One wants to find out an estimation of which aperture theoretically guarantees the best possible resolution?
--
regards
Janne Mankila, Finland
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
Which lenses will resolve 36mp? Which will struggle?
- Thread starter dholl
- Start date
--
regards
Janne Mankila, Finland
Robin Casady
Forum Pro
Easy, it gives a photographer an idea at what f/stop the image will start to suffer from diffraction effects.
--
Robin Casady
http://www.robincasady.com/Photo/index.html
CriticalI
Senior Member
Perhaps you should just listen to what some people are actually saying instead of repeating the same tedious mantra in response to what you THINK they are saying.
I quite understand what difference having more MP makes, and quite how irrelevant that is when the lens you own is simply not that good . The fact that it may be slightly better than awful does not suddenly make it good and certainly not capable of upping a print size.
If you are going to spend $3000 buy a 36MP camera and not exploit it by using glass that can maximise it's potential you are just wasting a bunch of money. You would be far better off buying an OMD with a nice prime.
Similarly if you never print that large, than 12MP is enough for pleasing and even stunning images without all the overhead of large files and at smaller print sizes, lens quality is less important as well. What will 36MP give you that actually matters if you print no larger than 12X8?
Putting cheap glass on a D800 is like putting a V8 in a Corolla. Faster, but why bother?
Regards,
Steve
I quite understand what difference having more MP makes, and quite how irrelevant that is when the lens you own is simply not that good . The fact that it may be slightly better than awful does not suddenly make it good and certainly not capable of upping a print size.
If you are going to spend $3000 buy a 36MP camera and not exploit it by using glass that can maximise it's potential you are just wasting a bunch of money. You would be far better off buying an OMD with a nice prime.
Similarly if you never print that large, than 12MP is enough for pleasing and even stunning images without all the overhead of large files and at smaller print sizes, lens quality is less important as well. What will 36MP give you that actually matters if you print no larger than 12X8?
Putting cheap glass on a D800 is like putting a V8 in a Corolla. Faster, but why bother?
--
Regards,
Steve
bobn2
Forum Pro
You haven't followed this at all. No, it doesn't do that. Lets have a look at my lens graphs again:
McHugh's calculator says that the D3 starts to 'suffer diffraction effects' at f/16, and the D3X at f/13.
Now look at the shape of the curves. can you see any special feature on the D3 curve at f/16 or the D3X curve at f/13 that would show the suffering of 'diffraction effects'.
The truth is this, they suffer 'diffraction effects' at all apertures, that is physics. Diffraction effects become the dominant determinant of resolution after f/5.6, for both cameras (when the line begins to fall). There is nothing on the graph to show the presence of McHugh's 'diffraction limit'. It does not exist, so the calculator gives no useful information. All it does is confuse people, and you are one of the people it has confused.
--
Bob
Mikael Risedal
Senior Member
ajajaj a mi madre ciega que parece agudo en el campo
buy a D800E and use cheap lenses and you avoid moire.
buy a D800E and use cheap lenses and you avoid moire.
But that is not what the calculator spits out. And it cannot spit it out because theoretically, ie, without any optical aberrations, the highest possible resolution is always as wide as possible. What it calculates is an f-stop where, in an optically ideal lens, diffraction reduces the resolution by some fixed amount (the precise value of this amount is dependent on the conditions used for the calculation).
That of course, is of some value but in reality optical aberrations are there and therefore the optimal aperture is different for different lenses. And thus this f-stop calculated by the diffraction calculator, which is calculated for an ideal lens, can be close to the optimal f-stop or it can be further away. And the resolution difference between the optimal f-stop and this 'diffraction limit' is also very much lens dependent.
ejmartin
Veteran Member
No, what I meant (but didn't state clearly enough) is where MTF drops to a fixed value, eg 50% or 10% of input contrast. Not from the peak, which is already degraded by lens imperfections, diffraction, etc.
