semilog
curmudgeonly optimist
Well, yes, "giving it time" (i.e., a longer exposure) is one way to increase SNR on a sensor.
Seriously, though, in high-end astronomy, microscopy, etc., what's limiting light sensitivity at this point is the number of photons falling on the sensors — not so much the sensors themselves, anymore. And for a given subject, the only way to get more photons is with a bigger numerical aperture.
Those high-end cameras already exhibit 90% quantum efficiency and low read noise. The chips also tend to be held under high vacuum and at low temperatures, below -40 C. They also tend to have larger pixels (6.5 um and up).
A huge amount of what's happening in sensor development now is optimisation of small pixels (below 5 um, which would be a 36 megapixel FF sensor). The electronics required to read and reset charge accumulated at each pixel take up space that is not light sensitive (shading factor). Much development right now, such as gapless microlens tech, is motivated by decreasing shading factor.
The major motive for doing this is to improve small sensors, especially in cell phones, since their shading factors tend to be a lot higher than with larger pixels. Some of those efforts are trickling into larger-sensor cameras, which is why you see increasing pixel counts as on the Canon 7D.
It is no an accident that the first mass market devices with back-thinned sensors are cell phones, not DSLRs.
Anyway, the consumer sensors are closing in on what is physically possible faster than many realise.
Last edited:
kshapero
South Florida Man
Fujifilm's FinePix X100 ships March 2011 for $1,000, we go hands-on
Engadget
Engadget
RayPA
Ignore It (It'll go away)
How about an audible, beeping distance scale, where A440 = 2M?
LOL! That's actually not a bad idea!
/
PKR
Veteran
Well, yes, "giving it time" (i.e., a longer exposure) is one way to increase SNR on a sensor.
Seriously, though, in high-end astronomy, microscopy, etc., what's limiting light sensitivity at this point is the number of photons falling on the sensors — not so much the sensors themselves, anymore. And for a given subject, the only way to get more photons is with a bigger numerical aperture.
Those high-end cameras already exhibit 90% quantum efficiency and low read noise. The chips also tend to be held under high vacuum and at low temperatures, below -40 C. They also tend to have larger pixels (6.5 um and up).
A huge amount of what's happening in sensor development now is optimisation of small pixels (below 5 um, which would be a 36 megapixel FF sensor). The electronics required to read and reset charge accumulated at each pixel take up space that is not light sensitive (shading factor). Much development right now, such as gapless microlens tech, is motivated by decreasing shading factor.
The major motive for doing this is to improve small sensors, especially in cell phones, since their shading factors tend to be a lot higher than with larger pixels. Some of those efforts are trickling into larger-sensor cameras, which is why you see increasing pixel counts as on the Canon 7D.
It is no an accident that the first mass market devices with back-thinned sensors are cell phones, not DSLRs.
Anyway, the consumer sensors are closing in on what is physically possible faster than many realise.
One popular way to increase S/N is through cooling the sensor package. This is done in many applications where a source of power is not an issue and very high resolution is needed. i don't think you'll be finding any SLRs with Peltier cooling any time soon. Are the sensors on your microscopes cooled? p.
semilog
curmudgeonly optimist
JayGannon
Well-known
I know one or two guys working in CMOS development who would disagree. With traditional thinking yes but theres still at least 10-15 years left in the hardware development field. On the otherhand I am not talking about miniature sensors I am talking about are in the 35mm range. I dont know if thats what you were classing as consumer (I am assuming you we're). And yes there is cooling issues but a not Pelitier/Sub Zero balance can be found with active but compact cooling systems. Chilled but not too cold
semilog
curmudgeonly optimist
LOL! That's actually not a bad idea!
Just what I need -- to carry a tuning fork in my camera bag!
PKR
Veteran
It would be great to be able to do that in a camera, but the batteries would be an issue, as would be the moisture.
