Mike Chaney's Tech Corner
November 23, 2024, 05:20:32 AM *
Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
News: Qimage registration expired? New lifetime licenses are only $59.99!
 
   Home   Help Login Register  
Pages: 1 [2]
  Print  
Author Topic: Tone Targeted Sharpening for JPGs  (Read 30538 times)
Terry-M
The Honourable Metric Mann
Forum Superhero
*****
Posts: 3251



WWW
« Reply #15 on: August 18, 2010, 06:37:41 AM »

Quote
I was just speculating that perhaps TTS was a bit simpler to implement than a bunch of selection tools.  Help me out, here.
Only Mike can answer that but in view of the fact the UI is simple, it is possibly the case.  Huh?
Terry
Logged
Terry-M
The Honourable Metric Mann
Forum Superhero
*****
Posts: 3251



WWW
« Reply #16 on: August 18, 2010, 07:12:26 AM »

Hi Box Brownie,
Quote
The examples posted are a wide range and in some cases look too strong in sharpening effect as I see artefacts but as I read it there are slider controls that can yield the right level for the subject(s).
Tell us which ones you think are overdone and we'll make an excuse of modify them  Grin
I'm still working my way up the learning curve with TTS and, to some extent, the samples shown were part of the leaning process, so not all of them will be "just right"  Roll Eyes
Yes, there is a slider but it's related to the selection of tones. At 100% there is maximum discrimination for the selected tone or, in the case of the exception modes, maximum exclusion.
In the case of a selected tone reducing the slider reduces the discrimination and a wider tone range is sharpened. The radius and % are entered in the normal Qimage way: up-down counter or directly.

About work flow: the jpegs were straight out of camera which had sharpening turned off, possibly some levels or curve adjustment and then TTS applied, all in the Q-U editor.
The raw samples were processed in Q-U, which includes some USM with camera EQ, and all adjustments made in the Q-U editor with TTS as the final stage. NB. Qimage applies edit filters in a sequence with sharpening last.
If you are preparing an image in an external editor, I would expect there to be no USM applied there but left to Q-U with TTS. I don't make any special adjustments for web images. There are options in the interpolation preferences for Anti-Aliasing with downsized images; I leave that at "Low".
Fred will  possibly add his two pennyworth too. Wink
Quote
I hope you are able to offer this insight in the aim that it might shorten my initial learning appreciation of the power of TTS
Just plough through a variety of images, try different TTS modes and values and enjoy the results, you are likely to see your images as never before  Grin
Terry

Logged
Fred A
Forum Superhero
*****
Posts: 5644



WWW Email
« Reply #17 on: August 18, 2010, 09:25:07 AM »

Hi Box Brownie,
I did want to comment about you mentioning about some sample images being a little over sharpened.
You are absolutely correct about some. 
I found, and terry did too, that the impact (wow factor) diminished as the screen image was reduced in size.
The 800 x 600 size which is acceptable for posting images to the Pbase screens seem to feel softer than they were at hi rez.
So for the sake of demonstrating, some were a little over the top in order to show the SELECTIVE capability; the ability to sharpen one item and not others.

I wouldn't dare print some of them like that.

Did you notice though, on that TTS web page, clicking on the JPG folder, that I posted three before and after sets?
If you select one of those images bringing it to 800 x 600, and then ticking the PREVIOUS and the NEXT words, back and forth, you can get a really good look at "A" vs "B".

I used to have a box Brownie camera a long time ago.

Enjoy!

Fred
Logged
mburke
Jr. Member
**
Posts: 61



« Reply #18 on: August 18, 2010, 11:59:42 AM »

This might be an obvious question, but can you apply the tts more than once? In other words on a building on the left side and a park bench on the right, possibly in two different edit steps?

Mike
Logged
Terry-M
The Honourable Metric Mann
Forum Superhero
*****
Posts: 3251



WWW
« Reply #19 on: August 18, 2010, 12:21:01 PM »

Hi Mike,
Quote
his might be an obvious question, but can you apply the tts more than once?
No is the short answer. The only way to do that would be to do the first TTS and then save-as  the filtered image. The second TTS would then be applied to the new image.