--
emil
--
http://theory.uchicago.edu/~ejm/pix/20d/
To illustrate this point further, assume we had an even higher resolution sensor. We have no reason to believe that the resolution vs. f-stop curve will have a materially different shape, it will be just stretched further up at all f-stops, proportionally more around its optimal f-stop but the peak position will barely shift at all. But the calculated f-stop at which this lens purportedly becomes diffraction limit will just shift with sensor resolution ever further to the left until at some point it is left of the optimal aperture. Which would put as into the strange situation that stepping down beyond this diffraction limit actually increases resolution.
McHugh's calculator says that the D3 starts to 'suffer diffraction effects' at f/16, and the D3X at f/13.
Now look at the shape of the curves. can you see any special feature on the D3 curve at f/16 or the D3X curve at f/13 that would show the suffering of 'diffraction effects'.
The truth is this, they suffer 'diffraction effects' at all apertures, that is physics. Diffraction effects become the dominant determinant of resolution after f/5.6, for both cameras (when the line begins to fall). There is nothing on the graph to show the presence of McHugh's 'diffraction limit'. It does not exist, so the calculator gives no useful information. All it does is confuse people, and you are one of the people it has confused.
Actually, our opinion on IQ matters precious little. It is this particular fact that lead us to point & shoot cameras with wild MP numbers quite early on. The sensor technology wasn't ready and still the marketing department demanded for more megapixels so they could sell the cameras. The megapixel came to a peace in the P&S market and the marketing department came up with a way to turn it into a positive - they're clever chaps and gals, they are.Of course it matters. We are the users, if our opinion on IQ doesn't matter then we're all being screwed over.J Mankila wrote:
This isn't a popularity contest. It doesn't matter what the "big public" says. It really amounts to squat. A common ignorant notion, luckily, doesn't change any of the facts.
Is this the reason why you think bigger pixels equals to better image quality?
PS: I quite abhor the "we ar the users, you shall serve us" mindset in the economical context. It's for lemmings scavenging for scraps of power. There's no sense in slicing every discussion down to the greatest common divisor.
To expect similar image quality with compacts and DSLRs is a bit dubious, really. The sensor size is different, which means the usual DOF is vastly different. This is something that you will have to note when assessing image quality, because it affects you greatly even if you weren't acknowledging it. And I believe that's what's happened.
The compacts also have a lot of noise reduction, which will make any photo look unreal. The reason compacts have such little sensors has to do with our beloved "big public" who wants pocketable (or jacketable) 40x zooms and 15MP at the same time. Blame them (or yourself, if you will) for the lacking image quality, not the D800.
Look at this comparison and tell me which has pure, lovely image quality.
http://www.jambor.ro/blog/2012/03/13/nikon-d800-vs-d700-high-iso-comparison/
This is what reality will look like if you started making bigger pixels:
Yes. Pure joy and happy memories fill me when looking at the photo, so you might be on to something, after all...
Please, listen to me. Increasing the megapixel count before the technology was ready was the reason for poor image quality. Now, when the technology seems to have been on a long winning streak, there's no point in keeping the pixel count low . If you're got a pole vaulter who can do five meters blindfolded, you're not going to put the bar on three meters.No, you are. "One of the most persistent and wide spread mythologies of the photographic world" is that "more megapixels = better pictures", that is what you're all saying here. That's nothing different to what the mainstream masses have been led to believe by the marketing hype and runs counter to many user experiences.Grevture wrote:
You are caught up in one of the most persistent and wide spread mythologies of teh photographic world. That is the simple fact
--
regards
Janne Mankila, Finland
bobn2
Forum Pro
OK, then in that case, your 'diffraction limit' works in reverse of the McHugh diffraction limit, that is, the more pixels the higher is the f-number at which the 'diffraction limit' occurs.
--
Bob
bobn2
Forum Pro
noirdesir has given you the answer to that. It doesn't do that, so that is not a practical use.
--
Bob
If you evaluate this fixed MTF threshold at the same output resolution for a given sensor size (or a fixed physical sensor area). When comparing sensors of different physical sizes at the same output resolution, this 'absolute' diffraction limit occurs at different f-stops (but at the same DOF).
