(aborted?) Sony 10.66 MP APS (D100, etc.) sized sensor

Anastigmat · Jul 22, 2004

Thom Hogan said:
Sorry, but your own argument goes against you, here. The 4/3 system
is a 2x factor, APS is a 1.5x factor. That's enough difference so
that if your argument that you can't get enough resolution out of
APS versus full frame is true, then it also applies that you can't
get enough resolution out of 4/3 versus APS.

Neither of them has as high resolution as 24mm x 36mm full frame sensors. So, if Nikon stays with the APS sized sensor, it would not be able to compete with Canon in the future, but it would be more than competitive against the E-1, since the D70 has lower noise at high ISO settings than the E-1. That is because the APS sized sensor is physically larger.

Thom Hogan said:
You don't support your notion of "better" by any details. Having
spent a lot of time handling DSLRs and doing competitive analysis
of them for one vendor, I can't agree. It's not a BAD camera, at
all. But there's little about that is unique, and it's UI is a
little haphazard.

Color accuracy, for example, is excellent for the Pentax. The reviewer on this site call its color accuracy and balance "film like". The D100 is notorious for inaccurate color balance. The *ist D is free of moire problems. The D70 is terrible in that regard. The *ist D has no back focus problem. The D70 does. The *ist D has a big and bright viewfinder. The D70 has one of the smallest viewfinders around, and it has a pentamirror. The *ist D has a pentaprism.

Thom Hogan said:
I'll repeat, for Pentax to succeed, they have to bet the
company. They have to make a better DSLR than the D70/DR at a lower
price that Nikon/Canon can't easily reach, they have to support
that with marketing that'll move it out the stores,

That is not a problem. They can easily make a better camera than either the D70 or the digital rebel. They are doing it now. As for pricing, they may need to get a price break from Sony. Sony may not give them the price break unless Pentax guarantees that it will buy a certain quantity. Pentax does not have to bet the company. As for marketing, that is a little tougher, since Pentax is not a household name to the average consumer.

Thom Hogan said:
they have to
have manufacturing ready to make more than 1m units a year, and
they have to be ready to follow that up with something that keeps
any advantage they gain in the market. I just don't see any of that
happening. First, they've proven they can make a decent DSLR that's
relatively equivalent to what's available, but they haven't shown
anything that indicates they can do better than Nikon and Canon.

As I said, they have demonstrated that they can eliminate the moire problem that plagues the D70 and the color inaccuracy of the D100. Pentax *ist D cameras are free of the AF and/or back focus problems that plague both the D70 and the Canon EOS 10D. Pentax has demonstrated its ability to build a better camera.

Thom Hogan said:
Second, I don't think they can get their price under what Canon
will do (I'll remind you that Canon has a stated goal of being #1
in every market they compete in--so that US$499 DR Jr isn't going
to be full frame!).

I know Canon does have the advantage on price because it makes its own chips. However, Pentax can make a better camera than Canon. If it costs more than the Canon, but if the consumers know it is better, than they will gladly pay the higher price. What I am really concerned is that Canon will come out with a 1.3x format 8mp 10D and sells it for $1K or slightly more. Not only will the Pentax be toast, but so will the Nikon D70. If that happens, Sony will lose its lucrative chip market. Sony will need to start making bigger sensors to supply Nikon and Pentax and/or try to sell its APS sensors to makers of point and shoot digital cameras.

If Sony won't make a larger sensor, then perhaps Pentax will turn to another source for sensors. In the mean time Pentax will not starve since it is making compact digital cameras and film cameras too, not to mention medium format cameras, lenses and binoculars.

Anastigmat · Jul 22, 2004

Thom Hogan said:
All these are decent cameras. Other than the size of the ZX-5,
though, none are particularly distinguished from their competitors,
IMHO. Where is the IS/VR? Where are the metering/focus/flash
advances? Where are the digital advances?

IS and VR are definitely useful features but these features are expensive and thus available only to the pro or well heeled consumer. To the average consumer, it is not important, and the lack of VR does not make a camera such as the ZX-5 less competitive. It would be nice for the MZ-S to have it and Pentax is apparently working on this.

As for flash metering, Pentax is better than Canon and Nikon. Both cannot get off the film plane TTL flash to work in a DSLR. Pentax can. So its owners can choose from a wide variety of Pentax and off brand flashes ranging from ringflashes, handle mount flashes, bare bulb flashes and regular shoe mount flashes. The D70 forces you to buy one of only two flashes available and then its bright preflash causes some subjects to blink, resulting in unflattering portraits with the eyes of subjects being closed. The Canon system give the users more options, but it is still not off the film plane TTL. As for metering, the *ist D has spot metering, the Canon 10D does not. In the following review:

http://www.luminous-landscape.com/reviews/cameras/pentax-istd.shtml

the *ist D is praised for its "exceptional" matrix metering and "As good as I've ever seen."

As for focusing, the same reviewer says, "Similarly, the multipoint auto-focus proved to be very quick and accurate."

The D70 has back focus problems. The 10D has either an Af or a backfocus problem depending on who you think is correct in their diagnosis. The Pentax *ist D is free from focus problems. As for "digital advances" the reviewer at dpreview.com praises the *ist D images exhibit "Very natural color response, very film like" That is indeed a digital advance because the D100 uses the same sensor and it is notorious for its inaccurate color balance.

