Unsharp Mask?

[CaseyScofield]

Just wonder'n...

For you 20D owners, in Photoshop, do any of you have a standard Amount/Radius/Threshold (Unsharp Mask) that you use on your photo's? I know it depends a lot on the photo, but I find myself adjusting it different ways each time...didn't know if there was a standard that alot of u use.

[Hellashot]

USM isn't camera dependent. So any camera users can chime in.

[PacAce]

USM isn't camera dependent. So any camera users can chime in.
Actually, it is to a point in that different cameras have different resolution sizes and that, in the end, will affect the amount of USM you can or should apply to an image. So anybody who's used to applying USM to a 6.3 MP image will soon find that the same USM values on a 8.2 MP image does not achieve the same degree of shapness as previously, with everything else being the same.

[Dubsta]

A good starting point is 300, 0.3 and 0.

[pcasciola]

A good starting point is 300, 0.3 and 0.As starting points, I generally use 300/0.3/0 as well. For shots that appear a little softer I start with 140/0.8/0.

[PacAce]

Yes, same here, but I use that just to pre-sharpen the image after conversion from RAW but before actually editing the file. Once I'm done editing the file and have resized the image to what I need, either for printing or for the web, I'll sharpen again for output.

[maderito]

I shoot a lot of available light, high ISO images. After applying noise reduction, there still remains noise, especially in the shadows. I've found that "pre-sharpening" sometimes exaggerates this residual noise.

Thus for high ISO shots, I complete the editing before applying any sharpening. I too will start with 300/0.3/0, but if noise becomes more prominent, I'll switch to alternative forms of sharpening, especially low contrast enhancement (USM with low amount and high radius, e.g. amount=10 to 20 and radius=50). That said, a well-exposed high ISO image not dominated by shadow tones responds well to standard sharpening approaches after noise reduction.

[pcasciola]

maderito,

I also shoot a lot of high ISO available light, and I'm really impressed with the new Noise Ninja plug in for Photoshop, which does USM along with it's noise reduction. I find being able to preview my noise reduction levels with USM to be a huge advantage, and it can do it all in 16-bit aRGB as well which is another big plus. It was around $80, but I feel like that was money well spent. Have you tried it yet?

[maderito]

maderito,

I also shoot a lot of high ISO available light, and I'm really impressed with the new Noise Ninja plug-in for Photoshop, which does USM along with it's noise reduction. I find being able to preview my noise reduction levels with USM to be a huge advantage, and it can do it all in 16-bit aRGB as well which is another big plus. It was around $80, but I feel like that was money well spent. Have you tried it yet?
Don't mean to hijack this thread into a different direction - but this point is worth discussing (at least to me :) ).

I've used the Noise Ninja and Neat Image plug-ins for Photoshop. I long ago convinced myself that there is no difference between the two products if you tweak noise profiles and output settings properly. So I've stuck with NI because I understand its underlying theory and application controls better.

My latest workflow for a high ISO image includes an initial and final run through NI, the first for noise reduction and the last for sharpening after all other image edits are completed. The noise profile has to be tailored for each pass. I agree that the sharpening in the NI/NN programs is excellent since you can independently judge effects on noise reduction and sharpening and find a proper balance between the two. The real beauty is that the nature of the technique (the underlying wavelet processing algorithms) permits noise reduction AND sharpening as relatively non-competing endpoints.

I choose to do the sharpening at the end to control the appearance of the final image. If I were working against a deadline with a lot of files, I would and do proceed as you suggest.

[robertwgross]

My latest workflow for a high ISO image includes an initial and final run through NI, the first for noise reduction and the last for sharpening after all other image edits are completed. The noise profile has to be tailored for each pass.

I think I agree completely, but I'm not sure why this has to be the case.

I've found (with Neat Image) that I have to generate a noise profile for each shot. What seems weird to me is that I can have three almost-identical shots, all taken together of nearly the same scene, and the noise profiles are fairly different. I can't argue about the results, but I'm not clear how it gets there.

Also, results and methods seem to be completely different depending on whether the original input image is digital or scanned from Velvia. Digital color noise is different from film grain, I guess.

---Bob Gross---

[Hellashot]

It also depends on what you want to do with the image. Is it for: onscreen viewing only? 4x6 printing? 8x10 printing?

I only save my RAW files on my computer. If I want to do something with a JPG or a TIFF image I convert it then sharpen it to what I am doing.

For onscreen viewing you usually sharpen it modestly. If you want to print for 4x6 you want to oversharpen it because of the small size you are printing to. If you are printing an 8x10 or bigger you want to sharpen less because you'll end up with it looking more the size it is at full size on your computer.

[CaseyScofield]

Thanks all for your valuable insight! Always learning...:D