Why not make the sensor larger?

On Jun 13, 11:26 am, "Victek" wrote:
It seems like the biggest weakness of mid-priced cameras is noise/iso issues
due to small sensors. Well, why don't the manufacturers use a larger
sensor? How much larger would it have to be to eliminate the worst of the
noise and provide a useful iso range? Would the increased cost make the
camera unmarketable? In other words, is there a real economic problem or is
it just ignorance?

Once someone other than Canon offers a FF sensor, you'll be amazed at
how fast the price drops.
But then why do most people, except for pros on the cutting edge NEED
a FF or larger sensor?

ray wrote:
[]
Perhaps you could elaborate - it seems to me that, for example, if a
DSLR uses a 2/3 size sensor it could probably be made 2/3 size - is
it that folks really expect a full size camera? Seems to me that all
you need to accomodate is the sensor size and the optics distance -
electronics are VERY compact.

They could, but instead they choose to use the old 35mm production lines
(which are already paid-for, of course) and just keep the same larger
size. More profit for the manufacturers. Even more disappointing that
Olympus with their "half-size" 4/3 system didn't offer half-sized cameras
and lenses. Some people find that even some of the current DSLRs are too
small for them.

David

On Jun 13, 1:31 pm, "David J Taylor" -this-
bit.nor-this-part.co.uk wrote:
Victek wrote:

[]

Thanks for information. I enjoyed reading the review of the Sony
DSC-R1. It's interesting to note that it costs more than many DSLRs
today. Perhaps moving to a DSLR is really the only way to overcome
the limitations of the "super-zoom" models (such as the S3-IS, or the
Panasonic FZ series)?

Victek,

It really depends what limitations are important to you. Low-light will
certainly be better with a larger DSLR sensor and expensive, large
aperture lens. Plus that will be a lot heavier to carry around if the
lens is reasonably telephoto. In the FZ5 I have a 432mm f/3.3 image
stabilised lens in a package weighing about 300g. f/2.8 at 432mm with the
Panasonic FZ20.

Nearest DSLR and similar lens? Nikon D40 + 55-200mm VR zoom weighing 520g
+ 335g, so nearly 3 times as much. This is an f/5.6 lens, and it doesn't
include wide-angle coverage. So you would require a second lens.
Providing a good wide-tele coverage is the 18-200mm VR, but that is still
only f/5.6 and the total weight is now 520 + 560g, so over 1kg. Perhaps
you're OK with that. I don't think that either of these lenses offer the
same macro capability as the Canon or Panasonic super-zooms do, and no
DSLR offers a swivel LCD or movie mode.

There may be a half-way house with the so-called 4/3 sensor used in the
Olympus DSLRs, but although the sensor is smaller than a conventional DSLR
(nearer "half-frame" size), the cameras are not lighter and the lenses are
not smaller. You don't want to ask the cost of their lenses either.

Not to split hairs but Olympus's E-410 is the smallest and I believe
the lightest DSLR on the market.
It's new kit lenses are also probably the smallest and lightest in
their class and far better optically than the likes of Canon, for
e.g. The costly Olympus lenses are the pro lenses that run $1000 or
more, but you can buy prosumer lenses from them for around $399-
$900.00.

"Prometheus" wrote in message
news
In article , Pete
writes
The only hope for a low-noise compact super-zoom is probably to cool the
sensor with e.g. a Peltier device?

Probably not by much since the main noise contribution, at least for the
shorter exposures under a few seconds, is the random distribution of the
photon flux.
--

How much of the noise in current compact super-zoom cameras is caused by
pushing the number of megapixels? Instead of constantly increasing
megapixel count the industry should actually be moving in the opposite
direction. There must be an optimal noise to megapixel balance for a given
sensor size that would also yield better low light performance and better
picture quality over all, but it seems that no one can get off the megapixel
pony g.

Victek wrote:
It seems like the biggest weakness of mid-priced cameras is noise/iso
issues due to small sensors. Well, why don't the manufacturers use a
larger sensor? How much larger would it have to be to eliminate the
worst of the noise and provide a useful iso range? Would the
increased cost make the camera unmarketable? In other words, is
there a real economic problem or is it just ignorance?

