K20D 35mm sensor?

J_K said:

OK, equivalent ISO will be higher, but S/N will be the same (since
pixels are appropriately larger).

Click to expand...

Correct.

J_K said:

Good point! You may be right about the plastic effect. And since the
diffusion time scales as the distance squared, it should give
(theoretically) a clear advantage to FF. As for CMOS vs CCD, I have
no idea, I am not working in that specific field.

Click to expand...

Neither am I.

I am doing single electronics and molecular electronics - not so applicable yet for a practical sensor design.

Regards,
Andrey.

I hope not.
--
Fongky

http://www.flickr.com/photos/fongky/

Dear all,

Lot of thanks to all of you who participated in the discussion.

For those of you who are really bored and have nothing else to read tonight, I’ve tried to make a short summary.

For any shot taken with ¾ sensor (shutter, f-number, ISO) an equivalent shot can be taken with FF sensor (same shutter, f-number*2, ISO*4). An equivalent shot will have the same FOV, DOF and S/N.

For some shots taken with FF sensor an equivalent shot can be taken with ¾ sensor with the same shutter, f-number/2, ISO/4. An equivalent shot can not be taken if f-number/2

FF has an advantage if you can afford longer shutter and/or you are aiming at shallow DOF: an equivalent combination of (shutter, f-number, ISO) can not be available for smaller sensor.

Though the universal scaling law ensures equal S/N for equivalent images, there are some absolute numbers which put the constraints on the equivalence: the minimum feasible f-number, the native sensor ISO, the registered distance, and the characteristic diffusion length.

1.The minimum f is 0.5. The closer to 0.5 the more problematic is aberration correction. 50 1:2,8 for FF is not the same as 25 1:1.4 for ¾. Generally speaking, even for equivalent shots one should expect better performance from equivalent FF lens.

2.The registered distance (from the lens mount to the sensor) is about 40 mm. It means that a “normal” 35 mm lens for APS must have a retrofocus design, while 50 mm for FF not. For wide angle lenses a smaller sensor calls for more extreme retrofocus design which (at least potentially) translates into more aberrations.

3.The current technology provides native sensor ISO in the range 100-200. With a given pixel size you can not collect as many electrons as you wish – the smaller pixels saturate faster.

4.Because of diffusion, there is always a chance that the photon landed on some particular pixel will generate an electron which will be eventually trapped in some pixel nearby. Though diffusion makes very little effect on the total noise, it substantially reduces the S/N ratio in low contrast areas with tiny details (like skin texture). The diffusion is more noticeable for smaller pixels (=smaller sensors).

Note that 1. and 2. have nothing to do with noise and the sensor design.

3. and 4. are about sensor design indeed, but that’s not what people actually have in mind when they say smth. like ‘bigger pixels = better image’.

Now, when you’ve read all this, go to your Christmas table and have a clean-memory party.

Best regards,
Andrey.

Perhaps, but isn't this mainly due to poor optical (and inexpensive) design? I still think that for an additional $250 cost to the camera, than an outstanding optical magnification of the VF is possible, even for those wearing eye-glasses. Yes the prism box would likely be significantly larger, but then, why not make it interchangeable while they're at it like on the old Nikon F1. Sometime I wonder when the older designs were lost in the digital age transition.

--
Roger

To heck with the moderator: I'm telling my MOM!

And my dad can beat up your dad!

Let's play nice, children; it's only pixels on a screen, now.

Thanks, and Merry Christmas or whatever you celebrate, everyone!
Cameron

Which would be about what I'd expect, for optimal performance. Basically take a K100 sensor and scale it up by a factor of 2.34 and you get 14MP. A sensor with pixels that big could have near D3 performance levels, and it would deliver 6MP in APS-c "crop mode"

J_K said:

Currently, very true. However, with decreasing price difference, more
and more users may find FF to be a better solution in the future.

Click to expand...

There's no reason to expect a decreasing price difference. Full frame sensor costs have dropped for a given capability, but not as fast the APS-C costs have.

A physically larger chip costs more because it takes up a larger part of the the 200 or 300 mm silicon wafer from which the individual chips are made.