Anastigmat said:
One of the things that changed in the move from 35mm SLR to digital
SLR is that the camera is no longer a light-tight box that holds
film. In the days of film, the difference in ultimate image quality
produced by an N55 and an F5 was nearly zero (okay, the F5 holds
the film a little flatter, which helps). The primary advantages to
an F5 were that it was more robust, offered more user control, and
could run that film through the camera faster. But, with care and
the same lens, image quality wasn't an issue. In the digital SLR
world, that's all changed. There can be a difference between the
low end and high end camera in a lineup, and it boils down to all
the digital stuff. I've seen no evidence that Pentax knows how to
do digital better than Fujifilm or Kodak, let alone Nikon and Canon.

I think you finally got what I have been saying. In the digital age, the sensor is the determining factor. The full frame 35mm sensor, because of its inherently lower noise and higher resolution, will always have the advantage over the smaller sensors. That is why Canon can deal Nikon (and Pentax) a serious blow by marketing a 1.3x 8mp 10D replacement for $1-1.2K. At higher prices (such as $2K), consumers would forsake such a camera and be content with less quality. But at such low prices, the Canon wins hands down. Pentax may then simply go up market and build a full frame digital and forget about trying to compete for market share at the low end. Pentax is contemplating a digital 645. Pentax has shown that it can make a camera that produces a better image than Nikon, even when both use the same chip. Pentax is a very capable camera maker and its corroboration with HP in building digital cameras is bearing fruit. Pentax has gained a lot of knowledge in image processing from HP.

Anastigmat said:
Sorry, but you're missing the market expectation here. The volume
right now is at the US$899-999 price point. Every survey I see says
the volume will be HIGHER at the US$599-699 price point, and I'm
pretty sure we're headed towards a US$499 DSLR in the not so
distant future.

Your survey does not take into account the possibility that a 1.3x format 8 mp camera will be sold at $1-1.2K. If such a possibility materializes then your survey goes out the window. Of course, it may not meterialize this fall, but it is definitely going to happen sooner or later. If the digital rebel is selling for $600, then why would consumers pay $1k for a 10D with the same sensor? If the 10D mkII has a bigger sensor, then the consumers would be more than willing to pay the higher price.

Anastigmat said:
I think you oversell their benefits. Does Pentax make decent
cameras? Sure. Do they make offerings that can compete at the high
end (the point I was trying to make)? I see little evidence of that.

Of course they cannot compete if they cannot get a full frame sensor. If they can get a reliable supply of a good sensor, they will be competitive. Nikon cannot compete at the high end either since it does not have a camera that compares to the 1DS.

Ron Parr · Jul 22, 2004

Joseph S. Wisniewski said:
Only if you placed one over that area. I'm going to try to
"draw"this with words, so bear with me.

Picture a CCD with a 50% fill factor, the cells are 71% of the
pixel pitch, the gutter between is 29%.

Now, place the microlenses at half the pixel pitch, and center them
so each focuses on a relatively small area in one quadrant of the
cell. Those bright points are 50% of the pixel pitch apart, and the
cell itself is 71% of the pixel pitch, so the focal points of four
lenses really will hit one cell. So you have four brightly
illuminated areas on the cell, and a cross shaped center "dark"
area. This doesn't matter, charge liberated in the bright areas
will fill the entire well.

I don't think I'm visualizing what you're describing.

I can visualize positioning the microlenses so that each lens overlaps a bit of photosensitive area on one side and a bit of shielded area on the other side, focusing the light on the photosensitive area. I could see that as a reasonable way to go although I'm a little dubious about how much real benefit there would be since the lens would need to be more than 50% over photosensitive area.

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

Anastigmat · Jul 22, 2004

Ron Parr said:
Anastigmat said:

Member said:

Nobody is arguing this position, so you can stop going after this
straw man.

Click to expand...

If sensor size is fixed, then the pixel would have to keep
shrinking every time there is an increase in resolution.
Unfortunately there are some who expect miracles from engineers but
engineers have to obey the laws of physics. To improve digital
camera image quality, engineers need to utilize every square
millimeter of real estate alloted to the sensor.

Click to expand...

As I've said many times, nobody is arguing that this can continue
ad infinitum. However, it is true there are many ways to improve a
sensor and that making it larger is just one of them.

Of course, they would have to back off from such an absurdity when pressed. But they keep insisting that the small sensor can yield high quality images given technological advances, as though such advances can overcome the inherent advantages of larger chips. All else being equal, the larger chip will always have the advantage, so it is irrational to keep trying to improve a small chip when a bigger chip can be fitted without increasing camera size.

Member said:
Your arugment has not, so far, been based upon insights into the
physics behind any fundamental limits. Your argument has been to
present the smallest known pixel.

The size limit on pixels is based on physics.

Member said:
Member said:

Not at all. I pointed out that a full frame sensor can better
handle 10-11 megapixels than an APS sized sensor.

Click to expand...

No. What you said was, "such a chip would not be the next
generation of DSLR sensor, since such a chip would almost certainly
be noisier, given its small size, than the current chip."

That is precisely because a full frame sensor can handle such high resolutions better than the APS sized sensor. If so, then they will be the next generation of photosensors, not the APS sized ones.

Member said:
You are clearly indicating that you think such a chip would be
noisier despite the improvements in chip design in the past few
years.

Yes it will, because whatever improvements made can be incoporated into the larger chip to make it even lower in noise and/or higher in resolution.

Member said:
Member said:

The question then becomes: how good is image quality for these
sensors with small pixels.

Click to expand...

Time will tell. While the marketing has occasionally gotten ahead
of the engineering in the megapixel race, manufacturers haven't so
far been totally stupid about it. When the 4MP 1/1.8" sensors came
out everybody thought they'd be worse than the 3MP 1/1.8" sensors.
The first batch actually were, but the current ones seem quite good.

