nicholas wong
Member
--
nws
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
nws
rick decker
Forum Pro
Nobody claims full-frame is "bad" here. What we have are two different groups. Traditionalists who place more emphasis on the past, the need to stay tied to the past and the value of the past, and tend to look backwards first or more then the second group. The progressivists who place less emphasis on tradition and the past than on the need to move forward and to question the past more closely than the others.
The images at PMA clearly showed that full-frame is not a necessity. Is it a nice-to--have? I don't know given the CA problems. It may be that full-frame will be seen as the Medium Format of Digital Photography in the not-too-distant future as technology advances.
More power to those who shoot full-frame as the will pave the way for improvements and advances. But it is not a priority for "yours truly".
Rick
http://www.lightreflection.com
http://www.silveroaksranch.com
http://www.pbase.com/rickdecker
The images at PMA clearly showed that full-frame is not a necessity. Is it a nice-to--have? I don't know given the CA problems. It may be that full-frame will be seen as the Medium Format of Digital Photography in the not-too-distant future as technology advances.
More power to those who shoot full-frame as the will pave the way for improvements and advances. But it is not a priority for "yours truly".
Rick
--
http://www.lightreflection.com
http://www.silveroaksranch.com
http://www.pbase.com/rickdecker
Michael Long
Senior Member
6MP/1.5x may be enough for some. I, for one, want all of the pixels I can get for high-quality landscape work.
At a certain point, due to diffraction, noise, or lens resolution (or a combination of the three), you'll be unable to effectively cram more pixels into an APS-sized form factor. At that point, your only option is to increase the size of the sensors real estate, either to FF 35mm, or to 6x4.5 or better.
And there is NO MF system with the wide lenses, long lenses, and handling characteristics of 35mm.
I also believe that FF prices will migrate down as manufacturing processes improve...
Captain Obvious
Forum Enthusiast
If the detail doesn't line up perfectly with the sensor, you need more than 2 samples.
FF is doing fine and the truth is that it provides all the advantages you can think of. Simply the bigger the better, just like with the film 35mm better than APS, 6x45 better than 35mm, 6x6 or 6x7 better than 6x45 and so on. Bigger chip means better dynamic range and less noise, so the ultimate format for DSLR is actually FF. Don't even compare the 4/3 or 1.5x crop factor of NIkon to the cuality of output that FF gives. Even as technology improves the FF sensor WILL ALWAYS be better quality wise.
...timeless, ordered, simply stated, or even implied...
BUHAHAHAA listen to what you're talking about....Canon lost track...that's funny becouse it seems like Canon is putting Nikon back into stone age. Nikon hasn't introduce a thing, and by the time they did it's gonna be too late for them as people are already switching gears. And the new DX line of lenses seems like a joke for a people who spend $8k + on their lenses, too dim anyway for what they're intended for.
--
...timeless, ordered, simply stated, or even implied...
rick decker
Forum Pro
Get the Sigma SD9..you will need all the sharpness you can get.
Check these pictures out..more sharpness than any Canon or Nikon shots I have seen:
http://www.pbase.com/sphomphanh/sd9_southwest
--
http://www.lightreflection.com
http://www.silveroaksranch.com
http://www.pbase.com/rickdecker
Check these pictures out..more sharpness than any Canon or Nikon shots I have seen:
http://www.pbase.com/sphomphanh/sd9_southwest
--
http://www.lightreflection.com
http://www.silveroaksranch.com
http://www.pbase.com/rickdecker
Gherry D Bender
Member
I guess we differ as to if you can really do a linear extrapolation on an mtf based upon two points, but at any rate I still believe that a 50% mtf at 80 lp/mm would be a very rare lens indeed. I looked at the mtf's at photodo and what really struck me was how poor zoom lens are. Anyway a good source that I found for understanding mtf is at
http://www.normankoren.com/Tutorials/MTF.html
He lets you be as mathmatically inclined as you wish.
http://www.normankoren.com/Tutorials/MTF.html
He lets you be as mathmatically inclined as you wish.