I suspect there are some real economic issues.

I don't know about the manufacturing of sensors, but I suspect they may
have the same kind of issues as IC's.

I have some knowledge of IC manufacturer (well like 15-20 year old
technology). In the manufacturing process chips (sensors?) were made on
round surfaces. The size of that circle was limited by the machinery that
can handle them in the manufacturing process under a vacuum. Making an IC
even slightly larger (say 10%) could result in a 50% or more reduction in
the number that could be made in a batch. That was quite an issue.

--
Joseph Meehan

Dia 's Muire duit

"Victek" wrote in
t:

How much of the noise in current compact super-zoom cameras is caused
by pushing the number of megapixels?

None, in many cases. Not in the image, anyway, even if it means more
noise in each pixel. The shot noise component of image noise depends on
the total mumber photons per fractional area of the entire image. More
pixels just gives more resolution. Read noises are dropping with newer
cameras, and they need to be divided

However, because of the higher noise at the pixel level, many
manufacturers use ridiculously heavy-handed noise reduction techniques
that take some detail with it, and leave an artificial texture.

Instead of constantly increasing
megapixel count the industry should actually be moving in the opposite
direction. There must be an optimal noise to megapixel balance for a
given sensor size that would also yield better low light performance
and better picture quality over all, but it seems that no one can get
off the megapixel pony g.

Your belief is an illusion, IMO. I have never seen any evidence to
support better imaging from bigger pixels. Bigger *SENSORS*, yes.

Take a 2MP image from yesteryear's compact, and upsample it to the same
size as today's 10MP (choose the best you can find of each, both of the
same sensor size). You will see sharper noise in the 10MP, with soft
noise, most likely of a higher intensity in the 2MP, detail in the 10MP
where there is blur in the 2MP.

You can simulate the same with binning or downsampling; take a 10MP image
from a compact with a sharp lens, make a duplicate and downsample it or
bin it to 1 or 2 MP, then upsample it back to 10MP. View the two next to
each other, with the same total size, at least 100% for the 10MP. The
binned/downsampled one is garbage when viewed at the same size. The only
time it is better is when the pixels oversample the optics to a
significant degree, and no real image detail is lost in the binning or
downsampling.

Bigger pixels, downsamples, and binnings are only advantageous with
sufficient optics for higher pixel counts when you are never going to be
able to see most of the detail, such as displaying on low-res digital
picture frame, or on a televison, or a web-sized image.
--


John P Sheehy

How much of the noise in current compact super-zoom cameras is caused
by pushing the number of megapixels?

None, in many cases. Not in the image, anyway, even if it means more
noise in each pixel. The shot noise component of image noise depends on
the total mumber photons per fractional area of the entire image. More
pixels just gives more resolution. Read noises are dropping with newer
cameras, and they need to be divided

However, because of the higher noise at the pixel level, many
manufacturers use ridiculously heavy-handed noise reduction techniques
that take some detail with it, and leave an artificial texture.

Instead of constantly increasing
megapixel count the industry should actually be moving in the opposite
direction. There must be an optimal noise to megapixel balance for a
given sensor size that would also yield better low light performance
and better picture quality over all, but it seems that no one can get
off the megapixel pony g.

Your belief is an illusion, IMO. I have never seen any evidence to
support better imaging from bigger pixels. Bigger *SENSORS*, yes.

Well, it's not the first time I've discovered I don't know what I'm talking
about. I'm interested to see what others have to say about this, too.

Victek wrote:

"Prometheus" wrote in message
news
In article , Pete
writes
The only hope for a low-noise compact super-zoom is probably to cool
the sensor with e.g. a Peltier device?

Probably not by much since the main noise contribution, at least for
the shorter exposures under a few seconds, is the random distribution
of the photon flux.
--

How much of the noise in current compact super-zoom cameras is caused by
pushing the number of megapixels? Instead of constantly increasing
megapixel count the industry should actually be moving in the opposite
direction. There must be an optimal noise to megapixel balance for a
given sensor size that would also yield better low light performance and
better picture quality over all, but it seems that no one can get off
the megapixel pony g.

