Well it looks like I will be keeping my 1DsIIIs a while longer. Combine that with no IS on the new 24-70 and no new 35mm f/1.4 and I am starting to think that Canon just doesn't want any more of my money!
S Dilworth wrote:
William Claff has crunched the numbers for many Nikons SLRs here. Note how close the D3S and D4 curves are to the ideal FX line, which represents the maximum theoretical performance for that sensor size. At very high ISOs there is no longer room for big leaps.
His chart tells us there will never in the future of the universe be a "full-frame" camera that performs above ISO 25600 better than the D3S performs at ISO 12800: the noise inherent in light itself prevents that possibility.
At the moment, Canon has a slight high-ISO read-noise advantage, Nikon has a slight quantum-efficiency advantage, and Sony has a big low-ISO read-noise advantage (all based on their latest sensors). Beyond that, they perform more or less the same.
It's exceedingly unlikely that the EOS-1D X would be significantly better than the D3S or D4 at high ISOs. And it's not even theoretically possible for it to be one stop better. You'll just have to somehow muddle along with that level of performance. I know you can manage it!
Future advances will lie in increasing the pixel count, improving low-ISO performance, and fixing the many ways in which our cameras are poorly designed for the tasks they are expected to perform.
None of this has much to do with making photos that people want to look at....Show more →
So if D4/D3s are near to perfect, how it comes that 5D2 is nearly as good at much higher pixel rates.
I am sorry. 35 years ago my physics teacher told me "There is nothing faster than light. Never in the future of universe that will be. It is just impossible! Einstein proofed it."
When I asked her if SHE could explain her answer was: "No. I never understood it myself!"
20 years later physicians "createt" new elementar pieces, moving faster then light, to keep their own theoretical system NOT break together. If the noise inherent in light itself prevents that possibility, we have to find a way to ignore that noise inherent. I do not care, if sensors can be better, or noise is calculated away mathematical as long detail is not lost (to much). The past has proofed again and again, that in this univers EVERYTHING is possible that can be imagined.
S Dilworth wrote:
William Claff has crunched the numbers for many Nikons SLRs here. Note how close the D3S and D4 curves are to the ideal FX line, which represents the maximum theoretical performance for that sensor size. At very high ISOs there is no longer room for big leaps.
His chart tells us there will never in the future of the universe be a "full-frame" camera that performs above ISO 25600 better than the D3S performs at ISO 12800: the noise inherent in light itself prevents that possibility.
At the moment, Canon has a slight high-ISO read-noise advantage, Nikon has a slight quantum-efficiency advantage, and Sony has a big low-ISO read-noise advantage (all based on their latest sensors). Beyond that, they perform more or less the same.
It's exceedingly unlikely that the EOS-1D X would be significantly better than the D3S or D4 at high ISOs. And it's not even theoretically possible for it to be one stop better. You'll just have to somehow muddle along with that level of performance. I know you can manage it!
Future advances will lie in increasing the pixel count, improving low-ISO performance, and fixing the many ways in which our cameras are poorly designed for the tasks they are expected to perform.
None of this has much to do with making photos that people want to look at....Show more →
Fortunately this is untrue.
For Bayer sensors there is not that much room for improvement. For any technology that no longer needs light absorbing CFAs before their photosite equivalent, there is a long way to go.
Ralph Conway wrote:
So if D4/D3s are near to perfect, how it comes that 5D2 is nearly as good at much higher pixel rates.
I am sorry. 35 years ago my physics teacher told me "There is nothing faster than light. Never in the future of universe that will be. It is just impossible! Einstein proofed it."
When I asked her if SHE could explain her answer was: "No. I never understood it myself!"
20 years later physicians "createt" new elementar pieces, moving faster then light, to keep their own theoretical system NOT break together. If the noise inherent in light itself prevents that possibility, we have to find a way to ignore that noise inherent. I do not care, if sensors can be better, or noise is calculated away mathematical as long detail is not lost (to much). The past has proofed again and again, that in this univers EVERYTHING is possible that can be imagined.
Thank you. I did. No, I tried. I am not intuitiv enough, I guess.
Here is an other thought of mine. When I started in photography I bougt an Tmax 3200 and asked for less grain. I was told: "That is impossible because of the nature of film. It is a problem of photography, that will never be served." This is just 20 years ago. Digital photography was in the beginning at video resolution and we safed 3-8 pictures in camera. Today DSLRs give us 4-6 times the resolution of 35mm film at ISO 3.200 with less noise than ISO 400 was in film times. And we safe 1000 images on a CF card smaller than a lens cap.
My first harddrive bought in 1989 was about 750 € for 20MB. 20 years later I bought my last one. 2000000 MB (2 Terrabyte) at € 105.
I do not fear there is no way to reduce noise and enhance IQ at highest ISO.
For Bayer sensors there is not that much room for improvement. For any technology that no longer needs light absorbing CFAs before their photosite equivalent, there is a long way to go.
"A long way" means another 2 stops at best, for red and blue light. Another 1 stop for green light, for a RGGB layout.
alundeb wrote:
"A long way" means another 2 stops at best, for red and blue light. Another 1 stop for green light, for a RGGB layout.
Hey, 2 stops (and it would be 2 real stops, not handwaving "the processor does magic" stops, means 4 times the shutter speed for the same noise performance, or the change in usable ISO from a 1D2 to a 1D4. And reds less likely to blow out for evening and night event shots. I am not going to turn my nose up at that
thedigitalbean wrote:
He might be referring to the faster than light neutrino report from last year:
Except that those aren't new particles - we've known about neutrinos for a long while, and that they travel near c. The weird data from the experiment needs work done and some verification, but that wasn't "made up to be faster" that was "shit this thing we've seen before goes faster, this breaks a lot of our current theory".
As for tachyons - if the experiment turns out to be accurate then those neutrinos will be seen as a type of tachyon. Pretty exciting times. But if that's what he meant he has a horrible way of expressing things.
Ralph Conway wrote:
Resolution is 9MP more. You exchanged D3s with D3x. What can not compare with 5D II in IQ at ISO higher than 1600.
Ok, so if we swap D3s to D3x and compare, then we have a camera that still outperforms the 5D2 noisewise (though by a smaller margin), except that now it outresolves it too
They look a bit waxy canon in cam jpgs, the ISO1600 has DNR waxy look like an over-cooked blu-ray disc.
But Canon samples have been poor since post 1Ds2 release.
Ralph