Today's 1/2.7" 3MP sensors seem about as good as the 1/1.8" 3 MP
sensors of a few years ago.

Today's approx. 2.7 micron pixels are a bit noisy (5 MP 1/1.8", 8
MP 2/3") but they may improve.

Hope is not a good design strategy. So, why wait for technology to improve so you can use the small sensors, when you can use the big sensor and get the image quality that you need RIGHT NOW?

Member said:
Member said:

Member said:

Again, nobody is arguing for infinitely small pixels so your
comments are addressing a straw man that has nothing to do with the
discsussion.

Click to expand...

Actually some people are indeed arguing that. So it is not a
strawman.

Click to expand...

Who? Seriously, who do you think you're debating with here?

I am debating against those who worship the small chip and claim that technology will keep improving the small chip so that there will never be the need for full frame sensors. I point out that small chips will always be lower in quality because there is a limit on how small a photodiode can shrink. Besides, even before that limit is reached, the image quality may not be adequate. Hence a larger sensor is always cheaper.

Member said:
Member said:

Am I not correct that a sensor with smaller pixels would be noisier?

Click to expand...

As a categorical statement this is not correct. As a ceteris
paribus statement it might be, but you should really qualify it as
such.

Why would a small pixel not be noisier if all else is equal?

Member said:
I the current context, the discussion is about a move from an old
sensor with large pixels to a new sensor with smaller ones that are
still pretty large, so your statement is neither correct nor
incorrect without further information. It's pure speculation.
That's the point.

Does the new sensor exist? How do you know that the new technology made the noise level the same while resolution has been increased?

Member said:
Member said:

Actually I said that sooner or later engineers are going to hit the
ceiling on resolution given a fixed sensor size because there is a
limit on how small a pixel can get. Some people of course disagree
with my claim. By disagreeing, they are arguing that pixels can
keep shrinking forever. If so, then there is no straw man. If
they are not arguing that, then I am correct that there is indeed a
limit to how small a photodiode can get. So, am I right about the
claim that there is a lower limit for pixel size or not?

Click to expand...

Nobody here believes that pixels can get arbitrarily small.
Really. That's not what the discussion is or has been about

Then that is simple, sooner or later it becomes impractical to hold on to the small sensor, and image quality can only be improved by enlarging the chip. We all agree then that the large sensor will always have the advantage over the small sensor regardless of how much technology has advanced. By so stipulating, there is no need for further arguments to the contrary.

Lucem · Jul 22, 2004

For the new 8MP 2/3" sensors Sony is employing double internal micro lenses, AKA (DIL)
I was wondering if that is what we are seeing in these pictures.
Only that this particular one seem to be using four instead.
I understood DIL as microlenses been stacked on top of other.

Anastigmat · Jul 22, 2004

Ron Parr said:
Anastigmat said:

It isn't being done because yield is indeed lower for larger chips,
but not as low as their price differential would suggest.

Click to expand...

It's hard to say since we don't know the actual prices. It's
probably true that Canon is skimming the market a bit because they
have such an edge on their competitors.

Sensors are probably more forgiving of defects than logic chips,
but if we can learn anything from logic chips it's that price goes
up at a really killer rate as chip size increases.

Play around with the following applet:

http://tech-www.informatik.uni-hamburg.de/applets/yield/

Of course, the reuslts will change based upon the random placement
of defects. With the random seed that I started with, a 10D sized
sensor has a yield of 18 good chips per wafer while a full sized
sensor (1Ds) has a yield of 1. This would imply a price
differential of 18X.

Let's assume that your calculation is correct. Let us also assume that a single wafer costs $1,000. That means the full frame sensor costs $1,000 each in material to make plus whatever labor needed to make it. The 1.6x format sensor would then cost $1,000/18 or $55 in material cost and also 18 the labor cost of the full frame chip. Since labor cost of the small chip is so low, let's ignore it for the moment. If the 1.6x sensor only costs $55 each, then it cannot possibly account even for the projected $600 cost of the Digital Rebel replacement. Let's suppose it costs Canon $100 to produce each Digital Rebel body, it can make $500-$55-dealer markup for every Digital Rebel it sells. That means it will make probably between $300-$450 per camera. Using the same formula, Canon can make the same amount of money per camera if it uses the same body and a full fram chip by pricing it at $2,000. The $400 extra should take care of increased production costs of the full frame camera and additional dealer markup. Compared to how much the EOS-1DS is selling for, one can see that there is a whole lot of room for price cutting if Canon wants to make the full frame digital camera the new standard. A full frame digital camera can be sold for as little as $2000 today!

So, why would a 1.3x format 8mp digital camera selling for $1k be impossible, as some posters have argued.

Terry Sessford · Jul 22, 2004

Ron Parr said:
Terry Sessford said:

I have no doubt that with advances in technology APS & 4/3rd sized
sensors will be capable of delivering more than enough dynamic
range and resolution for the vast majority of users

Click to expand...

The question is what "enough" means. You'll always have more
flexibility for digital zoom (cropping) and exposure compensation
with larger sensors, so it's a really a question of how much people
are willing to pay for this.

Enough to make no visible difference on, say, an A3 sized print, and how many users print larger than this?

Member said:
Sure - you need a larger and heavier lens when you have a larger
sensor if you want to achieve the same field of view, but if you
have larger, higher resolution sensor, you'll be able to crop and
avoid putting the really long lens in your bag in the first place.

