Sharpness VS resolution questions - a little confused

W

wildtracks

Suspended / Banned
Messages
222
Name
Martin Prothero
Edit My Images
Yes
OK. I have been trying to get my head around this (even though it is pretty basic) and I could do with some clarification please.

I have a D40 so 6MP. More pixels means larger printing size (and more crop or zoom-in of original file). - please start to correct me if I go wrong.

At screen resolution lets say 300ppi images taken with say a D40 and a D80 with the same lens, will look almost identical (allowing for the in camera sharpening etc) due to the fact that they are reduced to screen res. It is not until you zoom in to 100% that the better images make themselves obvious...

So, if the majority of your images are destined for screen viewing and not printing, then there is nor real advantage in increasing you pixels count? Obviously other reasons to have better cameras, but we're just talking pixels at te moment.

The questions continue...If I have a kit lens and sharpen the final images in PS to a point where they look good on screen...how is it that I can still tell the difference between a sharp image taken with a D80 using pro glass compared to a similar image taken using a D40 using the kit lens ON SCREEN?

I'm confused now. How dos a superior lens resolve better images if they are reduced to screen resolution?

I am missing something really obvious:thinking:
 
You're right to a point. Scaling an image down to screen size will hide all manner of problems with sharpness, noise, etc.

A good lens will resolve more detail and that will be noticed even when reduced to screen size. A macro shot of an insect's eye for example when reduced down will still hint at the compound cell detail from a sharply captured original. The same shot from a lower quality lens that didn't capture the detail will scale down to an almost flat colour.

Chromatic abberations will survive scaling to screen size as well, leaving a purple fringe on the edges of high contrast.
 
Thanks for that clarification, that helps.

But how does the better lens resolve more detail? What I mean is, does more pixels on a sensor allow more detail to be resolved and therefore pro glass on a lower res camera is a waste of money, or is it that a better lens somehow defines an image at pixel level?
 
Naturally the more pixels you have the finer the detail you can resolve on the sensor but even on a lower resolution sensor the better glass still produces a better image.

Think of it in terms of a photocopier. The copy is always a lower quality than the original so using the better the original the better the result. In camera terms the lens defines how good the original is.
 
Ahh. Thanks

So, when I check out flicker (rasone - the persons name) he is using the same lens as me yet has far crisper results, does that mean it is down to PS post processing or maybe that I have a slightly bad copy of that lens?

I have just received my 2nd copy of the 10-20 sigma and it is just as soft as the first one. I was wondering if it is the max resolution of my camera that is the limiting factor - but then my nifty shows far superior detail...
 
What post work are you doing when you prepare an image for the web?
 
Normally minimal processing. If I shot in raw then it needs a little bit more work but generally I will apply a bit of smart sharpen and very occasionally I might adjust the curves slightly.

Do you think I might benefit from a PS tutorial to get what I am looking for? If so, is there anywhere you recommend?
 
Do you sharpen after sizing the image down for the web?
 
No. Normally I sharpen before, do any colour changes and then resize at the end. I never save for web, just as a maximum setting Jpeg. I think my preset saves in sRGB.

I always figured the sharpening would be more effective with finer resolution in the image and imagined a bit like a mosaic - the more pieces you have the more detailed the adjustments can be.

Might just be barking up the wrong tree though?
 
You do need to do another sharpen after re-sizing and you'll notice that the results are much better for it. Re-sizing basically averages the pixels so a black/white edge will become grey and that leads to a soft result. Sharpening again recovers the detail lost due to the scaling. Be careful not to go mad tho, watch out for "halos" around high contrast edges - if you seem them lower the amount until they go away.
 
Many thanks for the tip. I have still yet to master the whole pixel radius, fade amount and tonal width thing. Is it me or do they not sound very logical?
 
I don't use Smart Sharpen so maybe someone else can chip with some good settings for that.

Try UnSharp Mask with amount=75%, radius=0.8, threshold=10 as a starting point.
 
Many thanks for all your advice. I think I'll get a book or search for an up-to date tutorial on-line to help 'refresh' my understanding!
Cheers PXL8
 
I tend to just select 'Bicubic - sharper' in the resizing window before I reduce the size of an image or 'Bicubic - smoother' when increasing an image size.


Seems to work o.k.
 
You must log in or register to reply here.
Back
Top