[PatBranch]

1. Edit raw files in Capture One (no sharpening)
2. Export as high quality adobe rgb 1998 jpegs
3. Edit exported jpegs in Photoshop (includes sharpening) - 'for print' photos
4. Apply landscape and portrait 'for web' actions to those 'for print'(resize, srgb, add logo, stroke) - 'for web' photos


Should I be sharpening the 'for print' files only after I resize them for printing? Also, should I be sharpening the web photos? I was thinking if I oversharpened the 'for print' photos, resizing them for the web would soften them a bit, but stay somewhat sharp (so they don't look overly sharpened on sharp monitors).

---by sharpening, I mean unsharp mask.

 

[H8R]

Limited experience here but...

Shouldn't you be sharpening before you re-size? Finer resolution to work with and you won't be sharpening jpg artifacts, etc.

 

[PatBranch]

That's what I've been thinking...

Maybe I should make the converted jpegs into psd's after photoshop. Would it be that much better to export the raw photos as tiffs, then edit them in photoshop and save them as psd's?

 

[H8R]

That's what I've been thinking...

Maybe I should make the converted jpegs into psd's after photoshop. Would it be that much better to export the raw photos as tiffs, then edit them in photoshop and save them as psd's?

Someone else has to chime in here...

For RAW files (Nikon .NEF) I'm working in Ubuntu Linux with Ufraw plug-in GIMP, then exporting a finished edit into jpg. The Ufraw preview opens first, I tweak exposure, etc then do the rest of the work in GIMP.

Seems like there should be a similar path in PS.

One thing that helped immediately is when I loaded the proper camera profile into Ufraw, (Nikon D70s) it then made the color easier to deal with.

 

[PatBranch]

Yeah, my Phase One has the 20D profile.

What are you doing besides exporting a jpeg that you open in gimp?

 

[Quo Fan]

I use Lightroom to process the RAW, then export it to Photoshop CS2 as a TIFF, edit as needed, save as JPEG.

 

[Strakar]

Limited experience here but...

Shouldn't you be sharpening before you re-size? Finer resolution to work with and you won't be sharpening jpg artifacts, etc.

It's done the other way around. First upsize to the desired size (web, print, etc) and only then do you "sharpen" the image. When you sharpen you aggravate jpeg artifacts (they are always there, shapening->upsizing will make them show up, and then make them big; the result is that you're creating false detail - maze artifacts, soft halos several pixels wide etc, i.e. random crap). If you do it the other way around, when you upsize, the artifacts are not pronounced and you can get away with smaller (proportionally to the new image size) radius USM, still sharpening the image while avoiding making the artifacts visible. Doing it before upsizing also usually kills the "per-pixel" sharpness you just gained (I'm using the "per-pixel" term loosely). In the end you have more control and effectiveness doing it as a final step.

The sharpening you apply should also vary according to the medium you print, I usually overdo it a bit for print, while for web I don't excede 70-80% 0.3 radius in USM.

 

[binary visions]

^ What he said

Sharpening should be the last step in your process before the image hits whatever display medium you want.

 

[PatBranch]

Thanks for the help! Should I save the exported jpeg (just edited) as a psd or something, to prevent further compression? What do others do?

 

[Reactor]

Thanks for the help! Should I save the exported jpeg (just edited) as a psd or something, to prevent further compression? What do others do?

I work with the Raw file(or PSD) right up until I take it to print or convert it to the format I'm I need. If I need to I can run the unsharp mask on it after it is converted to JPEG or whatever.

Once you convert it to a JPEG you've lost data. Even if you save it as a PSD the information is gone, and if you modify it and re-save it as a JPEG later you'll likely lose more data.

 

[PatBranch]

Yeah...

Should I export as tiffs then? And then after PS editing, save as psd or tiff?

 

[Strakar]

Yeah...

Should I export as tiffs then? And then after PS editing, save as psd or tiff?

I would say PSDs for added convenience regarding masking layers, etc etc.

But processing every photo differently is impossible if you have deadlines etc. In that case I batch process them (admitting that I got them right in the camera), and afterwards only go after the ones that standout the most, for a closer editing, etc.

 

[narlus]

i save as PSD, but that's only a problem if you need to get the files to others and they don't have photoshop (and therefore need TIFFs).

what i usually do this:

* open the file, make all adjustments in ACR (adobe's raw converter).

* if necessary, open in photoshop (i'm finding that i can do most image tweaks i need to in ACR, unless i need to run some actions, do cloning, blend exposures, etc).

* select the revised RAW images and choose 'tools > photoshop > image processor' and batch the files using the Manyk Sharpen/Resize/Sharpen action, which works well.

i've got that action sized for jpgs, but also have a full-size (under the 'image fit' step in the action) version if i want to sharpen full-res images prior to printing them.

hope this helps!

 

[PatBranch]

Thanks for that info and action! Will it affect the sharpening if I change that action to 500px wide? I think I'm going to set it to 500px and include that action in my 'for web' action.

Are you talking about a second action, because that's just for web?

 

[narlus]

not sure how that resize/sharpen is just for web; i've used it on full-size images and the prints look fine.

not sure if the algorithm is sensitive to different image sizes; check it out and experiment. i've not adjusted any of the parameters based on final output size, but there probably would be a difference (even if it's slight).

i don't fret about it too much though; i am just happy to find an action which works well, adn i can batch a bunch of files and be done w/ it.