A friend designed a sensor for core analysis in oil drilling. It was cooled down pretty low. The die cost was $30K/unit. I don't know what the yield was; likely very low as per the cost. He was to get a unit as a gift. The plan was to build a camera for his telescope. His kids would love the high-tech toy.
mervynyan
Mervyn Yan
imokruok
Well-known
Thanks for posting that. Just saw it Engadget too. Not surprising at all. The build quality looks quite good and the sensor is APS-C - no reason it would be competitively priced with nice P&S cameras. Of course, if we're lucky, $1,000 is retail and retailers will immediately discount.
Last edited:
semilog
curmudgeonly optimist
tbarker13
shooter of stuff
I find it sort of interesting that folks on this forum hear "$1,000" as a price tag and conclude that it will sell like crazy.
But the folks at Engadget hear the same price tag and declare that it will be a niche product and that it is "doubtful that Fuji will sell oodles of these given the limitations and price."
I have to side with Engadget on this one. I love the look of this new camera (of course I'd like to see how it actually performs) and could even see myself buying one someday. But I don't see a lot of folks shelling out $1,000 for a zoomless camera.
But the folks at Engadget hear the same price tag and declare that it will be a niche product and that it is "doubtful that Fuji will sell oodles of these given the limitations and price."
I have to side with Engadget on this one. I love the look of this new camera (of course I'd like to see how it actually performs) and could even see myself buying one someday. But I don't see a lot of folks shelling out $1,000 for a zoomless camera.
zumbido
-
I didn't know that was the exclusive context. At any rate, you clearly misunderstood what I was getting at, so ah well.
If you interpreted that as "bashing" you very clearly misunderstood what I was saying.
In a forum that's never short on laughable blanket assertions, that one is right up there.
gilpen123
Gil
Where's the cue?
dazedgonebye
Veteran
Price being reported as $1000.
http://www.petapixel.com/2010/09/21/fujifilm-finepix-x100-to-cost-1000/
oops. i see i was beaten to the punch on this one.
http://www.petapixel.com/2010/09/21/fujifilm-finepix-x100-to-cost-1000/
oops. i see i was beaten to the punch on this one.
semilog
curmudgeonly optimist
Could be. I know less about CMOS than I do about CCD.
Contrary to what many think, CMOS sensors have (in important respects) not caught up with the better CCDs, which is why CCDs still mostly rule in scientific imaging. The biggest benefit of CMOS is that it is cheaper to fabricate and that you can put additional circuitry on the same chip.
Nevertheless, there is a limit to what's physically possible and the biggest gains appear to be happening on sensors with tiny (~2 um) pixels. I'd love to have a beer with your friends.
Last edited:
Paul T.
Veteran
Well, none of us know, do we, but it's a fact that micro fourthirds are massive sellers - 40% of DSLR sales in Japan, I've read, which is a huge percentage for such recent arrivals, up against Canon and Nikon.
And while a lot of users will opt for the zooms, every GF1 user I've bumped into (Ok, a sample base of four or five) has the 20mm. I paid around £650 for my GF1, with £50 cash back - in comparison, I reckon £800 for the X100 is a bargain, given I can focus with the VF and not the damn screen on the back, given I don't have to mess around with an add-on finder, and given picture quality is likely better, with the larger sensor.
Again, if the focus is lousy in use, like the X1, that's a different matter, but this is a camera that should press the right buttons (or rotate the correct dials) for a lot of photographers.
PKR
Veteran
I read a story about a guy using a prototype sensor on a view camera. He was traveling in the south west with this big rig. The photo of his gear led me to think he must have had an assistant, as one person couldn't have packed the gear alone. The battery pack, likely SLAs were in a good size box. Some how I got the idea the sensor package was cooled.. maybe not.
I don't know what became of this "high tech" photo project. I recall that Bill Gates was funding the thing.
willie_901
Veteran
Well I was really busy until word got out among my clients that I used auto focus. Since then things have been really slow. So, here I am.
