[rizal2003]

Hi All,

Just want to seek for an advise here... What is the correct method to unsharp masking a photo? Do we do it with a full sized 3-4 MP image or to a down-sized image (640X480).

Rizal

[IanF]

I always sharpen a photo as the very last stage when working on a photo.

I crop --> resize (640x480 or 80x600 usually) --> adjust levels or whatever and then sharpen by whatever means and save a copy.

Then for web use I optimise the file using 'save for web' or using NeatImage and save an optimised copy.

[Alastair Rae]

Originally Posted by IanF

I always sharpen a photo as the very last stage when working on a photo.

I crop --> resize (640x480 or 80x600 usually) --> adjust levels or whatever and then sharpen by whatever means and save a copy.

Then for web use I optimise the file using 'save for web' or using NeatImage and save an optimised copy.

It does seem counter-intuitive to process with less data rather than more. I must give it a try it that way around and compare.

[scampo]

Originally Posted by rizal2003

Hi All,

Just want to seek for an advise here... What is the correct method to unsharp masking a photo? Do we do it with a full sized 3-4 MP image or to a down-sized image (640X480).

Rizal

Some cameras allow in-camera sharpening and the evidence is that this is well worth setting it to "high". Also, it is usually recommended to leave sharpening till last if you use PC software to do this. Have you looked at any of the sharpening software available? I have tried FotoWiz FocalBlade and Colorwasher and find they do a very good job exceptionally easily. When I have the cash, I'll be buying both! If you search in Google or Yahoo you'll find more details.

[IanF]

Of late I have actually been setting the in-camera sharpening to low (=no) in camera sharpening and using the computer software for all sharpening. On the possibly misplaced premise that the computer software is likely to be more sophisticated. I'm happy enough with the results I am getting of doing things that way, but uncertain if any real advantage is gained.

[scampo]

I did the same until I read a posting, with the usual first class explanation, from Jay. It might have been on the Yahoo forum, though! He felt that sharpening before aplying jpeg compression was likely to be better - and I think he has some "before and after" photos on his digiscoping website.

[IanF]

Thanks for that Steve, I may just swap between the settings next chance I get for a stationary subject so as to compare

[rizal2003]

Guys,

I often sharpen the photo twice, once in full size (3MP-Nikon 990) and second one in 640X480 and found the result is better.

Rizal

[IanF]

Thanks for that as well Rizal. I just gave that a go and I thinkn the double sharpened images as you describe look less grainy than the once sharpened ones. I think it needs more experimentation though.

[Tockman]

Hi Guys,
I'm new at Digiscoping,but have used Paint Shop Pro for some time, I have found it better to use unsharp mask etc,on only selected parts of a photo,(feathers etc) rather than the whole photo,any graininess that might effect the subject is usually lost in the feathers!

-------------------
Alan.

[IanF]

I have PSP too, though to be honest I hate fiddling with photos over much and I find the tool quite fiddly to use though the freehand tool though is great for working on selected areas. Normally though I prefer just to re-size and sharpen the whole photo.

[Ragna]

Ive tried sharpening twice first at 1200 x800 and using unsharp mask at about 100-120 and then resizing again to 1200x800 and using unsharp mask again,I ve tried this several times and compared this with using unsharp mask at full size at 200, sometimes it made a very slight difference sometimes i could see no difference.I also only sharpen the bird as the grain does'nt notice so much.

[rizal2003]

Ragna,

It needs a lot of patience to mask/highlight a bird only. How did you do it? Do you use the magic wand or manually masking?

Rizal

[Jay Turberville]

Originally Posted by scampo

I did the same until I read a posting, with the usual first class explanation, from Jay. It might have been on the Yahoo forum, though! He felt that sharpening before aplying jpeg compression was likely to be better - and I think he has some "before and after" photos on his digiscoping website.

Yeah - that was probably me. The topic had come up before and I did what I am amazed that more people don't do. I just did a simple test. Its easy with digital and you get quick feedback. The tests were done with my Coolpix 5000.

My results were conclusive for having sharpening set to HIGH. My thinking as to why goes along these lines.

1) The camera manufacturers provide different sharpening settings with the idea that they will be used with the camera lens alone. The "normal" setting for sharpening on the Coolpix line is just one step below the "high" setting. On other cameras, a "0" setting represents the application of some level of sharpening. So camera makers expect that some sharpening applied to the image is most often going to be a good thing. Now consider that no digiscoping setup is as sharp as the camera lens alone. So it seems unlikely that you could oversharpen the slightly soft digiscoping image. At the very least, "normal" sharpening would seem appropriate.

2) Most digiscopers save in JPEG format. JPEG introduces small artifacts. Sharpening tends to accentuate small details. So sharpening after the JPEG is stored might compound the noise by accentuating JPEG artifacts.

3) In camera sharpening may have a technical advantage since it is being applied to either RAW or uncompressed RGB data.

Theories are nice, but tests tell us what really matters. These are results from my tests (using Phred the Wonder Pheasant - 1).

http://www.jayandwanda.com/digiscope/sharpening1.jpg
http://www.jayandwanda.com/digiscope/sharpening2.jpg
http://www.jayandwanda.com/digiscope/sharpening3.jpg
http://www.jayandwanda.com/digiscope/sharpening4.jpg

And a super close-up of how sharpening affects the stuck pixel (actually a group of 4) on my CP5000. I include it because it really shows the noise increase which becomes a bigger issue with smoother out of focus areas than it is when looking at feather details and such.

http://www.jayandwanda.com/digiscope/hotpixelsharp.jpg

But don't use my tests as gospel. These tests are mostly a good justification to consider using in-camera sharpening. But to really know, you need to test your own equipment. Different scopes, cameras and eyepieces may yield different results. All you need is a feather and a little bit of time.

[Jay Turberville]

Originally Posted by rizal2003

Just want to seek for an advise here... What is the correct method to unsharp masking a photo? Do we do it with a full sized 3-4 MP image or to a down-sized image (640X480).

It probably doesn't matter that much. But I alway work on an image at full resolution so that I don't have to come back and repeat the work if I want a print or need a higher resolution version for some reason. Once done, I resize by exactly one half resolution. I then sharpen it slightly to compensate for the softness that introduces. I then resize to the web-sized image and do another sharpening. Each subsequent sharpening is progressively less intense.

But frankly, I've also turned around and resized in one step and then sharpened and the results aren't that much different. These things are easy to test and compare for yourself. And doing this kind of testing for yourself is what will provide the best answer.

[IanF]

Thanks for the detailed reply Jay

Your results are certainly worth exploring further. I'll do a bit more experimentation to see what works best for my set up.