The number of photons captured by a pixel is directly
proportional to the active area of the pixel. For properly
metered scene, a 20% diffuse reflectance spot will
deliver about 3200 photons per square micron to the
focal plane in the green passband regardless of exposure,
f-stop, focal length, or sensor size. And only a fraction
of those photons make it through the filters and get converted
to electrons by the pixel. Thus the larger the pixel,
the more photons it collects. Noise is proportional
to the square root of the number of photons converted.
So the majority of noise in most digital camera images
is photon noise (Poisson statistics).

Camera manufacturers used to make very small "full frame"
35 mm film cameras. There is no reason they couldn't make
equal sized digital P&S cameras with sensors the same
size as in DSLRs (like APS-C sensors). I believe we will
see that soon. They will cost a little more (chip manufacturing
goes up exponentially with the size of the chip), but
APS-C DSLRs are well below $1000 now, so APS-C P&S small
cameras could be reasonably priced and made now.
Manufacturers probably think such a camera might
reduce DSLR sales, and they would probably be right.
But I have a DSLR and I would buy one.

More on pixel size and sensors:
http://www.clarkvision.com/imagedeta...el.size.matter

http://www.clarkvision.com/imagedeta...rmance.summary

Roger

On Wed, 13 Jun 2007 19:52:00 -0700, Victek wrote:

Instead of constantly increasing
megapixel count the industry should actually be moving in the opposite
direction. There must be an optimal noise to megapixel balance for a
given sensor size that would also yield better low light performance
and better picture quality over all, but it seems that no one can get
off the megapixel pony g.

Your belief is an illusion, IMO. I have never seen any evidence to
support better imaging from bigger pixels. Bigger *SENSORS*, yes.

Well, it's not the first time I've discovered I don't know what I'm talking
about. I'm interested to see what others have to say about this, too.

For someone to say that your belief is an illusion, they really
should explain why and perhaps give an example or two, rather than
simply say that it's an opinion based on evidence not seen.

When superior sensors are used, they do have larger pixels. If
compared with another sensor having the same number of smaller
pixels, it will of course require a larger sensor, but it's not the
sensor *size* that determines the low light performance and picture
quality. For example, take a large sensor 8megapixel sensor that
has excellent low noise/high ISO performance as well as a large
dynamic range (both due to the large pixel size). If the sensor
manufacturer produced an otherwise identical sensor, but made it 1/4
the size, it would only have 2megapixels, but those pixels would be
the same size as the ones on the 8mp sensor. All other things being
equal, there should be no difference in noise characteristics and
image quality between the 2mp and 8mp sensors, even though there's a
tremendous difference in the sensor size.

Fuji's F30 has excellent low noise/high ISO performance for a P&S
and it does have a relatively large sensor, 1/1.8" IIRC. But a
significant reason why it does so well is that its sensor isn't
10mp, but was limited to only 6mp. Of course there are other
factors, such as the quality of the image processing engine, etc.,
but those apply to all cameras and are independent of sensor size.

In one respect I'd agree with John. Pixel sizes aren't generally
advertised or known by prospective camera buyers, but the sensor
size is usually included in manual spec's and reviews. Given that,
you probably wouldn't be too far off assuming that for a given
number of megapixels, the cameras having larger sensors would
produce better images and have better high ISO performance.

On Wed, 13 Jun 2007 21:02:31 -0700, Roger N. Clark (change username to
rnclark) wrote:

snip
Camera manufacturers used to make very small "full frame"
35 mm film cameras. There is no reason they couldn't make
equal sized digital P&S cameras with sensors the same
size as in DSLRs (like APS-C sensors). I believe we will
see that soon.
snip

An APS-C P&S would certainly be an interesting product, but it surely
couldn't provide the same lens specs as a small-sensor digicam without the
size, weight and cost going up astronomically.

My relatively ancient Pana FZ10 provides f/2.8 over its full 12X 35-432mm
zoom range (35mm equiv), and will macro down to 2 inches. If there is an
APS-C lens that will do that, I'd guess it costs well into 4 figures and
needs a mule to carry it.

Pete