But if you are going to crop and throw away pixels then why not simply have a smaller sensor in the first place with all its size, weight and cost advantages?

I just do not buy these contrived arguments for a full frame sensor.

Terry.

Anastigmat · Jul 22, 2004

Ron Parr said:
Anastigmat said:

I am afraid, however, that you are sacrificing ISO speed without
getting a smaller package. Digital camera bodies haven't shrunk to
fit the small ensor. The 10D for example, is bigger than any film
camera I have ever owned, including the Nikon cameras I have owned.

Click to expand...

Unsound reasoning here - extrapolating from a single case with no
facts to back up the generalization.

The size of the 10D and its sensor chip is evidence that a small chip does not result in a small digital camera. The Digital Rebel shrunk even though it has the same size chip as the 10D. So there is no evidence that a full frame chip would make a camera larger but there is evidence that a smaller camera is possible without reducing chip size. The Pentax *ist D has a bigger chip than the Digital Rebel, but it is a smaller camera. So, these are all facts which argue against a smaller chip resulting in a smaller camera.

Member said:
The computing part of the camera will undoutedly get smaller and
require less power, as computing devices have traditionally done
for the past 30+ years. Surely you don't think the sensor is
reponsible for the extra bulk?

Of course not. That is why I said that a large sensor will not increase camera size. It is borne out by facts.

Member said:
I suggest you look at the *ist review. I think you'll find it as
small or smaller than any Nikon film camera you've owned or could
currently buy.

Yes I know. That is why I disagree with your contention that the small chip is going to keep camera size down.

Member said:
Member said:

But why? Why shrink the pixel to get more resolution when you can
grow the sensor without increasing camera size or having to
redesign the lens mount to get a larger image circle. 35mm lenses
are already capapble of covering the full frame sensor, so why keep
shrinking the individual pixel if it is not even necessary from the
point of view of the photographer? I know the chip maker can make
more money by making small sensors, but why let the bean counters
at chip making companies dictate to us photographers how big of a
sensor we are allowed to have?

Click to expand...

It's a question of what you're willing to pay for. The technology
to mass produce large chips (greater than about 20mm on end) is
pretty exotic, even by semiconductor standards, so you're asking
consumers to help amortize the cost of the basic technology and
then pay for the generally low yields of large chips on top of that.

I don't expect prices for full format sensors to drop overnight. I am more than happy that professionals are paying for the cost of R&D for these chips right now when they pay $7-8K per camera. When the R&D has been paid for, then the price should drop. When prices do drop, the average consumer would definitely want a full frame camera.

Member said:
Eventually, we may see inexpensive big chips if demand is high
enough and manufacturers push on it hard enough (for all we know,
Canon could have been pouring money into this for years), but the
notion that manufacturers could EASILY make larger sensors at a
reasonable price point is naive, IMO.

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

I think we agree more than we disagree. I agree that big chips will drop in price. We disagree on whether there is demand for them. Consumers certainly demand big chips. The camera makers do not need to push these chips. All they need is to reduce prices, and the full frame sensors will sell.

A full frame sensor with 11 megapixels will have roughly the same amount of parts as an 11 megapixel APS sized sensor. So the chance that a random defect will occur in either of them would be about the same. So, they should have identical yields. The only difference in manufacturing cost is that far fewer full frame sensors would fit on a single wafer versus the number of APS sized sensors that will fit. I think the ratio is something like 36 APS sensors versus 24 full frame sensors. If so, then the difference in manufacturing cost would be a factor of 36/24 or 1.5. A full frame sensor would only cost 1.5 times as much to make as an APS sensor. Even if we double or even quadruple the cost factor of the large sensor, they would still only cost at the most $400-600 dollars more than an APS sensor. If full frame digital cameras are available at $400-$600 more than APS sensors, the APS sensor cameras will be obsolete overnight. I have no doubt that the consumer would opt for the full frame digital if they only have to pay a few hundred dollars more.

bmvmadeira · Jul 22, 2004

you can't make 100% perfect silicon waffer (not contaminated) so a bigger chip will allways be more sensitive to random defects besides the smaller number per waffer difference
I would still prefer a FF sensor. But I don't know if will happen.

FF will allways be better but Dx sensors can win based on a price/good enough combination.

Anastigmat said:
Ron Parr said:

Anastigmat said:

I am afraid, however, that you are sacrificing ISO speed without
getting a smaller package. Digital camera bodies haven't shrunk to
fit the small ensor. The 10D for example, is bigger than any film
camera I have ever owned, including the Nikon cameras I have owned.

Click to expand...

Unsound reasoning here - extrapolating from a single case with no
facts to back up the generalization.

Click to expand...

The size of the 10D and its sensor chip is evidence that a small
chip does not result in a small digital camera. The Digital Rebel
shrunk even though it has the same size chip as the 10D. So there
is no evidence that a full frame chip would make a camera larger
but there is evidence that a smaller camera is possible without
reducing chip size. The Pentax *ist D has a bigger chip than the
Digital Rebel, but it is a smaller camera. So, these are all facts
which argue against a smaller chip resulting in a smaller camera.

Anastigmat said:

The computing part of the camera will undoutedly get smaller and
require less power, as computing devices have traditionally done
for the past 30+ years. Surely you don't think the sensor is
reponsible for the extra bulk?

Click to expand...

Of course not. That is why I said that a large sensor will not
increase camera size. It is borne out by facts.

Anastigmat said:

I suggest you look at the *ist review. I think you'll find it as
small or smaller than any Nikon film camera you've owned or could
currently buy.