Robin Casady
Forum Pro
So, the practical question for the photographer becomes, at what MTF will this effect become noticeable? That would depend on the print size and viewing distance. If one could calculate that, one would know where in the graph the effect would make a difference to a photographer.You haven't followed this at all. No, it doesn't do that. Lets have a look at my lens graphs again:
McHugh's calculator says that the D3 starts to 'suffer diffraction effects' at f/16, and the D3X at f/13.
Now look at the shape of the curves. can you see any special feature on the D3 curve at f/16 or the D3X curve at f/13 that would show the suffering of 'diffraction effects'.
The truth is this, they suffer 'diffraction effects' at all apertures, that is physics. Diffraction effects become the dominant determinant of resolution after f/5.6, for both cameras (when the line begins to fall). There is nothing on the graph to show the presence of McHugh's 'diffraction limit'. It does not exist, so the calculator gives no useful information. All it does is confuse people, and you are one of the people it has confused.
--
Robin Casady
http://www.robincasady.com/Photo/index.html
Well, you can try out different apertures and see where the result turns into 'YES', can't you? Up to that point you can count on more resolution when increasing the f-number.
Having said that, I tend to think there are better alternatives, like testing each of one's own lenses for oneself.
--
regards
Janne Mankila, Finland
I believe this is the practical use that the calculator was put there for. It actually has very little to do with Bob's and John's arguments.
--
regards
Janne Mankila, Finland
Certainly no, just look at the example Bob has posted. The resolution on the 50 mm f/1.8 lens peaks at f/5.6, thus it falls when going to f/8 and further going to f/11 when measured on the D3. But the calculator spits out f/16. Thus you don't increase resolution up to this f/16 with the D3 (on that lens).
Wait a moment - who said anything about never making large prints? If you don't print big, you can make do with a P&S. If you don't print at all, you don't need a camera and can save the cash for plenty of things.
Your logic is faulty. Could you fix it?
The advantages of more megapixels are clear. Look at this graph and tell me what you see?
http://forums.dpreview.com/forums/read.asp?forum=1021&message=40897096
--
regards
Janne Mankila, Finland
Take a very good lens peaking at f/4 and take a very bad lens peaking at f/11. At f/13 (what this calculator finds for the D3x for a print size fully utilising the D3x resolution), the bad lens might be close to the f/13 performance, however, the very good lens will be much better at f/4 than at f/13. But from this calculator you would not know that the very good lens would be much better at f/4.
The practical use of this calculator is for systems with fairly slow lenses (in relative terms, ie, measured in DOF). If this calculated 'diffraction limit' is at a close the widest f-stop then this tells you at this lens likely is best wide open.
I understand the phenomenon in Bob's graph quite well, yes. But the issue, here, is that all that resolution is redundant, superfluous if the print size and viewing distance make it impossible to see a difference.
That is Robin's point, and that is missed by the lot of you. You're arguing a different definition.
--
regards
Janne Mankila, Finland
bobn2
Forum Pro
Certainly, on can make an assessment of what MTF50 one wants for a particular print size, but the calculator still doesn't help you with that. Say you wanted 29 lpmm for your chosen print size and viewing distance. That means on the D3 you can choose in the range f/2.8 to f/8. On the D3X you can choose in the range f/1.8 to f/13. The calculator doesn't tell you that, in fact it tells you the reverse, it says you have less choice with the D3X.So, the practical question for the photographer becomes, at what MTF will this effect become noticeable? That would depend on the print size and viewing distance. If one could calculate that, one would know where in the graph the effect would make a difference to a photographer.You haven't followed this at all. No, it doesn't do that. Lets have a look at my lens graphs again:
McHugh's calculator says that the D3 starts to 'suffer diffraction effects' at f/16, and the D3X at f/13.
Now look at the shape of the curves. can you see any special feature on the D3 curve at f/16 or the D3X curve at f/13 that would show the suffering of 'diffraction effects'.
The truth is this, they suffer 'diffraction effects' at all apertures, that is physics. Diffraction effects become the dominant determinant of resolution after f/5.6, for both cameras (when the line begins to fall). There is nothing on the graph to show the presence of McHugh's 'diffraction limit'. It does not exist, so the calculator gives no useful information. All it does is confuse people, and you are one of the people it has confused.
--
Bob
Keyboard shortcuts
- f
- Forum