I had wondered about that early on when I wanted to soften a background (use a - %) in addition to sharpening something else. In reality I have found few cases where it would be of use and the existing options can be used creatively to deal with a high proportion of images.
Terry
Logged
Terry-M
The Honourable Metric Mann
Forum Superhero
*****
Posts: 3251



WWW
« Reply #20 on: August 18, 2010, 01:01:15 PM »

Quote
select one of those images bringing it to 800 x 600, and then ticking the PREVIOUS and the NEXT words, back and forth, you can get a really good look at "A" vs "B".
I thought I'd do that with one of my recent raw images that has a very busy background, make sure Original Size is clicked below the image.
Before TTS: http://www.pbase.com/tjm04/image/127579841
After TTS: http://www.pbase.com/tjm04/image/127579842
Exactly the same radius & percent settings were used for each. The TTS version used Except Tone and a colour in the background, significantly different for the subjects, was used. The small image on PBase does not do full justice to the TTS 3D effect, but you should get the idea, click Previous & Next a few times.
Terry
Logged
Owen Glendower
Full Member
***
Posts: 185


« Reply #21 on: August 18, 2010, 01:39:52 PM »

Thanks, Terry, good example.  The difference in the white helmet is very apparent, but what's also significant upon closer examination is how TTS reveals the fine texture in the black jacket.  Toggle back & forth while looking closely at that area.  Impressive.
Logged
admin
Administrator
Forum Superhero
*****
Posts: 4220



Email
« Reply #22 on: August 18, 2010, 01:50:47 PM »

Just a quick note on topic here.  I developed TTS because I hate having to use lasso tools or "magic wands" to select areas, and when you use that selection method to sharpen, the results look "fake" to my eyes.  In the strawberry shot on the TTS page for example, you could have selected the strawberry and sharpened it but the effect really looks "photochopped" especially near the edge of the subject where the sharpening is going from heavy to none near the edges.  No matter how much you feather that, it's usually noticeable.  With TTS, you get to select the tones you want to sharpen and the effect naturally rolls off as tones diverge from what was chosen.  This not only takes care of edge anomalies but it also allows a sharpening curve within the selected area as tones shift around on your subject.  If you lasso a particular subject, it likely changes tone differently throughout and at a different rate at different places around the edges which TTS can account for and manual selection cannot.  And of course... clicking once with a dropper is a lot easier than using a lasso or magic wand.

Mike
Logged
Terry-M
The Honourable Metric Mann
Forum Superhero
*****
Posts: 3251



WWW
« Reply #23 on: August 18, 2010, 03:08:44 PM »

Owen,
Quote
The difference in the white helmet is very apparent
I hope you meant/see beige not white, the shirt on the girl on the right is white  Huh?
Terry
PS. note the hair wisps on the right, hair under the helmet, face etc. - litlle details that go towards the 3D effect.
Terry
« Last Edit: August 18, 2010, 03:13:16 PM by Terry-M » Logged
Owen Glendower
Full Member
***
Posts: 185


« Reply #24 on: August 18, 2010, 03:14:38 PM »

Yep, it's clearly beige, sorry.  I noticed the wisps of hair, too.  As you say, a small detail which has an important effect.  And the photo itself is a textbook example of an image which should NOT have overall sharpening applied.  Strange things can happen to the nicely-out-of-focus background when that's done.  Thanks again for posting it.
« Last Edit: August 18, 2010, 03:21:12 PM by Owen Glendower » Logged
Box Brownie
Newbie
*
Posts: 48


« Reply #25 on: August 18, 2010, 07:01:42 PM »

Thanks for the feedback guys  Smiley

Fred, ref my nick  Cheesy It may not have been a "Box Brownie" but my first camera, well my fathers, was a box camera and a great joy & wonder it was to take pictures with in the 1950's taking the film to the local chemist (drugstore to you cousins across the pond) and getting back the B&W pictures of family events.  I have had a camera of one type or another in my hands ever since  Wink  So the nickname came quite naturally?