Click to expand...

Yes I know. That is why I disagree with your contention that the
small chip is going to keep camera size down.

Anastigmat said:

Anastigmat said:

But why? Why shrink the pixel to get more resolution when you can
grow the sensor without increasing camera size or having to
redesign the lens mount to get a larger image circle. 35mm lenses
are already capapble of covering the full frame sensor, so why keep
shrinking the individual pixel if it is not even necessary from the
point of view of the photographer? I know the chip maker can make
more money by making small sensors, but why let the bean counters
at chip making companies dictate to us photographers how big of a
sensor we are allowed to have?

Click to expand...

It's a question of what you're willing to pay for. The technology
to mass produce large chips (greater than about 20mm on end) is
pretty exotic, even by semiconductor standards, so you're asking
consumers to help amortize the cost of the basic technology and
then pay for the generally low yields of large chips on top of that.

Click to expand...

I don't expect prices for full format sensors to drop overnight. I
am more than happy that professionals are paying for the cost of
R&D for these chips right now when they pay $7-8K per camera. When
the R&D has been paid for, then the price should drop. When prices
do drop, the average consumer would definitely want a full frame
camera.

Anastigmat said:

Eventually, we may see inexpensive big chips if demand is high
enough and manufacturers push on it hard enough (for all we know,
Canon could have been pouring money into this for years), but the
notion that manufacturers could EASILY make larger sensors at a
reasonable price point is naive, IMO.

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

Click to expand...

I think we agree more than we disagree. I agree that big chips
will drop in price. We disagree on whether there is demand for
them. Consumers certainly demand big chips. The camera makers do
not need to push these chips. All they need is to reduce prices,
and the full frame sensors will sell.

A full frame sensor with 11 megapixels will have roughly the same
amount of parts as an 11 megapixel APS sized sensor. So the chance
that a random defect will occur in either of them would be about
the same. So, they should have identical yields. The only
difference in manufacturing cost is that far fewer full frame
sensors would fit on a single wafer versus the number of APS sized
sensors that will fit. I think the ratio is something like 36 APS
sensors versus 24 full frame sensors. If so, then the difference
in manufacturing cost would be a factor of 36/24 or 1.5. A full
frame sensor would only cost 1.5 times as much to make as an APS
sensor. Even if we double or even quadruple the cost factor of the
large sensor, they would still only cost at the most $400-600
dollars more than an APS sensor. If full frame digital cameras are
available at $400-$600 more than APS sensors, the APS sensor
cameras will be obsolete overnight. I have no doubt that the
consumer would opt for the full frame digital if they only have to
pay a few hundred dollars more.

Peter Herth · Jul 22, 2004

Ron Parr said:
The question is what "enough" means. You'll always have more
flexibility for digital zoom (cropping) and exposure compensation
with larger sensors, so it's a really a question of how much people
are willing to pay for this.

"enough" means, that the mayority of the people are not willing to pay significantly more for quality improvements. Furthermore, while there are individual differences in what is considered a good price, for every technology, there is a point where increases in the price result in only small performance gains.
Peter

krusadr · Jul 22, 2004

Terry Sessford said:
Joseph S. Wisniewski said:

Anastigmat said:

DMillier said:

The oly E1 may point the way: whilst not a small camera it does aim
to be a pro camera and it is a tad smaller that a D1x...

Click to expand...

I hope the E-1 does not point the way for Nikon or Pentax.

Click to expand...

Click to expand...

I disagree (with Anastigmat), I hope that the E-1 does point the
way for all manufacturers, including Nikon and Pentax and, God
forbid, even Canon, but sadly I doubt that the latter will be the
case.

The vast majority of film SLR users switching to digital do not
need full frame sensors - the only benefit for them is total
backwards compatibilty with their existing legacy lenses.

How do you define need? Do you "need" a 3.2 Ghz processor or will a PII 933 do all you need? It's a bit of an arrogant statement. Even if they don't want to blow up their stock prints to poster size now, they might want to in a few years when such printers are affordable. Most fast processors are used by gamers and are as such needed in the persuit of pleasure. Photographers amateur and pro want the very best they can get, else why are we here having these discussions?

Member said:
Personally I view full frame sensors as expensive dinosaurs - blind
adherence to an old film standard that is no longer relevent in the
world of digital image capture.

Good point. That ranks with other insightful arguments of recent years

"A Pentium (as in the buggy Pentium I 66mhz) would only ever be used in high end graphic workstations and servers"

"A 10mb (Winchester) hard drive would be IMPOSSIBLE to fill."

and most famously of all..

"640k of ram is all you'll ever need"

Every time this type of argument about full frame comes up I gasp with amazement. Of course there is a place for small sensors and "prosumer" type formats like the E-1. Theres also room for tiny sensors in digicams.

However the 35mm format gained popularity with pro's due to it's convenient transportability despite its lower quality than MF. I would say that most pro's would love to get medium format quality from their 35mm size kit. To talk about dinosaurs is ridiculous.

Smaller sensors are perfectly fine IF the SLR system that houses them is signifigantly smaller to suit but at the moment this is not the case IMO.

If you only need an APS (ish) sized sensor and you want to invest in a line of lenses with a small image circle that is restricted to such a sensor then go ahead and buy such a camera. And if you think the size and weight savings are worth it then go ahead.

But please don't talk about a 35mm sized camera system with medium format quality being a dinosaur. The holy grail in camera equipment is surely the best quality (everything) for the smallest size/weight/cost and right now Canon is on that quest. I just hope Nikon decide to try and catch up. MF quility images in a 35mm sized system is the future like it or not and it's an exciting prosect for me and many other photographers I would imagine.

