Frankie
Speaking Frankly
eddie1960
Established
HAHA
I was thinking i'd love to have an b/w ir converted version (forget the colour ir)
i'm frequently called a curmudgeon though even by my freinds who frequently are older than me
Frankie
Speaking Frankly
Thank you.
Gazzah
RF newbie
PKR - I understand the increase in rez - but I still cant follow the arguments for increased DR when removing the bayer filter - or maybe I miss understood completly as there was talk of pixel binning - which would increase the DR but decrease the rez again.....
Guess I need to go back to high school physics.. (and english !)
Gary H
Guess I need to go back to high school physics.. (and english !)
Gary H
semilog
curmudgeonly optimist
Last edited:
morgan
Well-known
I agree. This isn't a camera that those people would most likely see as an alternative. Maybe more the ricoh/sigma crowd. I'm personally extremely excited about this camera and have already started socking away money. It's the viewfinder that makes it for me.
Gazzah
RF newbie
Semilog - Off topic but -
I took up surfing in my mid 40s, a freind used to wear a T-shirt over his wet suit that read
"DONT TRUCK WITH THE CURMUDGEON!"
He is 20 years older than me and called me "Pup"... I think Im catching up now though!
I took up surfing in my mid 40s, a freind used to wear a T-shirt over his wet suit that read
"DONT TRUCK WITH THE CURMUDGEON!"
He is 20 years older than me and called me "Pup"... I think Im catching up now though!
Frankie
Speaking Frankly
Exactly.
In the digital age, areas are better measured by square-pixels as opposed to square-foot or square-mm...assuming each pixel is large enough to collect sufficient photons without [noisy] boosts.
Pixels size had evolved from 12u (Dalsa) to 9u, to 7.2u, to 6.8u, to 6u, and now 5.2u...and will get smaller yet.
As I had opined often, a truce will occur when FF is reached...or new [larger image circle] lens development will have to happen. No M-mount lens collector wants to hear that.
If you want hardcopy, then the pixel-for-dpi ratio must be good enough...thus the R-D1 2000 x 3000 sensor is good enough for an ~8 x 10 if a 300dpi printer is used.
The 12Mp APS-C sensor used in the X100 would have to be ~3000 x 4000 @ 6u, and a good 8 x 10 or even an 11 x 14 is possible.
That kind of pixel count is also more than good enough for display on a 1080p (1920 x 1080) LCD/Plasma penal. I suspect that is why the camera has both mini-USB and HDMI ports built-in [under the right-side hatch].
My home TV has long been my revolving gallery.
Frankie
Speaking Frankly
JoeV
Thin Air, Bright Sun
Regarding the price, if you compare the $1000 guesstimate against a GF-1 or EP-2 with the Lumix 20-f/1.7 lens attached, you're coming pretty close. I think this camera, if the current design and specs hold out (they are, after all, not finalized) will fly off the shelves.
Regarding speculation of FF sensors coming down in price, I don't think that'll happen any time soon, and it's not a marketing decision, but rather it's about the economics of semiconductor manufacture:
1) Logic transistor size and density continues to shrink, making chips smaller in size, faster, generating less heat, and able to make more per wafer.
2) Chip cost is mainly wafer-based. So there's an optimal balance between die size, selling price, and maximizing profits. A thousand small chips per wafer, selling cheaply, may or may not be as profitable as 300 larger chips per wafer that can be sold for more money.
3) Full-Frame chips, by definition, can't be shrunk in size, since the format specifies a fixed size. Therefore the yield per wafer is fixed, hence revenue per wafer is fixed. They can shrink transistor density in FF chips, thus selling the newer ones for more money because they have more megapixels; that's their only path to increased revenue for FF chips.
Constrast these with the economics of computer chips. You don't know or care how big or small your processor's die size is. But you can bet that each new process technology enables smaller die sizes, yielding more die per wafer, that operate faster (shorter path lengths, etc.), yielding more profit per wafer.
~Joe
Regarding speculation of FF sensors coming down in price, I don't think that'll happen any time soon, and it's not a marketing decision, but rather it's about the economics of semiconductor manufacture:
1) Logic transistor size and density continues to shrink, making chips smaller in size, faster, generating less heat, and able to make more per wafer.
2) Chip cost is mainly wafer-based. So there's an optimal balance between die size, selling price, and maximizing profits. A thousand small chips per wafer, selling cheaply, may or may not be as profitable as 300 larger chips per wafer that can be sold for more money.
3) Full-Frame chips, by definition, can't be shrunk in size, since the format specifies a fixed size. Therefore the yield per wafer is fixed, hence revenue per wafer is fixed. They can shrink transistor density in FF chips, thus selling the newer ones for more money because they have more megapixels; that's their only path to increased revenue for FF chips.
Constrast these with the economics of computer chips. You don't know or care how big or small your processor's die size is. But you can bet that each new process technology enables smaller die sizes, yielding more die per wafer, that operate faster (shorter path lengths, etc.), yielding more profit per wafer.
~Joe
semilog
curmudgeonly optimist
In theory, I agree. But as Yogi Berra said, In theory, theory and practice are the same. In practice, they aren't.
A few things that modify this simple calculus:
• To a first approximation, no one uses a tripod and good technique to reduce vibration to negligible levels. The spatial frequencies that are actually captured in handheld photography seldom exceed 1000 lp/image height. More pixels won't help.