Mike, yes agree entirely with the sharpening philosophy you have managed to create with TTS.  I am a real idiot when it comes to layers & masks as such my PPing tends towards (I hope) carefull usage of capture sharpening using USM LCE settings and for web after resizing Smart Sharpening.  Overall this gives me the pleasing results I like to see.  But the way TTS appears to make subject isolation imitating a narrower DoF but still ensuring the whole of the primary subject is in sharp focus looks like a wonder  Cool

TTS (and I will get on and check it out by the weekend) in a way makes me think of the way the eye works ~ we distinguish distance and subject isolation in part I think by the tonal drop off between nearer/further subjects ~ may be wrong thinking but Huh?

The pith helmeted subject and partner make the TTS point quite nicely ~ but and sorry if I missed it are you resizing to web size before or after applying the TTS because I thought with conventional PPing the received wisdom is to make any final output sharpening after resizing and based on the feedback above TTS is in effect a step sharpening, isn't it???

For the record I always have my in camera settings to the very lowest sharpening & contrast, the others I leave at default midrange, I like to control these parameters in PPing.

TIA for any further thoughts & feedback.
Logged
Terry-M
The Honourable Metric Mann
Forum Superhero
*****
Posts: 3251



WWW
« Reply #26 on: August 18, 2010, 07:14:19 PM »

Hi Box Brownie
Quote
but and sorry if I missed it are you resizing to web size before or after applying the TTS because I thought with conventional PPing the received wisdom is to make any final output sharpening after resizing
The TTS was applied to the raw file and the Q-U used to make web/e-mail copies directly from the raw. I don't know about the "received wisdom" but I like the simple life an some received wisdom can be a myth, like always re-sizing to 300ppi for printing.
Mike would be the authority on sharpening after downsizing and whether it makes a significant difference, but you would need to make a downsized tif first and then a jpeg for web, otherwise you be losing something doing jpeg to jpeg. A lot of effort for what gain? You could do some tests  Wink
Terry
Logged
Terry-M
The Honourable Metric Mann
Forum Superhero
*****
Posts: 3251



WWW
« Reply #27 on: August 18, 2010, 07:36:36 PM »

Quote
The pith helmeted subject and partner make the TTS point quite nicely
Just an idea for Qimage owners (Pro or SE)
You can download my images if you like and use Qimage Comparator with the Swop function (CTRL S).
That shows the differences far more clearly than the PBase view; I don't really know why. Make sure it's the "original" size, 800px wide.
Terry
Logged
Owen Glendower
Full Member
***
Posts: 185


« Reply #28 on: August 21, 2010, 01:28:41 AM »

Just a quick note on topic here.  I developed TTS because I hate having to use lasso tools or "magic wands" to select areas, and when you use that selection method to sharpen, the results look "fake" to my eyes.  In the strawberry shot on the TTS page for example, you could have selected the strawberry and sharpened it but the effect really looks "photochopped" especially near the edge of the subject where the sharpening is going from heavy to none near the edges.  No matter how much you feather that, it's usually noticeable.  With TTS, you get to select the tones you want to sharpen and the effect naturally rolls off as tones diverge from what was chosen.  This not only takes care of edge anomalies but it also allows a sharpening curve within the selected area as tones shift around on your subject.  If you lasso a particular subject, it likely changes tone differently throughout and at a different rate at different places around the edges which TTS can account for and manual selection cannot.  And of course... clicking once with a dropper is a lot easier than using a lasso or magic wand.

Mike
Appreciate it, Mike.  It does indeed help to know the "innards," as you say elsewhere.  And you're certainly right that sharpening a selected area can result in something which just doesn't look quite right.  About the only time I feel forced to use a selection tool is when it's necessary to apply fill light only to a specific area, like a heavily-shadowed recessed window...a situation where overall fill light doesn't work, in other words.  Not that Fill Light in Qimage hasn't saved my bacon a couple of times!  Occasionally, though, it needs to be selective, like sharpening.
Logged
Pages: 1 [2]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Valid XHTML 1.0! Valid CSS!
Security updates 2022 by ddisoftware, Inc.