Member said:
Member said:

But Pentax is a wild card. Remember this?

http://www.cameraquest.com/jpg5/p11025a.jpg

110 format SLR, with the entire camera scaled down to fit. I can
seriously see them trying something like this in the "4/3 system".

Click to expand...

This is an excellemt point made in the past by myself and others. If such a sytems were introduced today with a nice 5 Mp sensor I'd probably buy one. Fun fun fun.

http://forums.dpreview.com/forums/read.asp?forum=1000&message=9363102

Member said:
Yep, I remember this well - I used to own a Pentax 110 SLR plus
lenses about twenty or so years ago. It was a lovely little system
that was perfect for hiking etc, and it produced surprisingly good
results. But it is apparent that, size-for-size, digital can
surpass film in terms of quality and interestingly the size of the
E-1 4/3rd sensor is within a mm ot two of the size of the old 110
film format. In the past I have stated more than once in these
forums that I never could understand why Pentax did not resurrect
their old 110 film SLR and stick a 2/3 or 4/3 size CCD in it!

Always great to hear your opinions and thoughts Joe, they are
nuggets of gold in these forums.

Terry.

krusadr · Jul 22, 2004

Terry Sessford said:
Joseph S. Wisniewski said:

Anastigmat said:

DMillier said:

The oly E1 may point the way: whilst not a small camera it does aim
to be a pro camera and it is a tad smaller that a D1x...

Click to expand...

I hope the E-1 does not point the way for Nikon or Pentax.

Click to expand...

Click to expand...

I disagree (with Anastigmat), I hope that the E-1 does point the
way for all manufacturers, including Nikon and Pentax and, God
forbid, even Canon, but sadly I doubt that the latter will be the
case.

Well I agree with Anastigmat. We need inovation at the top end. Just because 'you' don't need or understand other's need for progress doesn't mean there shouldn't be any.

If you like small sensors (29% of the 35mm format) and you think an E-1 is significantly smaller than a full-image-circle system then go and buy one. I think it's way too big a compromise for the barely noticible size advantage. It's also not cheap. I actually suspect Oly could have made it much smaller but it wouldn't look as good for marketing reasons.

Then you can go take pictures and as your lens purchases will lock you into the 4/3 sensor you won't feel the need to comment in threads dealing with the exciting advances and price drops of 1.3x and FF sensors. At least by then the E1 will be priced more competitively with it's other consumer format brethren.

I find it highly amusing the way the E-1ers pop up every time sensor size is discussed. I imagine that in 12 or 18 months when these threads are discussing the latest large sensors and their wonderful high ISO performance, superb dynamic range, enlargement capability etc etc those E1-ers will be strangely quiet, locked away in their E1 forum enclave. The forums will be all the worse for it. I miss them already.

Roll on the 1.3x 10DII.

Member said:
The vast majority of film SLR users switching to digital do not
need full frame sensors - the only benefit for them is total
backwards compatibilty with their existing legacy lenses.
Personally I view full frame sensors as expensive dinosaurs - blind
adherence to an old film standard that is no longer relevent in the
world of digital image capture.

Member said:

But Pentax is a wild card. Remember this?

http://www.cameraquest.com/jpg5/p11025a.jpg

110 format SLR, with the entire camera scaled down to fit. I can
seriously see them trying something like this in the "4/3 system".

Click to expand...

Yep, I remember this well - I used to own a Pentax 110 SLR plus
lenses about twenty or so years ago. It was a lovely little system
that was perfect for hiking etc, and it produced surprisingly good
results. But it is apparent that, size-for-size, digital can
surpass film in terms of quality and interestingly the size of the
E-1 4/3rd sensor is within a mm ot two of the size of the old 110
film format. In the past I have stated more than once in these
forums that I never could understand why Pentax did not resurrect
their old 110 film SLR and stick a 2/3 or 4/3 size CCD in it!

Always great to hear your opinions and thoughts Joe, they are
nuggets of gold in these forums.

Terry.

Anastigmat · Jul 22, 2004

bmvmadeira said:
you can't make 100% perfect silicon waffer (not contaminated) so a
bigger chip will allways be more sensitive to random defects
besides the smaller number per waffer difference

Stop making excuses for the chip makers please. Unless they can show us the real numbers, no one has any idea what the actual yield is. A lot of us are assuming that the yield is very bad because cameras with full frame sensors are so expensive. The yield for Kodak's sensor is good enough that its full frame digital is selling for nearly half the price of the Canon. And Canon's sensor has lower megapixel count (less chance for random defect that would result in a bad chip), and its CMOS sensor has fewer circuits because of its simpler design. Therefore Canon should get better yield than Kodak. Yield isn't the reason why full frame sensors are so expensive. The lack of competition is.

bmvmadeira said:
I would still prefer a FF sensor. But I don't know if will happen.
FF will allways be better but Dx sensors can win based on a
price/good enough combination.

It will happen. It is a matter of time. Competition can bring it about much more quickly. Without AMD, we may still be using much slower chips and paying higher prices for processors.

krusadr · Jul 22, 2004

HAE said:
This could actually be a version of the sensor in the D1x/h. Around
a year ago Bjorn Rorslett published a macor photo of the D1h sensor
on this site clearly showing four microlenses for each pixel,
indicating each pixel actually is made up of four photo sites. The
D2X has odd rectangular pixels and is double the D1H in pixel count
(2 photo sites per pixel?).