• All but the very best (read: expensive) lenses are already Nyquist oversampled on a 12 megapixel sensor. Again, more pixels won't help.
Sure, some pros and crazy amateurs use tripods and have immaculate technique, and they are of course overrepresented in forums like this. But the vast majority of people don't own a tripod, and the vast majority of people who own tripods seldom or never use them.
• Additionally, as print sizes increase, so (usually) does viewing distance, and the 300 dpi requirement can, except for very high-quality (exhibition) prints, be relaxed.
For most people, including a majority of "amateur photographers" and a pretty large number of pros, the already available APS-C sensors provide more IQ than needed. As these sensors are continuing to improve, the need for FF equipment is, for most people, less and less pressing.
I acknowledge that there are a not-insignificant number of people who can benefit from FF, and even more so that there are many gearheads who just like cool stuff.
But I also think that if you put a 7D or a D7000 or even an E-PL1 in the hands of most people, they will already get far better IQ than they will be able to use, and they'll be able to make bigger prints than you think they will. If these prints aren't good enough, it generally won't be because of sensor limitations, but because of camera movement.
• I haven't even mentioned focus accuracy. One of the tremendous benefits of smaller sensors (APS-C, 4/3) is that DOF is increased, relaxing the requirement for perfect focus accuracy.
Last edited:
PKR
Veteran
frank, i.e. [id est] = frank, that is = frankie
Assorted marketing people in my field often worried about what I might say. Often, just one sentence killed their whole spiel. AND, I am not politically correct.
I got it the first time.. I was just..jesting. And being succinct is a good thing. There is less chance of being misunderstood.
And, knowing they are necessary, I'm not especially fond of "marketing people".
tonyjuliano
Wooden Indian
Frankie
Speaking Frankly
The Bayer pattern assigns every other pixel to a different colour...RGBGRG...and 50% of the area [sensitivity] to green. Thus resolution for any colour is only half spatially.
In our industry (aerial imaging) all real cameras [starts at $1 million++] are panchromatic [CCD without Bayer]...ancillary smaller IR/R/G/B cameras cluster around the main panchromatic camera are used to collect data for colourizing the imagery...the camera firmware files colour data pixel-for-pixel. The end-user can select bands in any combination...RGB for natural colour, G/R/IR for infra-red imagery [mostly used in forestry...deceased trees show up red/brown]
The latest is a Zeiss RMK-DX, using a monolithic 11,200 x 11,700 7u Dalsa sensor...imagine 131Mp in one shot.
Image resolution is defined as Ground Sampling Distance (GSD)...pixel footprints. The higher you fly, the larger the GSD. If Bayer is used [every other pixel spatially] then 2X GSD...mine must be smaller than yours to compete.
Last edited:
dazedgonebye
Veteran
j j
Well-known
They already did so in previous cameras.
PKR
Veteran
Well if you allow for a given pixel population and Bayer filtering using RGGB (red-green-green-blue) for the make up of a file. The actual pixel sites needed to produce and image require a RGB to deliver. So if we have 3 pixels necessary; this makes the actual sensor population = to divide by 3. So, your 12MP camera is really delivering 12/3= 4MP usable for color file creation. The marketing folks, got a hold of the pixel race early in digital camera marketing, and established the "standard" of using the larger number.
So, if you don't need RGGB (and the extra green is used because our vision is more green sensitive) to create a color file, this frees up more pixels for a B+W file.
Phase One has the first b+w sensor in common use..
http://www.phaseone.com/
Any increase in DR can be made by allowing some of the pixel population to be used for upper bandwidth blowout. Fuji was a pioneer in this field with the S pixel - R pixel arrangement in the S3, S5 Pro cameras.
There is going to be a trade-off in pixel use. But overall, with a B+W sensor you could, for a given pixel population, have improved resolution and bandwidth over the same population of pixels used to produce a color image.
There may be some errors in the above. corrections are welcome.
Last edited:
Frankie
Speaking Frankly
Your points are all valid. I didn't mention tripods, etc., etc....to avoid off-my-topic issues.
Bottom line is, indeed, the human eye...said to be able to resolve 1 arc-minute or ~0.3mm/m...thus further the viewing distance, larger the pixel size tolerated.
I couldn't see pixels in my 1680 x 1050 @ 0.27mm dot pitch computer screen at 0.75m, just as I couldn't see pixels on my 1920 x 1080 plasma @ 0.6mm dot pitch sitting 3m away.
One will have to have super visual acuity and burying the nose into an 8 x 10 to see dots...hence most printers made are only 300dpi.
Frankie
Speaking Frankly
To summarize...
Hmmm... 12mp aught to be just about enough.
Yeah! Mine has to be bigger than yours!
AND, I have fondled 131Mp cameras and held court with Zeiss engineers...although I have never met Herr Doktor Kaufmann.
Gazzah
RF newbie
PKR - I know the theory of the bayer filter, and pixel binning, and even the S/R senors (I have an s3 Pro) - I just dont grok it.....
Guess they didnt cover things like that in science class back in the early 70s - it is also part of the reason I back using film agan, I can understand it ( sort of... ;( )
Gary H
Guess they didnt cover things like that in science class back in the early 70s - it is also part of the reason I back using film agan, I can understand it ( sort of... ;( )
Gary H