The Photographic Apocalypse is Nigh

[KevinBarrett]

Just found this on DPReview: Apparently Sony has introduced an image sensor for cell phones that out-resolves our precious a700! My headline is sensational, yes, so feel free to comment and leave your own impressions.

http://www.dpreview.com/news/0811/08111 ... mphone.asp

[Javelin]

I always beleived thaty cell phones will take over the P&S market Probably why Sony is so hot to establish themselves in the DSLR area. Cell phones will soon come with mini projectors too which will improve the small display problem, I expect them to take over the laptop market too eventually.

[Winston]

Pico Projection...they have been working on this for some time.

[Dr. Harout]

How about the lens? ... Or should we attach one of our lenses to the cell phone?
I am always with Megapixels. Some might argue or stand against, insisting that a 12-14 MP is way better than 24 MP. Someday, and in the very near future 46-48 MP on FF DSLR will become reality and with excellent IQ (and very very low noise = negligible). Meanwhile I'm enjoying my A700

[jcoffin]

There are some pretty basic laws of physics that get in the way of getting drastically higher resolutions from smaller sensors. Red light has a wavelength of about 750 nanometers or .75 microns. To capture light of this wavelength, you need a sensor with an active area about the same size or larger -- and that will remain true regardless of improvements in signal to noise ratio.

That does translate into a fairly big number (e.g. a 4x5 millimeter sensor with ~30 megapixel resolution), but that number is a bit like the speed of light: something we can never exceed (short of finding new laws of physics that show things working a lot differently than we currently believe possible) and realistically, it's not very likely that we'll get very close to it anytime soon either.

[pakodominguez]

How about the lens? ... Or should we attach one of our lenses to the cell phone?

Sony will sell the whole module (sensor plus lens and probably processor too)
dam: I just got the G phone with just 3MP...

[KevinBarrett]

There are some pretty basic laws of physics that get in the way of getting drastically higher resolutions from smaller sensors. Red light has a wavelength of about 750 nanometers or .75 microns. To capture light of this wavelength, you need a sensor with an active area about the same size or larger -- and that will remain true regardless of improvements in signal to noise ratio.

That does translate into a fairly big number (e.g. a 4x5 millimeter sensor with ~30 megapixel resolution), but that number is a bit like the speed of light: something we can never exceed (short of finding new laws of physics that show things working a lot differently than we currently believe possible) and realistically, it's not very likely that we'll get very close to it anytime soon either.

Just an amusing figure based on this data: a full-frame 35mm camera could reach nearly 1.3 gigapixels.

[Greg Beetham]

Now if it just came with a popup flash for WL studio flash and maybe a battery grip w-tripod thread, oh and a CF slot would be good for storing the ex-jine JPG and RAW images @ 5 frames/sec along with a set of interchangable lenses, not to mention the P A S M mode dial for controlling the DOF and esposure just the way you want and...gee almost forgot the selectable focus points, but then, who needs focus points when you have unlimited DOF, just point it out the window when you wake up in the morning and take a photo of everything, then you can do a "Blade Runner" ie. zoom, track right, enhance, track left, zoom, enhance, and find stuff that wan't even visible in the original image ....
Greg

[Dr. Harout]

... and use it underwater too (up to 50m depth).

[KevinBarrett]

Love that Blade Runner zoom!

While I would never feel pixel-envy for a cell-phone (can't show up with a smart-phone to shoot a wedding, can I ?), I was just flabbergasted by this figure.

[jcoffin]

Just an amusing figure based on this data: a full-frame 35mm camera could reach nearly 1.3 gigapixels.

Right -- just keep in mind that this is only looking at the sensor side. For that to do any good, you'd have to have a lens with a circle of confusion of approximately .8 microns or so. The usual definition for the diffraction limit is: d = 1.22 LN (the 'L' should really be a lambda), where 'L" is the wavelength of the light, N is the f/number, and 'd' is the size of the airy disk produced (i.e. the smallest size of a focused spot). In this case, we want to rearrange that to: N = D/(1.22L). We're using, L = .750 microns and D = .8 microns, so for the lens to produce an image of that resolution, we'd have N=0.8/(1.22*0.75), or a smallest usable aperture of about 0.87. Keep in mind, however, that this would be roughly equivalent to something like f/5.6 or f/8 on a typical current camera/lens combination -- the point at which diffraction, rather than the lens, limits the resolution that can be achieved.

Then again, with sensor wells this small, you could only take a decent picture at ISO 0.5, so a lens that fast would be barely enough to let you avoid a tripod in broad daylight...

[Greg Beetham]

Just an amusing figure based on this data: a full-frame 35mm camera could reach nearly 1.3 gigapixels.

Right -- just keep in mind that this is only looking at the sensor side. For that to do any good, you'd have to have a lens with a circle of confusion of approximately .8 microns or so. The usual definition for the diffraction limit is: d = 1.22 LN (the 'L' should really be a lambda), where 'L" is the wavelength of the light, N is the f/number, and 'd' is the size of the airy disk produced (i.e. the smallest size of a focused spot). In this case, we want to rearrange that to: N = D/(1.22L). We're using, L = .750 microns and D = .8 microns, so for the lens to produce an image of that resolution, we'd have N=0.8/(1.22*0.75), or a smallest usable aperture of about 0.87. Keep in mind, however, that this would be roughly equivalent to something like f/5.6 or f/8 on a typical current camera/lens combination -- the point at which diffraction, rather than the lens, limits the resolution that can be achieved.

Then again, with sensor wells this small, you could only take a decent picture at ISO 0.5, so a lens that fast would be barely enough to let you avoid a tripod in broad daylight...

Ha Ha...but an aperture of 0.87 what? cm, millimeters, microns...if it's microns that's a mighty small aperture, probably work ok if the sensor had only one pixel....what a hoot
btw isn't the wavelength of light measured in nm? oh I've got it .750micron = 750nm
Greg

[Winston]

I believe his aperture (0.87) is a dimensionless f-stop (f/.87).

[jcoffin]

Ha Ha...but an aperture of 0.87 what? cm, millimeters, microns...if it's microns that's a mighty small aperture, probably work ok if the sensor had only one pixel....what a hoot
btw isn't the wavelength of light measured in nm? oh I've got it .750micron = 750nm
Greg

The aperture is an f/number, just like usual.

People who work with such things on a regular basis seem to measure wavelengths in angstroms most of the time. You've done the conversion correctly (at least I hope so -- it's the same conversion I was using). Of course, I could have gone the other direction, leaving the wavelength in nanometers, and instead converted the pixel well size to nanometers instead. Any unit will work, as long as you're consistent throughout.

I guess if the sensor only had a single pixel, an aperture of .87 microns wouldn't be too bad, but if you wanted to be able to resolve between even two pixels, it would be pretty worthless.