The wafer is a fixed, fairly high, cost; the more chips you can get per wafer, the more you have to spread the cost of the wafer over. (Since the ones that work have to carry the cost of the ones that don't work, there are advantages to small and simple in terms of the percentage of working chips you get per wafer, too; this is usually referred to as 'yield'. Low yield = high prices.)

The other thing is that the larger the physical components -- transistors, etc. -- of the chip are, the more power it takes to run them. This is the other major economic incentive running the miniaturization process; not only do you get more chips per die, you get to use smaller batteries, or get more operating time, or use less active cooling hardware, etc. Pretty much everyone making portable devices -- whether ipods, cameras, or satnav -- cares about this a lot.

So, at a given fabrication technology generation, you can get more sensors into a larger physical die (the actual silicon part of the chip = the die). But you get fewer working chips per wafer, and as the fabrication tech advances, the cost per wafer goes up. (Generally offset by the size of the dies going down, so you get many more chips per wafer, but a 'full frame' camera sensor is a fixed physical size; you can't shrink the size of the die for a full frame sensor.)

So as fabrication technology improves, the cost advantage of the APS-C sensor will get larger, not smaller.

Also consider that Pentax has gone for sensor stabilization; a larger sensor is heavier, and needs more powerful motors and more power to move it around to stabilize it. That chews into the camera power budget and the required volume for the sensor (larger motors, more space to swing a larger sensor, etc.) So larger sensors aren't attractive from a shake reduction point of view.

Also, full frame = heavier lenses. People don't like huge heavy lenses. (People may like the results, but who likes the fact that the great long telephoto with which they have taken so much lovely nature photography weighs thirty pounds?)

Then consider that the human retina is smaller than an APS-C sensor and does a much better job.

Exploiting fabrication improvements to include a better dynamic range of sensors in a single APS-C die seems much more likely to be economically successful than going for the same thing in a full frame sensor, because the APS-C sensor size is relatively less disadvantaged by being required to be a fixed physical size by the progress in chip fabrication technique.

--
StephenG

Pentax K100D
Fuji S5200
Fuji E900
PCLinuxOS

cause its part of the amarican dream....

Those that do don't read or post in this forum (nt)

sjgcit said:

Can't any of you just wait a month?
--
StephenG

Pentax K100D
Fuji S5200
Fuji E900
PCLinuxOS

Click to expand...

rogerstpierre said:

Perhaps, but isn't this mainly due to poor optical (and inexpensive)
design? I still think that for an additional $250 cost to the
camera, than an outstanding optical magnification of the VF is
possible, even for those wearing eye-glasses. Yes the prism box
would likely be significantly larger, but then, why not make it
interchangeable while they're at it like on the old Nikon F1.
Sometime I wonder when the older designs were lost in the digital age
transition.

Click to expand...

Isn't the reason that interchangeable finders are no longer provided is that they would probably play havoc with the metering accuracy?

If stray light from a normal eye level viewfinder can mess up metering, imagine what an open waist level finder would do!

I guess that is why Live View is being developed and fitted to the newer models.

The only other way is to instigate real time metering during the actual exposure (after the mirror is up), the only challenge being that there would be no pre-indication of the exposure setting, unless a small blackout period is made to meter beforehand to give a reading (a bit like pre-flash with P-TTL).

If I remember correctly, real time OTF (off film) metering was a feature of the Olympus OM2 and the Pentax LX. My OM2n certainly had the feature, I believe it was the first camera with TTL and OTF. The whole OM series was the brainchild of Yoshihisa Maitani over a 30 year period, this interview with him makes very interesting reading.
http://www.geocities.com/maitani_fan/om_interview_2.html

As a concept, I feel that the OM system was a much more complete vision than anything Pentax had, the LX, which was perhaps the finest of Pentax's achievements, was very much a reaction to the OM challenge.

With the demise of the OM system and his retirement in 2002 went a whole style of entrprenuership that doesn't (probably and sadly cannot) exist today.