Is that a typo or has there been an announcement about the D2X ?

HAE said:
If you do that math the sensor in the
D1X/H should have 10.6 MP. I think it was Thom Hogan that suggested
that that The D1X/H and D70 sensors are Nikon modifications of Sony
designs. So this could be the D1X/H sensor with information
released to the public after Nikon's execlusive on it expired ???
Just a thought.

Ron Parr said:

I just noticed this one, although there are references to this
dating back to 2001. According to acrobat, the document was
created in July 2001:

http://www.adelco.it/online/company/Sony/pdf/ICX263AL.pdf

The 1.6 FPS rate is discouraging and would seem to limit how high

upmarket this sensor could move.

It is capable of 9FPS by skipping 1 out of every 6 vertical

pixels for a total of 441 lines. This seems a bit awkward since
it's too slow for video and too low res. to be useful for high
quality stills.

One has to wonder if Sony is continuing to develop this product,
perhaps waiting for the technology to improve the frame rate, or if
it was completely shelved in favor of the 6MP sensor that found its
way into the D100 and *ist.

Is this the prototype for the next generation of low to mid priced
digital SLRs, or just a relic?

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

Click to expand...

Ron Parr · Jul 22, 2004

Anastigmat said:
Of course, they would have to back off from such an absurdity when
pressed. But they keep insisting that the small sensor can yield
high quality images given technological advances, as though such
advances can overcome the inherent advantages of larger chips. All
else being equal, the larger chip will always have the advantage,
so it is irrational to keep trying to improve a small chip when a
bigger chip can be fitted without increasing camera size.

Assuming you're talking about 35mm film derived SLR bodies with reduced size sensors: The reason for pursuing improved reduced size sensors is cost. Problems with yield and limits on reticle sizes can create a steep price curve when moving beyond roughly APS size.

I sure hope that Canon finds a way to produce a 1.3X or larger sensor at a reasonable cost, but given the challenges I'm not going to criticize people who are focusing on reduced size sensors.

BTW, I should mention that the relatively low cost of the Kodak 24x36 sensor is an anomaly in the sense that they have been partnering with fabs (first Tower, now somebody else) that are experimenting with chip stitching as a way work around reticle size limits. (They don't stitch finished chips, but stitch the images on the wafer.) It remains to be seen if these tecniques will be effective in the long run or if they come at a price in terms of quality. It seems that they are using a larger line width process which gives them more slack but also has price in terms of quality. I'd say the jury is still out on Kodak's experiment.

Anastigmat said:
Anastigmat said:

Your arugment has not, so far, been based upon insights into the
physics behind any fundamental limits. Your argument has been to
present the smallest known pixel.

Click to expand...

The size limit on pixels is based on physics.

Not really.

Anastigmat said:
Anastigmat said:

Anastigmat said:

Not at all. I pointed out that a full frame sensor can better
handle 10-11 megapixels than an APS sized sensor.

Click to expand...

No. What you said was, "such a chip would not be the next
generation of DSLR sensor, since such a chip would almost certainly
be noisier, given its small size, than the current chip."

Click to expand...

That is precisely because a full frame sensor can handle such high
resolutions better than the APS sized sensor. If so, then they
will be the next generation of photosensors, not the APS sized ones.

All other things being equal, a larger sensor would do better, but a new 10-11 MP APS sized sensor could still do better than an old 6 MP APS sized sensor.

Anastigmat said:
Anastigmat said:

You are clearly indicating that you think such a chip would be
noisier despite the improvements in chip design in the past few
years.

Click to expand...

Yes it will, because whatever improvements made can be incoporated
into the larger chip to make it even lower in noise and/or higher
in resolution.

You said, "the current chip," so it's sensible to assume that you are talking about actual current chips and that "current chip" isn't referring to some hypothetical chip that doesn't even exist.

Anastigmat said:
Hope is not a good design strategy. So, why wait for technology to
improve so you can use the small sensors, when you can use the big
sensor and get the image quality that you need RIGHT NOW?

Erm... The technology for making such sensors at prices most people would be comfortable paying probably doesn't exist, so it's a question of where you want to pin your hope. Historically, cost per unit area of silicon has not gone done, while performance per unit area has gone up. For sensors this can't continue forever, but for large SLR sensors, there's no evidence yet that steam is running out.

Anastigmat said:
I am debating against those who worship the small chip and claim
that technology will keep improving the small chip so that there
will never be the need for full frame sensors. I point out that
small chips will always be lower in quality because there is a
limit on how small a photodiode can shrink. Besides, even before
that limit is reached, the image quality may not be adequate.
Hence a larger sensor is always cheaper.

I understand that there are some Nikon folks who have an interest in rationalizing reduced size sensors as a way of explaining away Nikon's lack of options. If it entertains you to taunt them, go ahead.

However, large sensors are not cheaper in any typical interpretation of the word. In principle, larger sensors will let you do things that you can't do with smaller sensors. That's certainly true, but it's also true of larger trucks, larger guns, and larger houses. Still, we don't typically think of those things as cheaper.

Anastigmat said:
Anastigmat said:

Anastigmat said:

Am I not correct that a sensor with smaller pixels would be noisier?

Click to expand...

As a categorical statement this is not correct. As a ceteris
paribus statement it might be, but you should really qualify it as
such.

Click to expand...

Why would a small pixel not be noisier if all else is equal?

Ceteris paribus a smaller pixel will be noisier, but you need to say the cetris paribus part if you're in a discussion where ceteris are rarely paribus.

