HELP!! Need extraordinary sharpening of image

Due to a lens malfunction, I have an image needed by a client that is very out of focus. Any suggestions for sharpening? It's going to need more than the typical unsharp mask. I think it could somehow be salvaged with some intricate work, and would appreciate any thoughts on the approach. Duplicate layer, high contrast, blending modes -- I'm trying everything and must deliver in the morning. Can anybody out there help?

Thanks.

Terry

From:

Due to a lens malfunction, I have an image needed by a client that is very out of focus. Any suggestions for sharpening?

Hard to fix with USM ... can you post a jpeg so people can see the problem?

This is my first time with posting onto the groups, and I'm unfamiliar on attaching a JPG. I'll keep trying, but the photo is a grip and grin shot of 3 people. Imagine shooting through a heavy diffusion filter! Any thoughts? I'm now trying to use the "find edges" filter on another layer, hoping something may work with blending, etc.

wrote:

This is my first time with posting onto the groups, and I'm unfamiliar on attaching a JPG. I'll keep trying, but the photo is a grip and grin shot of 3 people. Imagine shooting through a heavy diffusion filter! Any thoughts? I'm now trying to use the "find edges" filter on another layer, hoping something may work with blending, etc.

Typically, you'd toss the picture onto a web site, and then copy and past the URL into a post replying to the last message in your thread.

Truly out of focus can't be focussed, but it maybe can be altered to look, ah, better, if not sharper.

--
John Mcwilliams

wrote:

Due to a lens malfunction, I have an image needed by a client that is very out of focus. Any suggestions for sharpening? It's going to

need

more than the typical unsharp mask.

An obscurish part solution is making a duplicate layer,
running the Filter->Other->High Pass(6 ish? Play around) on this layer and then setting layer properties to overlay. The higher the High Pass is set the darker the edges will become.

Perhaps setting an action with interaction (whatever that little red box on the actions panel is) on the High Pass would be the best way to experiment with this.
This does mess with your colours quite badly iirc.

Best wishes,
Dave

Thanks Dave, I'll give it a spin. My other attempts are in vain, so I appreciate your feedback. I can deal with the color shifts; just really need to at least add some edge sharpness.

Terry

The higher the High Pass is set the darker the edges will become.

Sorry, that was completely wrong. More contrasty? Sorry, it's late here.

It gets kinda' complicated but Deke Mcclelland describes a process in the PS 7 Bible. Here is synopsis

Dupe one of color channels
Filter>Stylize>Find Edges
Invert
Filter>Other>Maximum
Filter>Noise>Median
Filter>Blur>gaussian
Return to composite view
Convert the mask to a selection
Filter>Sharpen>Unsharp Mask - 500%, Radius 2, Threshold 0

Basically what this does is take the elements that have sharp distinctions and makes a mask of them. When the unsharp mask filter is applied, the smooth areas are masked out so they do not get grainy.

As always, trial and error must be employed. Sometimes, what you got is as good as you can get.

Good luck.

wrote:

Thanks Dave, I'll give it a spin. My other attempts are in vain, so I appreciate your feedback. I can deal with the color shifts; just really need to at least add some edge sharpness.

Terry

You could try unshake from www.unshake.co.uk

Andrew

"Psicotix" wrote in message
SNIP

This does mess with your colours quite badly iirc.

Only if you leave the High pass layer in RGB. If you desaturate it (or channel mix to monochrome) it will only affect the contrast of the chosen spatial frequency. One can improve multiple frequencies with multiple HP layers and, after flattening, the procedure can be repeated multiple times.

What the OP really needs, is an attempt with a deconvolution method. Recommendations depend on the image (size/noise), but the timeframe/deadline seems too short/soon to find the best solution.

Bart

"Bart van der Wolf" wrote

What the OP really needs, is an attempt with a deconvolution method. Recommendations depend on the image (size/noise), but the timeframe/deadline seems too short/soon to find the best solution.

my thought exactly.
I wonder if de-focusing in camera lenses (if that's the term I'm looking for) is consistent enough across various types, focal lengths, etc., etc. that a solution with one or two parameters is possible, so that a plug in could be devised with the usual slider or two.
If so, I'd be surprised if someone hasn't done that.

In article ,
"ggull" writes:

"Bart van der Wolf" wrote

What the OP really needs, is an attempt with a deconvolution method. Recommendations depend on the image (size/noise), but the timeframe/deadline seems too short/soon to find the best solution.

my thought exactly.
I wonder if de-focusing in camera lenses (if that's the term I'm looking for) is consistent enough across various types, focal lengths, etc., etc. that a solution with one or two parameters is possible, so that a plug in could be devised with the usual slider or two.
If so, I'd be surprised if someone hasn't done that.

There are several such products, such as Focus Fixer and Focus Magic, as well as high-end expensive scientific applications. But none of these will do what the OP apparently wants, which is currently possible only on TV shows.

--

Warren S. Sarle SAS Institute Inc. The opinions expressed here SAS Campus Drive are mine and not necessarily
(919) 677-8000 Cary, NC 27513, USA those of SAS Institute.

ggull wrote:

"Bart van der Wolf" wrote

What the OP really needs, is an attempt with a deconvolution

method.

Recommendations depend on the image (size/noise), but the timeframe/deadline seems too short/soon to find the best solution.

my thought exactly.
I wonder if de-focusing in camera lenses (if that's the term I'm

looking

for) is consistent enough across various types, focal lengths, etc.,

etc.

that a solution with one or two parameters is possible, so that a

plug in

could be devised with the usual slider or two.
If so, I'd be surprised if someone hasn't done that.

At our studio muckups where sharpening is necessary occur either when a photographer has moved the camera focus point and/or focused incorrectly. 'de-focusing in camera lenses' is a very polite way of describing their incompetence ;)

ds

"Psicotix" wrote ...

ggull wrote:

I wonder if de-focusing in camera lenses (if that's the term I'm

looking

for) is consistent enough across various types, focal lengths, etc.,

etc.

that a solution with one or two parameters is possible, so that a

plug in

could be devised with the usual slider or two.
If so, I'd be surprised if someone hasn't done that.

At our studio muckups where sharpening is necessary occur either when a photographer has moved the camera focus point and/or focused incorrectly. 'de-focusing in camera lenses' is a very polite way of describing their incompetence ;)

Well, I was referring to the action of the lens on the image, not the operator on the lens. But feel free to borrow the term :-).