Actually I have gigs today and tomorrow where I will use AF on my D700 to manually set the optimum focus point for interior shots.
tapesonthefloor
Well-known
This hybrid viewfinder makes many things possible. Let's imagine something together:
Engadget confirmed for us that the MF-ring is actually a MF-ring. This is excellent news. As most of us probably know, manual focus in a digital camera is near useless without a quick and easy way to confirm focus. dSLRs deal with this by having decent optical viewfinders and a focus dot. EVIL cameras deal with this by having a variety of software solutions, none of which I find satisfying, and most of which I find maddening. Is it the Olympus models that have the zooming focus method that only works with their new AF lenses? This, in my opinion, is terrible. If I'm ever going to be happy focusing manually with a digital camera, I have have have to have my eye up to an optical viewfinder. Maybe I'm already an old man at thirty, but this is simply a necessity for me.
With the X100 and some software, I can have a bit of everyone's cake, and ... eat... them? ...okay, ignore that part.
So, if I'm reading all this emergent data right, I have the choice between seeing only the EVF or seeing the OVF with some choice EVF bits merged. Since they're using a prism, they won't be able to project lowlights over the OVF, correct? the EVF is a light source, meaning only bright things (frame lines, data/numbers, etc) will be able to be superimposed on the OVF image. So, perhaps instead of the traditional rangefinder patch, when I'm in MF mode the EVF can project a zoomed-in, contrasty patch of the image into the center of my display so I can do a contrast focus on my own. Perhaps it will be similar to the new Pens in that the patch appears automatically when I touch the zoom ring. The difference will be, however, that instead of switching to a full zoom mode in the most jarring way possible (like the Pens), the camera will be able to project a bright patch onto the optical image. Augmented reality.
I can't picture exactly how this "contrast zoom" patch might look when the EVF can only project bright areas—it might look like a part of your composition is sitting in front of a bad blue-screen in the 70s—but it would still be enough to focus as quickly as one would with a proper RF once one was used to it.
I'm going to stop rambling now, but I have my fingers crossed that this triggers someone else's creativity. I really don't think we've even started to touch on what's possible here with this hybrid viewfinder and some creative firmware. This is probably the future.
Engadget confirmed for us that the MF-ring is actually a MF-ring. This is excellent news. As most of us probably know, manual focus in a digital camera is near useless without a quick and easy way to confirm focus. dSLRs deal with this by having decent optical viewfinders and a focus dot. EVIL cameras deal with this by having a variety of software solutions, none of which I find satisfying, and most of which I find maddening. Is it the Olympus models that have the zooming focus method that only works with their new AF lenses? This, in my opinion, is terrible. If I'm ever going to be happy focusing manually with a digital camera, I have have have to have my eye up to an optical viewfinder. Maybe I'm already an old man at thirty, but this is simply a necessity for me.
With the X100 and some software, I can have a bit of everyone's cake, and ... eat... them? ...okay, ignore that part.
So, if I'm reading all this emergent data right, I have the choice between seeing only the EVF or seeing the OVF with some choice EVF bits merged. Since they're using a prism, they won't be able to project lowlights over the OVF, correct? the EVF is a light source, meaning only bright things (frame lines, data/numbers, etc) will be able to be superimposed on the OVF image. So, perhaps instead of the traditional rangefinder patch, when I'm in MF mode the EVF can project a zoomed-in, contrasty patch of the image into the center of my display so I can do a contrast focus on my own. Perhaps it will be similar to the new Pens in that the patch appears automatically when I touch the zoom ring. The difference will be, however, that instead of switching to a full zoom mode in the most jarring way possible (like the Pens), the camera will be able to project a bright patch onto the optical image. Augmented reality.
I can't picture exactly how this "contrast zoom" patch might look when the EVF can only project bright areas—it might look like a part of your composition is sitting in front of a bad blue-screen in the 70s—but it would still be enough to focus as quickly as one would with a proper RF once one was used to it.
I'm going to stop rambling now, but I have my fingers crossed that this triggers someone else's creativity. I really don't think we've even started to touch on what's possible here with this hybrid viewfinder and some creative firmware. This is probably the future.