It is significant that both companies have struggled, thankfully successfully, to transition from those lovely metal masterpieces to the hi-tech composite computers we use today, even now, they are probably the closest (emotively) of the various camera producers.
--
Richard Day - 'Carpe Diem!'
Gloucester UK

richardday said:

If I remember correctly, real time OTF (off film) metering was a
feature of the Olympus OM2 and the Pentax LX. My OM2n certainly had
the feature, I believe it was the first camera with TTL and OTF.

Click to expand...

I don’t understand how of-the-film metering could be possible.

At fast shutter speeds the exposition is controlled by adjusting the gap between curtains, while the shutter moves at fixed speed. To set the proper shutter opening one should measure exposition before the shutter is fired. One obvious solution would be to measure the light reflected from the shutter, but this, I believe, will make the things quite messy for flash TTL in slow speed synchronization mode.

Regards,
Andrey.

TTL FLASH was a MINOLTA invention! OFT metering could be too.. i remember reading about it years ago... when the X700 was released in 82, it had OFTand TTL flash, they said it was their invention...remember Leica and minolta were partners for about 20 years, R3 Electronic was a M XE1, R4 etc was the old XD, acumate focussing screens were used under licence by Hasselblad..

XM MOTOR what a beast it was.. C&N had nothing like it at the time 1975.. those were the days..
--
Know The Truth, and The Truth will Set You Free

Rick104 said:

TTL FLASH was a MINOLTA invention! OFT metering could be too.. i
remember reading about it years ago... when the X700 was released in
82, it had OFTand TTL flash, they said it was their
invention...remember Leica and minolta were partners for about 20
years, R3 Electronic was a M XE1, R4 etc was the old XD, acumate
focussing screens were used under licence by Hasselblad..
XM MOTOR what a beast it was.. C&N had nothing like it at the time
1975.. those were the days..
--
Know The Truth, and The Truth will Set You Free

Click to expand...

The OM-2 had OTF flash metering and I got mine in the 1976. I always thought (but could be wrong) it was an Olympus invention.

Dave

RaSoFe said:

richardday said:

If I remember correctly, real time OTF (off film) metering was a
feature of the Olympus OM2 and the Pentax LX. My OM2n certainly had
the feature, I believe it was the first camera with TTL and OTF.

Click to expand...

I don’t understand how of-the-film metering could be possible.
At fast shutter speeds the exposition is controlled by adjusting the
gap between curtains, while the shutter moves at fixed speed. To set
the proper shutter opening one should measure exposition before the
shutter is fired. One obvious solution would be to measure the light
reflected from the shutter, but this, I believe, will make the things
quite messy for flash TTL in slow speed synchronization mode.

Click to expand...

Read about it here: http://www.geocities.com/maitani_fan/om_2.html
--
Richard Day - 'Carpe Diem!'
Gloucester UK

DLBlack said:

Rick104 said:

TTL FLASH was a MINOLTA invention! OFT metering could be too.. i
remember reading about it years ago... when the X700 was released in
82, it had OFTand TTL flash, they said it was their
invention...remember Leica and minolta were partners for about 20
years, R3 Electronic was a M XE1, R4 etc was the old XD, acumate
focussing screens were used under licence by Hasselblad..
XM MOTOR what a beast it was.. C&N had nothing like it at the time
1975.. those were the days..
--
Know The Truth, and The Truth will Set You Free

Click to expand...

The OM-2 had OTF flash metering and I got mine in the 1976. I always
thought (but could be wrong) it was an Olympus invention.

Click to expand...

You probably had one of the first OM2s, I believe the first models appeared in 1975. I had an OM1 and two OM2-n bodies, one silver and one black, I think I bought them in 1977 or 78. They were fabulous cameras and lenses. I finally parted with most of my OM kit in 1995, my eldest son kept the OM1, 3 lenses and some other accessories.

Was it the first OTF system? - Well it seems it probably was, although it was never disclosed if anyone else had done it prior to Olympus. If they had, they either never patented it or bothered to challenge it. The fact that Pentax utilised the same concept in the LX in 1980 suggests that it was never patented by anyone, so fell into the common usage domain.
--
Richard Day - 'Carpe Diem!'
Gloucester UK

richardday said:

Read about it here: http://www.geocities.com/maitani_fan/om_2.html

Click to expand...