Anastigmat said:
Anastigmat said:

I the current context, the discussion is about a move from an old
sensor with large pixels to a new sensor with smaller ones that are
still pretty large, so your statement is neither correct nor
incorrect without further information. It's pure speculation.
That's the point.

Click to expand...

Does the new sensor exist? How do you know that the new technology
made the noise level the same while resolution has been increased?

Nobody knows, which is precisely why I'm discouraging categorical comments such as your original one.

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

Ron Parr · Jul 22, 2004

Terry Sessford said:
Enough to make no visible difference on, say, an A3 sized print,
and how many users print larger than this?

I always laugh when people talk about print sizes with the implicit assumption that consumers never crop images.

Member said:
But if you are going to crop and throw away pixels then why not
simply have a smaller sensor in the first place with all its size,
weight and cost advantages?

I just do not buy these contrived arguments for a full frame sensor.

Wow... Now THAT's a contrived argument for a small sensor. Seriously - I think you understand why peoplle need or want to crop images and what difference is between having the flexibility to crop and having a tiny sensor. Pretending not to understand this as some kind of rhetorical device is just silly.

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

Ron Parr · Jul 22, 2004

Peter Herth said:
"enough" means, that the mayority of the people are not willing to
pay significantly more for quality improvements. Furthermore, while
there are individual differences in what is considered a good
price, for every technology, there is a point where increases in
the price result in only small performance gains.

Exactly. The issue is that it's quite difficult to claim to have insights into what "enough" will be 5-10 years down the road when we don't know what the price/performance tradeoffs will be and when, despite the bluster around here, none of us really knows what Joe consumer's preferences are.

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

Ron Parr · Jul 22, 2004

Anastigmat said:
The size of the 10D and its sensor chip is evidence that a small
chip does not result in a small digital camera. The Digital Rebel
shrunk even though it has the same size chip as the 10D. So there
is no evidence that a full frame chip would make a camera larger
but there is evidence that a smaller camera is possible without
reducing chip size. The Pentax *ist D has a bigger chip than the
Digital Rebel, but it is a smaller camera. So, these are all facts
which argue against a smaller chip resulting in a smaller camera.

On the contrary, all the facts indicate that other factors are currently pushing camera sizes up but these other factors are ones that have fairly predictable trends towards shrinking over time.

Anastigmat said:
A full frame sensor with 11 megapixels will have roughly the same
amount of parts as an 11 megapixel APS sized sensor. So the chance
that a random defect will occur in either of them would be about
the same. So, they should have identical yields.

You have the wrong model of how chip defects arise. The number of "parts" has little to do with it.

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

Ron Parr · Jul 22, 2004

Anastigmat said:
bmvmadeira said:

you can't make 100% perfect silicon waffer (not contaminated) so a
bigger chip will allways be more sensitive to random defects
besides the smaller number per waffer difference

Click to expand...

Stop making excuses for the chip makers please. Unless they can
show us the real numbers, no one has any idea what the actual yield
is. A lot of us are assuming that the yield is very bad because
cameras with full frame sensors are so expensive.

The reasons for the assumption of low yield follow from extremely well studied models of chip yield and basic mathematics for the chances of hitting wafer defects. All of this was worked out long before most people starting thinking about digital camera sensor sizes, so people aren't making excuses for chip makers; they're informing you about fairly basic economics of chip manufacture.

The only variable here is how tolerant sensors are of defects. Clearly, they're more forgiving than logic gates, but we don't know how much.

Anastigmat said:
The yield for
Kodak's sensor is good enough that its full frame digital is
selling for nearly half the price of the Canon.

Kodak's suppliers are doing some experimental stuff with fabrication. It's obvious from the quality of the first round, that they're still getting the bugs out. It will be interesting to see if this really turns out to be a viable method and if these techniques become widely adopted.

Anastigmat said:
And Canon's sensor
has lower megapixel count (less chance for random defect that would
result in a bad chip), and its CMOS sensor has fewer circuits
because of its simpler design. Therefore Canon should get better
yield than Kodak.

The megapixel count probably has very little to do with it.

Anastigmat said:
Yield isn't the reason why full frame sensors
are so expensive. The lack of competition is.

Given your other comments, it seems that you don't really have a great handle on how the chips are made, so I think it would behoove you to refrain from making such strong statements.

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

Ron Parr · Jul 22, 2004

Anastigmat said:
Let's assume that your calculation is correct.

The example showed an 18X price differential, so if you assume that the sensor is the most expensive part of the camera (a very reasonable assumption that is true even for small consumer cameras), then the difference in market price beteween the 300D and the 1Ds is actually must less than what would be projected.

Anastigmat said:
So, why would a 1.3x format 8mp digital camera selling for $1k be
impossible, as some posters have argued.

Based entirely on yield and the defect rate assumed in the applet, one would expect 1.3X sensors to cost 2X as much to produce and a selling price of nearly double the corresponding 1.6X camera.

However, this ignores that exotic equipment and techniques may be needed to produce chips larger than 20mm or so in length.

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

(aborted?) Sony 10.66 MP APS (D100, etc.) sized sensor

Forum Pro

Forum Pro

Ron Parr

Guest

Forum Pro

Senior Member

Forum Pro

Terry Sessford

Guest

Forum Pro

Forum Enthusiast

Leading Member

Senior Member

Senior Member

Forum Pro

Senior Member

Ron Parr

Guest

Ron Parr

Guest

Ron Parr

Guest

Ron Parr

Guest

Ron Parr

Guest

Ron Parr

Guest

Keyboard shortcuts