FF is going the way of LF and MF went: reserved for few. The main stream is de facto APS.
baruth
--
baruth
Michael Long
Senior Member
Not bad, but call me back when they get a few more pixels. And a better body. And better lenses.
Til then I'll stick with my 1Ds...
You must be referring to the profitable segment of the photography market.
I wasn't aware that "Mainstream" was something to aspire to.
Larry
Well, it depends. If, at a certain resolution, a new aberration becomes resolved, you cannot.
On the other hand, for a diffraction limited (i.e. perfect) optics you can calculate analytically the MTF curve. And at Fstop=8, it's close to what is observed for best lenses. So, take lenses like
Contax: Planar T* 85/1,4, Planar T* 50/1,7
Canon: EF 200/1,8L USM , EF 50/2,5 Macro
Pentax: SMC-F 50/2,8 macro, SMC-F 50/1,4
LeicaM: Summicron-M 50/2,0
etc. At Fstop=8, the highest point of each MTF curve is more or
less the same: 10 lp/mm -> 95%; 20 lp/mm -> 90% 40 lp/mm 75%.
For 80 lp/mm, the respective value would be ca 50%.
If a lense is close to the theoretical limit at 40 lp/mm, it shows
that abberrations are of much smaller geometrical size than the line width. Consequently, at 80 lp/mm we are also close to the theoretical limit (though, maybe not as close as at 40 lp/mm).
Probably very few will have it throughout the field.
Nice tutorial. However, somewhat confusing is that black and white stripes with sharp boundaries are used for illustrations. Actually, sinusoidal profile (density is a sinusoidal function of coordinate) is used for MTF-s.
Larry,
I am referring to digital camera market. There is only one serious FF camera on the market 1Ds. 14n is not. No other maker is considering it. On the other side, the 1.5x sensor is universal, including Canon.
People admire 1Ds as they admire Hassie. But how many of them really even consider buying 1Ds or Hassie, inclusing the pros?
baruth
I am referring to digital camera market. There is only one serious FF camera on the market 1Ds. 14n is not. No other maker is considering it. On the other side, the 1.5x sensor is universal, including Canon.
People admire 1Ds as they admire Hassie. But how many of them really even consider buying 1Ds or Hassie, inclusing the pros?
--
baruth
Peter Phan
Senior Member
Well, obviously there's a market large enough to support Hassie, 'Blad, Rollei, Mamiya, Contax 645, Pentax 645, Pentax 67, etc. And you can bet that many of these users are looking seriously at 1Ds. So I don't understand where people get off saying that 1Ds is admired but not bought, that the 1Ds doesn't sell, etc. Not only does 1Ds tap into the 35mm market, but it taps heavily into the medium format market, too.
Peter Phan
Senior Member
My mistake for mentioning Hassie twice (Hassie, 'Blad). Should have said Bronica.
Michael Long
Senior Member
Funny, I count Pentax, Contax, Canon, and Kodak as those who've "considered" it. And you should probably ask Kodak if they think they're "serious" or not.
They simply used a less expensive body in an attempt to make their camera more accessable.
But then again, I'm also glad we have someone such as yourself inside Nikon, who "knows" they're not considering it either.
If you'll look around, you'll find its quite a few.
I think you're disgruntled because a 1Ds system was priced out of your and what you perceive to be everyone else's price range, and I think you're making the fatal flaw of assuming FF technology is ALWAYS going to be so.
I myself have no doubt that the technology will get cheaper, and will percolate downward as it does. For people on a "Rebel" budget, however, APS sensors will always be cheaper...
Meryl Arbing
Senior Member
I would have thought that a line/pair would be equivalent to two lines of pixels since you would need to have at least two pixels (to distinguish a point of detail. It would be obvious that detail smaller than a single pixel could not be imaged and detail the same size as a single pixel only needs on pixels to register it but it is the second pixel that provides the necessary contrast.
