning curve at the fab, and also because they move to new process technologies that make the chips smaller.
(on the 18u process -- I'm not sure if Intel is currently down to
13u, but I believe it needs to be to be equivelent to the size of
an 18u PIII)?
Intel is down to 0.13micron for the newest P4, which makes it still a big larger than the 0.18micron P3. 0.18micron P3: 106mm^2, 0.13micron P4: 146mm^2. (Intel doubled the cache size to 512KB when they moved from 0.18 to 0.13 micron with the P4, which added an extra 13 million transistors for a 30% increase If they had kept the cache size the same, the 0.13micron P4 would be about the same size as a 0.18micron P3.)

All of these, however, are smaller than the first Pentium/60 which was 0.8micron and a whopping 295mm^2.

--Ron ParrFAQ: http://www.cs.duke.edu/~parr/photography/faq.htmlGallery: http://www.pbase.com/parr/
 
Yes.

Did anyone notice in the specs sheet that it said "mechanical shutter"?

Hmmm...
It's interesting to me that this imager is pretty close to half of
the area of a 35mm full frame image. It's not that difficult to
imagine a 4000 x 3000 pixel imager with the same cell size being
produced for the pro market for use in SLR's.
I'm betting that the current product is intended for the Pro/SLR
market given its area. Doubling the are of the chip would push the
price into the stratosphere, if it could even be produced at all.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
-- Ulysses
 
Yes.

Did anyone notice in the specs sheet that it said "mechanical
shutter"?
Interesting.

The Data sheets have been strangely disappearing from Sony's web page. I still have the 6MP one, but the 5MP one just disappeared.

My understanding is that all of CCDs need a mechanical shutter if you want to reduce blooming. That's why even consumer digital cameras have them.

I'm a little confused about something in the 6MP sensor specs. It assumes a mechanical shutter says that it's designed on the assumption of the frame readout system. This suggests that it's a full frame device intended for SLRs.

However, on the second page, they indicate that it has interline transfer.

--Ron ParrFAQ: http://www.cs.duke.edu/~parr/photography/faq.htmlGallery: http://www.pbase.com/parr/
 
Olympus and Kodak are working on a new Digital SLR that will be announced at PMA 2002. It will use the new Kodak 4/3" 5.1mp, 6.8-µm KAF-5100 CCD. The price will be less than $2000 ..
 
Funny, I didn't notice it, though I just assumed that a minimum of re-engeering would go into the SD9. I guess that a mechanical (electro-mechanical?) shutter is the easy way to go, especially for an SLR. Am I missing something?

About the chip area: I just found it a teensy bit convenient that it almost perfectly matched half the area of 35mm. So doubling that (the equivalent of 2 cores side by side) isn't really out of the question. That's around 900 mm^2 which, is indeed a lot of real estate for a chip, but not out of the question down the road. I just figured it would ease the engineering task of converting an SLR to digital.

Tom
Did anyone notice in the specs sheet that it said "mechanical
shutter"?

Hmmm...
It's interesting to me that this imager is pretty close to half of
the area of a 35mm full frame image. It's not that difficult to
imagine a 4000 x 3000 pixel imager with the same cell size being
produced for the pro market for use in SLR's.
I'm betting that the current product is intended for the Pro/SLR
market given its area. Doubling the are of the chip would push the
price into the stratosphere, if it could even be produced at all.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
--

Ulysses
 
The Data sheets have been strangely disappearing from Sony's web
page. I still have the 6MP one, but the 5MP one just disappeared.
It's still there. They've just moved things around a bit:
http://www.sony.co.jp/~semicon/english/img/sony01/a6802960.pdf

Notice that this one mentions the use of a mechanical shutter in relation to how smearing would be reduced, frame readout, etc.

The 6Mp initial spec sheet is not a full spec sheet, so maybe I was making assumptions based on incomplete info. Still, I'd like to learn why this was such a standout feature in the initial specs.
My understanding is that all of CCDs need a mechanical shutter if
you want to reduce blooming. That's why even consumer digital
cameras have them.
Yes, that would seem to be correct.
I'm a little confused about something in the 6MP sensor specs. It
assumes a mechanical shutter says that it's designed on the
assumption of the frame readout system. This suggests that it's a
full frame device intended for SLRs.
Your questions would seem to mirror my own as I look more closely at this.
However, on the second page, they indicate that it has interline transfer.
What are you thinking that this implies?-- Ulysses
 
Funny, I didn't notice it, though I just assumed that a minimum of
re-engeering would go into the SD9. I guess that a mechanical
(electro-mechanical?) shutter is the easy way to go, especially for
an SLR. Am I missing something?
The circuitry for an electronic shutter takes up extra space on the CCD. If you don't want an electronic shutter and don't need the ability to pull video off the chip, then you can skip this circuity and use the space for larger receptive areas. All digital SLRs do this.
About the chip area: I just found it a teensy bit convenient that
it almost perfectly matched half the area of 35mm. So doubling that
(the equivalent of 2 cores side by side) isn't really out of the
question. That's around 900 mm^2 which, is indeed a lot of real
estate for a chip, but not out of the question down the road. I
just figured it would ease the engineering task of converting an
SLR to digital.
900mm^2 will probably be beyond the realm of what can manufactured cost effectively for anything other than special purpose applications for quite some time. I think they used the aspect ratio to match the aspect ratio of film (as done in other digital SLRs). I doubt it has anything to do with plans to make a 35mm size CCD in the near future.

--Ron ParrFAQ: http://www.cs.duke.edu/~parr/photography/faq.htmlGallery: http://www.pbase.com/parr/
 

Keyboard shortcuts

Back
Top