Sharpening often adds brightness to interior areas and when the sharpened image is viewed those areas look brighter. But in the process, tonal or color gradient get wiped out. The end result is a sort of gritty look. Leaves with detail at full size become single color/tone blobs.
The real problem is that the detail including tone and color gradations is lost in the crunch. There is no way to compress an image 6x1 or more without losing this detail.
Downsizing is a curse with no true cure. For web viewing, I think it best to present the image with halo free sharpening and forget about how to preserve the detail of the original.
The boatyard unsampled TIFF above is sharp enough for web just as it is shown before sharpening. After all, web viewing is crippled viewing.
Also, 1280 Pixels? The landscape forum has a pitiful max of 850.
In the end it seems to come down to personal taste. I like a crispy image and do not worry about some minor halos, others prefer less sharpened images without any sharpening artifacts.
If the limit here was 850px i would emigrate somewhere else .
Phillip Reeve wrote:
In the end it seems to come down to personal taste. I like a crispy image and do not worry about some minor halos, others prefer less sharpened images without any sharpening artifacts.
If the limit here was 850px i would emigrate somewhere else .
Yep, I quit landscape, but some are posting 36mp images there and get pretty crispy looking after the big crunch.
The etiquette seems to be to make it easy to view on slow connections or small screens. Maybe this is a by product of travelers who use their laptops from campgrounds as many landscapers do.
I find the smallest decent size for my 21mpixel images is 1900 displayed at 26 inches wide.
The result is the bottom row. I guess I missed something?
(And, out of interest, did you have time to work anything with the Alt Iceland Tour? I haven't posted in that thread but I'm medium interested. That may turn into highly interested depending on How, When and price.)
Hm Jonas, the difference is huge, and the only factor I could think of causing such a difference, is if the downsizing is not really Nearest Neighbor, even if you ask for it.
I tried the recipe but with bicubic downsizing instead, and then got very close to your result. I use CS5. Can you try with bicubic and see what happens then? Do you see any difference?
In any case, the first step must NOT be NN. Using NN on anything but a whole number ratio, will cause periodic displacements of detail. So bicubic for the first small resize
I like resizing in steps and have now started resizing height and width separately and only do sharpening once the file is in final size. The default sharpening is 0,2 150-200 and sometimes two passes if I feel the image needs it. Some images require different sharpening for different parts so sometimes I use adjustment layers to control the amount of sharpening depending on areas an for eliminating halos.
And sometimes I feel lazy and do direct exports from LR with (screen && high).
alundeb wrote:
Hm Jonas, the difference is huge, and the only factor I could think of causing such a difference, is if the downsizing is not really Nearest Neighbor, even if you ask for it.(...) So bicubic for the first small resize
Hi
That made it. I didn't get the exact same result as you but close enough. Thank you for the tip(s)!
AhamB wrote:
Capture sharpening is standard. It will be negated anyway when you downsize.
Btw, your idea of "microcontrast boosting" isn't bumping clarity to 100, I hope?
WTH is "capture sharpening"? AFAIK in-camera sharpening only applies to jpegs. There may be some in in-camera" Tiffs tho... I dunno. I thought not.
Clarity as Adobe describes it and is obvious by looking at it has nothing to do micro-contrast. It's a mixture global contrast boosting and "smart sharpening".
I mean using one of the many micro-contrast boosters in place of sharpening when doing the stepped reduction method and using "Best For Smooth Gradients" when doing the reductions. The amount that it "sharpens" the image is of course controllable but no matter how much you apply you never get halos unless they already exist - Like they do in the supplied TIFF. So, I can't show it using the supplied image because he already sharpened it producing halos. Here's a quicky example I did but where his processing produced halos I selected them out for the final step.
Also no, halos do not go away when downsizing an image. In fact they usually accentuate and become more obvious.
Phillip Reeve wrote:
@bif if we based it on the raw we would compare raw converters and personal taste, not re-sizing algorithms. Anyway, here is the unsharpened tif, i am courious about your method
OK, I missed this post before I did my edits - which actually I did only about 30min. after you posted it but didn't reply till just a little while ago.
I'll try it again with this file maybe if there's more interest. My down-scaling doesn't add any halos at all BTW. In PS the "best for smooth gradients" actually blends nearest neighbors. It might be similar to the 0.4 pixel gaussian blur that was used on the previous page.
Bifurcator wrote:
WTH is "capture sharpening"? AFAIK in-camera sharpening only applies to jpegs. There may be some in in-camera" Tiffs tho... I dunno. I thought not.
Capture sharpening is a term for the sharpening you apply (at full size) to negate the effect of the AA filter that most cameras have.
Bifurcator wrote:
I mean using one of the many micro-contrast boosters in place of sharpening when doing the stepped reduction method and using "Best For Smooth Gradients" when doing the reductions. The amount that it "sharpens" the image is of course controllable but no matter how much you apply you never get halos unless they already exist (...)
Would it be possible for you to share the method? I would like to try something like this out. Step by step instructions even?
Bifurcator wrote:
Because it's NOT unsharpened. Look at it at 200% It has sharpening applied to it prior to his uploading of it.
indeed, but he posted an unsharpened one later. there is no need to ever look at something at 200% unless you have poor eyesight, it doesn't add anything to looking at 100%.
That's your own opinion. It doesn't apply to me at all. I often NEED to see it at 200%, 300% and 400% to know what's going on, how things work, what the original image contains, etc. etc.
I'll post a step by step in a little bit... an hour or two most likely.
Bifurcator wrote:
That's your own opinion. It doesn't apply to me at all. I often NEED to see it at 200%, 300% and 400% to know what's going on, how things work, what the original image contains, etc. etc.
I'll post a step by step in a little bit... an hour or two most likely.
not really an opinion, i did say it depended on your eyesight. there is no change in the information displayed at 100% versus 400%, anything above 100% is just pretending you have a lower res monitor. if one can distinguish the individual pixels on your monitor at normal viewing distance, then blowing something up beyond 100% doesn't accomplish anything for them. it also means that you are stuck seeing stair step patterns in diagonals no matter what downsizing algorithm you use.
OK, here's my attempt, with a slightly tweaked variant on my usual quick-and-dirty re-sizing script. Using the "convert" command line tool from GraphicsMagick:
with pre-sharpening at various scales, followed by a slightly under-sampling downscale and final re-sharpening. I've recently switched to a MacBook Pro with retina display, which has really thrown off my ability to tell what images will look like on "normal" resolution displays, so I'm a bit unsure about my sharpening-for-web procedure these days.
'Oversharpening at FM is as bad as the HDR-frenzy was a couple of years ago.'
Thank you, these all look very very digital.
Highlights lose tonal range and tone differentiation, saturation and acutance amps up unnaturally and three dimensionality suffer badly - as everything is preternaturally 'sharp' and the viewer's eye is affronted as a result. I guess it is an aesthetic choice.
alundeb wrote:
No. Blowing something up beyond 100% with NN is lossless reverisble. Just use NN back to the original size.
sorry, i meant being able to distinguish individual pixels at normal viewing distance on your monitor means that your are cursed to see stair step patterns no matter what.
sebboh wrote:
sorry, i meant being able to distinguish individual pixels at normal viewing distance on your monitor means that your are cursed to see stair step patterns no matter what.
Ah, yes.
To all of you who have a problem with my and others crazy pixel sharp images:
Just back off from the monitor a little. At a certain distance, everything falls into place, the detail looks fine and sharp but not unnatural, no stair steps, tonal gradations look fine, the 3D feel is back.
.