To 2.5 pixels per resolved line pair would seriously degrade the expected resolving potential of the snesor.
At 2 pixels/line pair, for example a 6 Mp Canon 10d (3072 x 2048 pixel) image would have 1536 x 1024 line pairs of pixels and with a 22.7 mm x 15.1 mm sensor (for the 4/3 cameras) that would give an absolute resolution of not more than 67 lpmm. Reolution would go down from there because of lens limitations..camera shake...etc.
The same calculations done at 2.5 pixels/line pair would give even worse results (@ 54 lpmm) which is extraordinarily bad.
Yet, there are people who are more than satisfied with the results from thier consumer level digital cameras and even claim parity (if not superiority) over film cameras.
How do you reconcile the figures which don't seem to make much sense compared with experience?
To 2.5 pixels per resolved line pair would seriously degrade the expected resolving potential of the snesor.
At 2 pixels/line pair, for example a 6 Mp Canon 10d (3072 x 2048 pixel) image would have 1536 x 1024 line pairs of pixels and with a 22.7 mm x 15.1 mm sensor (for the 4/3 cameras) that would give an absolute resolution of not more than 67 lpmm. Reolution would go down from there because of lens limitations..camera shake...etc.
The same calculations done at 2.5 pixels/line pair would give even worse results (@ 54 lpmm) which is extraordinarily bad.
Yet, there are people who are more than satisfied with the results from thier consumer level digital cameras and even claim parity (if not superiority) over film cameras.
How do you reconcile the figures which don't seem to make much sense compared with experience?
Universal? Maybe in your universe.
Last time I checked Canon had a 1.6, 1.3 and 1.0, and Kodak will shortly have 1.0.
There's nothing magic about 1.5, it came about from a convergence of cost issues and technology limitations. Now it seems to gained some cult following as the "blessed " format.
Lots of people have and will continue to buy Medium Format and Large Format because that's what the job requires. Everyone I know shooting with these systems either has a digitial solution in place or is seriously considering getting one this year.
Those of us who have lots of money invested in lenses won't mind spending a few grand to take full advantage of them. If Nikon doesn't want our money, somebody else will. If Nikon had a 1Ds equivalent I'd go buy one today because it would pay for itself in a month.
Larry
W
Walter Freiberger
Guest
In theory, Yes. In the real world, No. 4/3 can achieve up to 15mp resolution at reasonable noise levels. APS can achieve up to 30mp. And FF up to 60mp. Who needs 60mp resolution? Have Canon's FF lenses enough resolving power? Don't think so.
Chip sizes will have to be reduced in the next few years....
Walter
Michael Long
Senior Member
How did you arrive at -- or make up -- these numbers?
Michael Long
Senior Member
...of aliasing. Very rarely do you have perfect registration of the pattern the lens is projecting with the sensor mosiac. As such one "black" line may span two rows to a greater or lesser degree. The result of which could well be two gray lines, one maybe darker than another.
Picture a piece of graph paper. Now run a line across it at random with a marker. How many squares contained the line perfectly? How many sets of squares did the line overlap to some degree?
Pick the number of line pairs you want to resolve and call it N. To get there you're going to need N*3 or even N*4 pixels in the real world.
Which is also, BTW, why you needs sensors with greater resolving power than the lens can provide to get the maximum amount of detail.
Picture a piece of graph paper. Now run a line across it at random with a marker. How many squares contained the line perfectly? How many sets of squares did the line overlap to some degree?
Pick the number of line pairs you want to resolve and call it N. To get there you're going to need N*3 or even N*4 pixels in the real world.
Which is also, BTW, why you needs sensors with greater resolving power than the lens can provide to get the maximum amount of detail.
Similar threads
Keyboard shortcuts
- f
- Forum