Luminous Landscape Forum

Raw & Post Processing, Printing => Digital Image Processing => Topic started by: undavide on March 05, 2012, 05:57:37 am

Title: 500% Amount in USM
Post by: undavide on March 05, 2012, 05:57:37 am
Hello,
does someone know why the Amount in the Photoshop's Unsharp Mask filter (expressed in percentage) goes up to 500%?
Is there any reason to exceed the 100%?
Mind you, I'm not asking whether it should be wise or not to apply more than 100% USM in a picture - I'm wondering why the Unsharp Mask filter has been engineered by Adobe developers with such a ceiling value.
Thanks in advance

Davide Barranca
www.davidebarranca.com
Title: Re: 500% Amount in USM
Post by: ErikKaffehr on March 05, 2012, 06:45:48 am
Hi,

Sharpening with large amount is meaningful with small radius. If I recall it correctly it was suggested in a example in the original "Real World Image Sharpening book by Bruze Fraser and Jeff Schewe that high amount 500) would be used with very small radius to compensate for OLP filtering on early Canon cameras.

Best regards
Erik


Hello,
does someone know why the Amount in the Photoshop's Unsharp Mask filter (expressed in percentage) goes up to 500%?
Is there any reason to exceed the 100%?
Mind you, I'm not asking whether it should be wise or not to apply more than 100% USM in a picture - I'm wondering why the Unsharp Mask filter has been engineered by Adobe developers with such a ceiling value.
Thanks in advance

Davide Barranca
www.davidebarranca.com
Title: Re: 500% Amount in USM
Post by: undavide on March 05, 2012, 06:57:34 am
Hi Erik,
thanks for the answer but I'm afraid you've misunderstood my question:

Quote from: undavide
Mind you, I'm not asking whether it should be wise or not to apply more than 100% USM in a picture - I'm wondering why the Unsharp Mask filter has been engineered by Adobe developers with such a ceiling value.

I mean, the 500 ceiling is arbitrary? If the unit of measure is in percentage, why is that the maximum allowed value is not 100%?
My question in this case is from a developer point of view if you will, not from a retoucher one.
For the Radius parameter, this is clearly expressed in pixels as the radius of the Gaussian blur used to compute the difference (Unsharp Mask needs a blurring kernel, it works that way). It's just not clear to me why the Amount range goes up to 500%
Thanks anyway Erik!

Davide
www.davidebarranca.com
Title: Re: 500% Amount in USM
Post by: Nigel Johnson on March 05, 2012, 10:22:24 am
I mean, the 500 ceiling is arbitrary? If the unit of measure is in percentage, why is that the maximum allowed value is not 100%?
My question in this case is from a developer point of view if you will, not from a retoucher one.
For the Radius parameter, this is clearly expressed in pixels as the radius of the Gaussian blur used to compute the difference (Unsharp Mask needs a blurring kernel, it works that way). It's just not clear to me why the Amount range goes up to 500%

Davide

I haven't got my copy of 'Real World Image Sharpening' handy, however as far as I remember 100% equates to applying an additional amount of contrast to the edge that is the same as the unsharpened edge contrast and values greater than 100% apply higher amounts. Obviously when a small radius is used the actual amount of brightness change applied to a given pixel will be proportionally less hence the recommendation mentioned by Erik.

Regards
Nigel
Title: Re: 500% Amount in USM
Post by: AlanPezzulich on March 07, 2012, 06:14:35 pm
The high numbers allow you to use the unsharp mask for local contrast adjustments. Raise both sharpness and radius for this effect.

Title: Re: 500% Amount in USM
Post by: hjulenissen on March 08, 2012, 05:49:47 am
If a sharpening filter is a highpass filter, 0% and 100% could correspond to infinite attenuation and unity gain. Then those percentages would have a sense in a dsp-manner. I dont know if that is the case.

-h
Title: Re: 500% Amount in USM
Post by: smahn on March 23, 2012, 05:36:28 pm
USM that only goes to 100 would be like an amp that only goes to 10.