Workflow in Photoshop- best sequence

(eventually setting up a macro maybe). My question is – can anyone tell me or point me to a site that gives best advice about how to generically approach the editing- right now I tend to follow a rough

Here is an article by Roger N. Clark (very knowledgeable) about workflow. While one person’s workflow will not work for someone else this article can act as a guide for further learning and also help to generate search terms for Google/Wiki searches.

http://www.clarkvision.com/photoinfo/digitalworkflow/index.h tml

correct sequence, eg USM towards the end rather than at the beginning.

My understanding of sharpening at this point is that it is one of the more destructive of the transformations so it should be delayed until late in the workflow. I like to edit in the wide-ish gamut color space Adobe RGB with the intention of targeting four destinations, printer, stock, web and extracted smart objects. The edited file is saved as an uncompressed Tiff or PSD in Adobe RGB. The printer version I convert to CMYK and apply ‘smart sharpen’ and other printer related adjustments and save. For the stock version I leave in Adobe RGB, then smart sharpen, then convert to minimally compressed jpg, then save. For the web version I convert to sRGB then ‘smart sharpen’, convert to ~ 100kB jpg, then save. Finally for any smart objects I am able to extract I save with minimal edit (preferably raw) then save. The effect of this is that I sharpen towards the end of the workflow and in the color space I intend to save. Take this with a grain of salt, as other can attest, my knowledge of color management is limited and I am just learning Photoshop myself. HTH

Good luck,
Ron

wrote:
[snip]

My understanding of sharpening at this point is that it is one of the more destructive of the transformations so it should be delayed until late in the workflow.

[snip]

The printer version I convert
to CMYK and apply ‘smart sharpen’ and other printer related adjustments and save. For the stock version I leave in Adobe RGB, then smart sharpen, then convert to minimally compressed jpg, then save. For the web version I convert to sRGB then ‘smart sharpen’, convert to ~ 100kB jpg, then save.

[snip]

The effect of this is that I sharpen towards the end of the workflow and in the color space I intend to save.

[snip]

There appears to be an emerging view that sharpening isn’t something done at one place in the workflow, but in a number. I think of it in 3 places, as follows:

http://www.barry.pearson.name/photography/temp/sharpening_3. gif

Each stage of sharpening is done just enough to satisfy the objectives of that place, and not encroach on the rest of the workflow.

—
Barry Pearson
http://www.barry.pearson.name/photography/

There appears to be an emerging view that sharpening isn’t something done at one place in the workflow, but in a number. I think of it in 3 places, as follows:

The problems for me with multipoint sharpening (and KatWoman should be proud) is that I like the ‘look’ of ‘Smart Sharpen’ better than the other methods.
Maybe I spend too much time staring at before and after screens.

Thanks Barry,
Ron

"Barry Pearson" wrote in message
SNIP

There appears to be an emerging view that sharpening isn’t something done at one place in the workflow, but in a number.

I’m aware of that concept, and of some of the gurus that preach it. I also don’t agree with them.

One very obvious, to me anyway, type of output that does best without any *prior* sharpening is down-sampling. Any high spatial frequency, whether present in the original capture or created by capture AKA ‘creative’ sharpening, in excess of what is possible to be resolved at the final size will create aliasing artifacts.

While there are precautions (low-pass filtering) that can/should be taken when down-sampling, they are not perfect. Not applying sharpening in the first place, will improve the end result.

Another reason to avoid cumulative sharpening steps, although 16-bit/channel processing will allow to get away with more abuse, is because rounding errors will also accumulate. This can lead to edge artifacts.

Finally, whether sharpening needs to be applied at all will be very workflow dependent. For example, most stock bureaus request unsharpened images, and sharpening is hard to undo (unless it’s on its own layer).

Beside those warnings, I’m a big fan of applying ‘Smart sharpening’ tuned to the output modality.

—
Bart

"Bart van der Wolf" wrote in message

"Barry Pearson" wrote in message
SNIP

There appears to be an emerging view that sharpening isn’t something done at one place in the workflow, but in a number.

I’m aware of that concept, and of some of the gurus that preach it. I also don’t agree with them.

One very obvious, to me anyway, type of output that does best without any *prior* sharpening is down-sampling. Any high spatial frequency, whether present in the original capture or created by capture AKA ‘creative’ sharpening, in excess of what is possible to be resolved at the final size will create aliasing artifacts.

While there are precautions (low-pass filtering) that can/should be taken when down-sampling, they are not perfect. Not applying sharpening in the first place, will improve the end result.

Another reason to avoid cumulative sharpening steps, although 16-bit/channel processing will allow to get away with more abuse, is because rounding errors will also accumulate. This can lead to edge artifacts.

Finally, whether sharpening needs to be applied at all will be very workflow dependent. For example, most stock bureaus request unsharpened images, and sharpening is hard to undo (unless it’s on its own layer).
Beside those warnings, I’m a big fan of applying ‘Smart sharpening’ tuned to the output modality.

—
Bart

I don’t have CS2 so I never used smart sharpen
I find different things work on different images depending on many factors like how much sharper you need? how large the image will display or print? type of subject? contrasty vs flatter. sometimes with people subject you don’t want real sharp at all!

I even tried the sharpen tool in the Roxio CD creator (it’s not bad) I used the hi pass method
I got some downloaded actions
I use regular unsharp mask most often
and I always sharpen on a copy or do it on a duplicate layer if the picture is too fuzzy make them re-shoot it!!
I mean good sharp lenses and fast cameras, images look nie with no sharpening in PS if they are really good to start
It is always preferable to get it right in the camera in the first place

"KatWoman" wrote in message
SNIP

I don’t have CS2 so I never used smart sharpen

It is *very* useful, especially with digital captures.

I find different things work on different images depending on many factors like how much sharper you need? how large the image will display or print? type of subject? contrasty vs flatter. sometimes with people subject you don’t want real sharp at all!

Exactly, it’s workflow, and IMHO to a lesser extent subject, dependent. So recipes like capture/creative/output sharpening are too ‘unfocussed’ to be called a recipe.

SNIP

I mean good sharp lenses and fast cameras, images look nie with no sharpening in PS if they are really good to start It is always preferable to get it right in the camera in the first place

That’s correct, but it’s also at the basis of sharpening issues that find their cause in digital capture. Most digital cameras are based on a one-shot full color method that starts out as a Color Filter Array (CFA). It’s a grayscale mosaic of Red, Green, and Blue filtered light, which has to be ‘demosaiced’ into a full color image.

It’s the CFA capture and demosaicing that introduces unsharpness (it has to in order to avoid aliasing artifacts). It can also result in (most visible) lower Red resolution compared to Green or Luminance resolution.

This requires more than just cranking up edge contrast, it requires resolution enhancement. There are methods know in astro photography and signal processing that can do just that. Smart sharpening is a nice compromise between slow resolution enhancement, and a tolerable speed process that emulates that.

This is an example:
< http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jp g> The difference with edge contrast enhancement is that features get smaller/sharper with resolution enhancement, because the blur from neighboring pixels is pulled back into the pixel that caused the exposure.

—
Bart