Thanks for the link!
From the link:

*************

The curtain is printed with computer-generated random digital pattern which simulates the average reflectance of a broad variety of film surfaces.
*************

Oh!
Very smart indeed. Now I see how it worked.

OTF is great feature, but nowadays people expect no less than 1005 zones from AE sensor, AE linked to AF, AE-based Scene Recognition System etc.
I don’t think one can go that far with the diffusive reflected light.
May be, that was the reason why Pentax abandoned TTL flash.

Even if people really don’t need scene recognition, the marketing guys will hypnotize them until they do.

Regards,
Andrey.

Hello Richardday, heres a link.. http://en.wikipedia.org/wiki/Minolta

YEAH EVERYONE THINKS ITS AN OLY INVENTION... MINOLTA WAS BIGGER THAN WHAT EVERYONE THOUGHT..

1972: Minolta signs an agreement to cooperate with Leica in SLR development.
1973: The Minolta CL is the first fruit of this agreement.

1976: The Leica R3 is introduced. Minolta produces the R3, R4, and R5 models in the Leica R series. Subsequent cameras are built in Germany by Leica themselves.

1981: Implementation of Minolta's invention and patent of TTL OTF Through The Lens Off The Film exposure metering: The Minolta CLE is the first 35 mm rangefinder camera to feature TTL metering and aperture priority auto-exposure. The Minolta X-700 manual-focus SLR is introduced; this model is sold until 1999 and is enormously successful. The Minolta XD-11 (Model E) is the first Minolta product branded with an updated logo in caps, which was in use until the 2003 merger with Konica.

Cheers
--
Know The Truth, and The Truth will Set You Free!

How can any company survive after this..lol

"Unfortunately for Minolta, its autofocus design was found to infringe on the patents of Honeywell, a U.S. corporation. After protracted litigation, Minolta in 1991 was ordered to pay Honeywell damages, penalties, trial costs and other expenses in a final amount of 127.6 million dollars (source: NY Times)."

I oringinally thought it was only 5 million...
--
Know The Truth, and The Truth will Set You Free!

Rick104 said:

Hello Richardday, heres a link.. http://en.wikipedia.org/wiki/Minolta

YEAH EVERYONE THINKS ITS AN OLY INVENTION... MINOLTA WAS BIGGER THAN
WHAT EVERYONE THOUGHT..

1972: Minolta signs an agreement to cooperate with Leica in SLR
development.
1973: The Minolta CL is the first fruit of this agreement.
1976: The Leica R3 is introduced. Minolta produces the R3, R4, and R5
models in the Leica R series. Subsequent cameras are built in Germany
by Leica themselves.
1981: Implementation of Minolta's invention and patent of TTL OTF
Through The Lens Off The Film exposure metering: The Minolta CLE is
the first 35 mm rangefinder camera to feature TTL metering and
aperture priority auto-exposure. The Minolta X-700 manual-focus SLR
is introduced; this model is sold until 1999 and is enormously
successful. The Minolta XD-11 (Model E) is the first Minolta product
branded with an updated logo in caps, which was in use until the 2003
merger with Konica.

Click to expand...

It is possible to patent a variation of an existing technology. Minolta introduced their version in 1981, at around the same time as the Pentax LX (1980 I believe).

The Olympus OM2 was introduced in 1974. Minolta never challenged the Olympus design, so, Olympus were the first to implement the concept AFAIK.

I'm not denying that Minolta were a great camera maker, I regret their passing just as much as anyone else, who would have predicted it a few years ago when the Dimage A1 & A2 models were so popular, their delay to make an interchangeable lens DSLR and attempt to rely on the bridge format was probably their ultimate downfall. Unfortunately, time moves on and some are left behind. Let's face it, even Pentax are no longer the same company, though hopefully their essence will remain intact. So far, so good, only time will tell.
--
Richard Day - 'Carpe Diem!'
Gloucester UK