 

[H8R]

It's done the other way around. First upsize to the desired size (web, print, etc) and only then do you "sharpen" the image. When you sharpen you aggravate jpeg artifacts (they are always there, shapening->upsizing will make them show up, and then make them big; the result is that you're creating false detail - maze artifacts, soft halos several pixels wide etc, i.e. random crap). If you do it the other way around, when you upsize, the artifacts are not pronounced and you can get away with smaller (proportionally to the new image size) radius USM, still sharpening the image while avoiding making the artifacts visible. Doing it before upsizing also usually kills the "per-pixel" sharpness you just gained (I'm using the "per-pixel" term loosely). In the end you have more control and effectiveness doing it as a final step.

The sharpening you apply should also vary according to the medium you print, I usually overdo it a bit for print, while for web I don't excede 70-80% 0.3 radius in USM.

I see. Sorta. I put the RAW into GIMP via the Ufraw plug-in and sharpen it a tad (if I think it needs it). If I think it's sharp enough I keep it in Ufraw and then just export it from there. This way I'm not sharpening a jpg, I'm sharpening at the source. There are no jpg artifacts to make pronounced in sharpening because they aren't there yet.

Am I high?

 

[PatBranch]

Maybe a little. I'm pretty sure jpegs need sharpening (if you are exporting to that sometime, for web, for print ?)

 

[PatBranch]

not sure how that resize/sharpen is just for web; i've used it on full-size images and the prints look fine.

not sure if the algorithm is sensitive to different image sizes; check it out and experiment. i've not adjusted any of the parameters based on final output size, but there probably would be a difference (even if it's slight).

i don't fret about it too much though; i am just happy to find an action which works well, adn i can batch a bunch of files and be done w/ it.

It said it's for web images.

edit: Did you just take out or replace the 600px resize, for printing?

 

[narlus]

i put in 3502 in both pixels field (the size an uncropped, full-res image is out of my camera).

 

[PatBranch]

I emailed him about changing the size.

Patrick Branch to Manyk Photography
[I]http://news.deviantart.com/article/20250/[/I]
Would changing the resize too 500px affect the sharpening?

Manyk Photography to me
The only thing you might want to change is how many steps of USM are applied at the end. Instead of the default 3 steps, you might want only one or two. But, in terms of the overall action, you should be fine setting the size to 500.

Patrick Branch to Manyk Photography
Okay, thanks! So...about only doing one or two finshing USM steps....could/would the smaller size be over-sharpened with those 3 steps?

Manyk Photography to me
Well, anything COULD be oversharpened ... but the sharpening is on a layer that can be adjusted. I would suggest trying it first, then setting the sharpening layer to 100%, if you see bright lines or halos around edges, then that is too much.

And if you do see halos, you could redo the action to reduce the last 3 USM steps to 2 ... or even one. Then set the sharpening layer to 100% again and see if that helps by removing 1 or 2 USM steps.

Or ... completely separate the last 3 USM steps to another action on their own. And then fiddle. That way you have the entire reducing action minus the last steps. And you can experiment with the last 3 on their own instead of having to run the entire action over and over.

 

[narlus]

nice one, thanks!

 

[binary visions]

I see. Sorta. I put the RAW into GIMP via the Ufraw plug-in and sharpen it a tad (if I think it needs it). If I think it's sharp enough I keep it in Ufraw and then just export it from there. This way I'm not sharpening a jpg, I'm sharpening at the source. There are no jpg artifacts to make pronounced in sharpening because they aren't there yet.

Am I high?

Not high, but there are two reasons to sharpen.

One is at the beginning of the workflow, you might sharpen to counteract the effects of the anti-alias filter in your camera. The AA filter reduces sharpness a little to help counteract things like moire. That should be a pretty light sharpening.

The final sharpening, which is usually significantly more aggressive, and is usually the one people are talking about when they ask "did you sharpen that?" should be done immediately before the presentation of the image (displaying online, printing, whatever), and AFTER all the resizing has been done. While it's true that you shouldn't sharpen compressed images, up until the presentation of the image online, you really shouldn't have done any significant compression to it. If you're working with it in an uncompressed file type (or, frankly, even a JPG at the highest image quality setting), you won't have introduced many/any compression artifacts.

Resizing your image reduces the sharpness, so sharpening before you resize is only partially useful. Most images still benefit from a final sharpening immediately before presentation.

 

[Strakar]

I see. Sorta. I put the RAW into GIMP via the Ufraw plug-in and sharpen it a tad (if I think it needs it). If I think it's sharp enough I keep it in Ufraw and then just export it from there. This way I'm not sharpening a jpg, I'm sharpening at the source. There are no jpg artifacts to make pronounced in sharpening because they aren't there yet.

Am I high?

I understand where you're coming from, but the problem is that no matter how you handle the data, the image recorded has already stuff that doesn't belong in the photo. The thing is, there are other artifacts besides those caused by jpeg compression, straight out of the camera you have noise, demosaicing artifacts and might even get to see moire and purple fringing. All this image defects accentuated through sharpening.

But even if originally you had the perfect photo, without any deffects, there would be another problem. If you sharpen before upsizing, the "crispness" you were after by sharpening the photo will be lost after the upsizing. An example:

- Original photo, you have two shades of gray side-by-side; let's admit that ideally all pixels inside each shade are the exact same color.

- You sharpen the photo at the original size, and again ideally, the two shades of gray get separated by two lines of pixels (one darker and one lighter). Your eye does not identify the new line and the image looks "crisp";

- You upsize the image. The 2 pixel line that before looked crisp will be upsized to a much softer gradation of shades in the upscaled image (there are severall algorithms for upscaling, but this will always be tough), making the original sharpening less efficient, and loosing the "crispness" effect.

You not only loose the sharpness of the image but also every sharpening halo gets amplified too, and therefore degrades somewhat the image.

 

[H8R]

Ok. I get it. I'm high.