Post Processing Sharpening - the Mad way!

[Mad Aussie]

I was asked recently if I wouldn't mind walking a few members here through how I go about sharpening my images when I find that necessary.

What follows below is the way I go about this task. This is not the only way, neither do I suggest it's the best or right way. It's just what I do for reasons that should become apparent as you read through.

I should say that I do usually leave my sharpening to last or at least one of the final parts of my post processing. No point concerning myself with sharpening parts of an image I will later crop out or sharpening up artifacts, dust etc that need to be removed, cloned etc.

So ... here's our image ...
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen1.jpg

You'll notice the whole owl is a little soft. Look carefully at his eye. That's definitely too soft for my liking and that's what we will concentrate on here.


Step 1 - Duplicate the layer
In this screenshot I have highlighted the Layers Palette in Photoshop. You can see 2 layers there. One is the original image and the 2nd is a duplicate of that layer. They are therefore Identical.
Use the tabs Layers/Duplicate layer to achieve this.
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen2.jpg


Step 2 - Make a Layer Mask
In this shot I have highlighted the layer mask at the bottom of the layers palette and also the layer itself which shows up as a white box next to the image layer.
To achieve this simply select the top layer in the palette and click that layer mask button at the bottom.
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen3.jpg


Step 3 - Selecting the lower layer
What I'm going to do is sharpen the lower layer and use the layer mask on the upper layer to allow us to see through to the bottom layer. Think of two identical photos, one on top of the other. You get scissors and cut out the eye of the owl on the top photo. Now you see the eye of the owl on the bottom photo only. Any changes you make to that bottom layer will only show through the hole in the layer mask we'll create soon. Believe it or not, this is how some photographers like Ansell Adams used to do post process some of their works by cutting negatives in a similar way.
So .. to select the lower layer, first click on the 'eye' that I've highlighted by the top layer in the layer palette. This turns that layer off so we can see the layer below properly for the time being.
Make sure you have selected the lower layer also by clicking on it.
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen4.jpg


Step 4 - Sharpening the eye
I've highlighted the owls eye in this shot to illustrate that I don't care what happens to the rest of the bird at all when I sharpen now. I only care what that eye looks like. Remember, when we turn on that top layer the eye is all that will be visible through the 'hole.'
Now use you favoured method of sharpening. I use unsharp mask most of the time. You may prefer smart sharpen for instance and that's fine.
Sharpen that eye to the level you are most happy with. Don't worry if it creates artifacts or makes the feathers etc look crappy.
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen5.jpg


Step 5 - Using the Layer Mask
Here's the cool bit. The Layer Mask. Turn on the top layer again by clicking that little eye icon.
Select the layer Mask by clicking on the highlighted white box that is the Layer Mask. A white border should appear to show you have selected it.
Now ... to cut a 'hole' (or erase parts of that top layer) you use a black brush. So using the brush tool, select black, choose the flow you want ( I use about 30% so it takes a few passes to get full black) and paint over the eye. As you do so, you'll notice the eye getting sharper because the bottom layer is starting to show through where you paint.
If you make a mistake or change your mind about how much you painted, simply select white and paint back over and the mask will be put back.
You'll notice, by my highlighting, that I chose to paint over both the eye and the beak.
Also ... if you look at your layer mask (white box) you'll see the shape of what you painted in black. You can see the eye and beak in my mask in my highlight.
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen6.jpg


Step 6 - Finishing up
In this shot you can see I've highlighted the top layer again.
The reason for that is I decided I might sharpen the rest of the owl now, just a little. So I select the top layer as highlighted and applied an unsharp mask to a small degree but not enough to take away that soft feel to the feathers.
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen7.jpg


And that's it! Flatten the image and save it as needed.

[Mad Aussie]

Here's the finished product ...
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen8.jpg

And the original again to compare ...
http://i166.photobucket.com/albums/u104/Mad-Aussie/09%20photographyca/sharpening/sharpen1.jpg

[zenon5940]

Thanks MA for this demonstration. You explain real well the process and the illustrations help understand. You sure have an eye for little details.

Thank you.

Clement

[EJC]

Ditto here too MA. This is very helpful and easy to follow. Thanks!

cheers

[kat]

Yay! Thanks so much MA!

[Mad Aussie]

Hope it helps all.

[AntZ]

Thanks MA.
Certainly not something you would want to do to every image. I wonder how the result would differ from selective sharpening available is something like lightroom?

[Mad Aussie]

Thanks MA.
Certainly not something you would want to do to every image. I wonder how the result would differ from selective sharpening available is something like lightroom?
I find the lightroom method very clumsy myself.

The added advantage with this method I use is that while you have the two layers you can also be more selective with the saturation. For example, I could have saturated the eyes on the lower layer or even changed their colour entirely if I really wanted. Not that I'd do that to a bird.

