Gaussian blur affects every pixel. Dust and scratches, presumably, looks for isolated areas of very high local contrast and eliminates them — probably by a combination of gaussian blur and other techniques. The dust and scratches filter should not affect at all the areas of the image not having very high local contrast.
If you don’t change the threshold, Dust&Scratches is the same as Median. And Median is nothing like Gaussian Blur.
If you do change the threshold, then Dust&Scratches only affects areas with rather strong edges — according to the threshold you set.
thanks Michael and Chris:
funny I never pay attention to threadhold. Can you go over that Chris? I’m not thinking about edges am I; aren’t I thinking about dust spots?
And why do the results look the same to me (as gaussian) — a blurry look.
I’m not thinking about edges am I; aren’t I thinking about dust spots?
the dust spots have edges, no?
That’s my mom, quite a few years ago.
Linda
While we all enjoy answering your questions, it is not possible for a beginner to get a basic understanding of Photoshop just by Q & A sessions in forums.
By all means go on asking questions ย yours are all reasonable (not like some of the really stupid questions that appear from time to time here). But make more use of the Help files.
For example:
To use the Dust And Scratches filter Choose Filter > Noise > Dust & Scratches. If necessary, adjust the preview zoom ratio until the area containing noise is visible. Drag the Threshold slider left to 0 to turn off the value so that all pixels in the selection or image can be examined. The Threshold value determines how dissimilar the pixels should be before they are eliminated.
Note: The Threshold slider gives greater control for values between 0 and 128ยthe most common range for imagesยthan for values between 128 and 255. Drag the Radius slider left or right, or enter a value in the text box from 1 to 16 pixels. The Radius value determines the size of the area searched for dissimilar pixels. Increasing the radius blurs the image. Use the smallest value that eliminates the defects. Increase the threshold gradually by entering a value or by dragging the slider to the highest possible value that eliminates defects.
Your questions betray a lack of basic image editing principles which only a bit of reading can remedy. Spend some time studying: read the Manual; read the Help files; read the "How to…" articles. Seek tutorials on the internet and buy yourself a beginners book on Photoshop and read it till you understand at least the principles and terminology.
But, having said all that, please keep coming with you questions ๐
Don’t listen to him Linda (he can be a bit well, you know, from time-to-time)
Your question is IMHO a great question and probably based on some onscreen observations.
The matrix of effects at different setings is a great way to find out to your own satisfaction
As a little observation adding a bit of blurring (no matter how refined or selective or both) may need a tiniest little tweak on USM as well just to sharpen things up once the blemishes have gone (IMHO)
MIchael:
thanks, so you combine both D&S and Gaussian? did not think of that. But is it necessary, should not the D&S be enough?
HA! I agree with you Deebs here!
Thanks Josh, your insights are great, but (it’s a personal thing) I often refer to many books, but in the end, I like the human feedback in conjunction with the books– but that clip you just posted was great — as is M. Sullivans work — which you wont see in any book anywhere.
That clip comes straight from the help files… F1, one click away… I see where John is coming from, and Deebs too.
I often duplicate the layers that will be affected, in order to mask the filter work.
What a wonderful reply Linda!
Good luck with your project
It sounds as if you have good analytical skills that will find the best working option for the task to do
Photoshop tools often work best in combination, or in repeated applications. Experiment!
deebs
Did you notice the title of the thread?
In fact Michael and Chris said it all in their first posts. And Dave clarified it even more.
We all know you are trying to help but your philosophising and the obscurity of your answers is often quite baffling, even for the more experienced among us.
Keep trying ๐
Beg pardon JJ but it is within context of the flow of the thread.
It relates to composition of functionality touched upon by using several filters to obtain an effect.
The combination of blur followed by sharpening is, I think, well known the sequencing or ordering of filters is also (IMHO) important. For example gaussian blur followed by USM does not generally give the same results as USM followed by blur.
A vizualisation of it really does help IMHO and the parallels to the original topic and subsequently discussions seem, to me, rather obvious?
In logical or mathematical terms f compose g in this case is not symmetrical as g compose f gives different results.
I hope this is clear 8)
I am sure it will be soon ๐
A very, very easy way to understand it (it works with year 6 pupils) is
putting your socks on before your shoes is not the same as putting your shoes on before your socks.
Of course, we have to abstract the language a little to something like f*g <> g*f
but the essence is the same