Not every images needs this sharpening of course. Some of my shots need nothing, some just a little overall, and some need this method whether it be needed or just my taste.

[Greg_Nuspel]

MA excellent method, thanks for sharing your knowledge. :1st:

[Mad Aussie]

You're welcome Greg.

[Marko]

Thanks so much for that MA - You rock! :highfive::goodvibes

Just as an FYI, some people like to know the percentage of shots that require sharpening. I shoot in RAW and I use sharpening tools on 99.5% of all my images. I cannot remember the last time I did not sharpen (even if it's only a tad) an image.

[Mad Aussie]

Thanks marko

I do usually sharpen a little at least on about 90% of my shots I guess. It depends on what I want. Some shots, using that method I've described above, have extensive sharpening in selective areas obviously.

[kat]

Here's my question. Correct me if I am wrong, but would one use different amount (%) of sharpening depending on the size you are printing? How would you know without a print actually in front of you? Boy..I hope this doesn't come out as stupid!!! :p

[Mad Aussie]

Here's my question. Correct me if I am wrong, but would one use different amount (%) of sharpening depending on the size you are printing? How would you know without a print actually in front of you? Boy..I hope this doesn't come out as stupid!!! :p
Perfectly good question.
I usually use a similar % (using unsharp mask) but I use a different radius depending on the size of the photo I'm sharpening.
I make the decision on how much the ahrpening is taking effect as I use the %. If I'm up at 80 -100% and not seeing enough sharpening I usually increase the radius.

[kat]

Perfectly good question.
I usually use a similar % (using unsharp mask) but I use a different radius depending on the size of the photo I'm sharpening.
I make the decision on how much the ahrpening is taking effect as I use the %. If I'm up at 80 -100% and not seeing enough sharpening I usually increase the radius.

That makes perfect sense! Thank you!!! :thankyou:

[Marko]

I use considerably more (probably around 30%) sharpening when the image is for print.

[kat]

Now here's the next question. The bigger the print..won't you see the noise from sharpening? Or is there noise. I've never sharpened my photos before print..always thought everything was sooo soft..now I know why! And here i was blaming the lens/camera and myself!!!! :)

[kat]

And now I'm confused again. Do you use unsharp mask marko? do you mean it's at 30% in that?

[Marko]

I sometimes use unsharp mask and I sometimes use smart sharpen.

Whatever I use, when the shot is destined for printing (something like an 8x10) that shot will likely require 30% more sharpening (or more) than if that same shot was destined for the web.

Please keep in mind that this is my personal recipe..and I'm giving average numbers (sometimes I sharpen more, sometimes less), and by NO means should anyone accept this as gospel. YOUR camera and PP technique may require you to use your own recipe.


Now here's the next question. The bigger the print..won't you see the noise from sharpening? Or is there noise. I've never sharpened my photos before print..always thought everything was sooo soft..now I know why! And here i was blaming the lens/camera and myself!!!!

IF you look at the print from 1 inch away, yes you may see the effects of the sharpening. I forget the formula for the correct distance to look at prints according to their size...but if you respect that distance, the sharpening should not be noticeable.
Of course you CAN oversharpen and you'll see these artifacts. A VERY general guide might be 100-200% sharpening for an 8x10 with a radius around 1.0 and a threshold around 0. Again, your recipe may vary depending on the camera, image and PP technique.

The only point I really wanted to make is that in general, in my experience, print images require a significant boost in sharpening regardless of which sharpening technique you use.


Hope that helps - Marko

[kat]

Okay, I got yah! I think it's time to start playing around with the sharpening! :) Thank you!

[tomorrowstreasures]

Great tutorial MA!!!
:highfive::highfive::thankyou:

page two of this thread felt like i was reading Swahili. i cannot wait for all of this math stuff to make sense to my brain. exposure, exposure, exposure.

[Mad Aussie]

Ahsante Sana TT (thank you in Swahili)
Glad you liked it.

[Mad Aussie]

I think the correct viewing distance is twice the length of the longest side of the photo? From memory anyhow. And that of course was only ever a general rule of thumb.

Basically our little cameras of around 10mp or so can take a useful photo for a billboard that's several meters long and high. But if we stood right next to it the photo is going to look like crap in a big way. But view it from the highway or whatever at many meters back and it can look perfectly for a billboard. My brother had several of his shots used for billboards and his camera is a 6mp Canon D60.

At 10mp, it is said that a photo of around 8x10, 8x12 ish (it's worked out on the resolution of the camera) would be the largest in 'true photo quality' which in rough terms means there would be no degeneration due to blowing the photo up to this size. Above that and some quality loss is evident and viewing from further back becomes necessary. That's roughly the theory anyhow.