Best USM Tool I've seen

Mastrianni · Nov 24, 2002

Jan Esmann said:
Jan Esmann said:

Editing in non-linear working-space embeds various kind of errors
to the images, e.g from USM you get the white halo.

Click to expand...

This is not true. Regular USM constructs a more blurred image and
subtracts that from the original you want to sharpen - and this is
irrespective of wether the space is linear or not. The halos are
created because the constructed unsharps mask (the more unsharp
image to subtract from the original) is just that - more unsharp
and therefore subtracts to wide blurs from the original and creates
negative edgelines. Now the math involved in this - and avoiding it

has nothing to do with the linearity of the working space what so

ever. It has something to do with how light spreads in relation to
what colour it originates from, what intensity of light is on it
and so forth. Hence a sharpener that is to avoid edgelines has to
adjust its algorith to all those factors pixel for pixel. It has
nothing to do with the working space, but everything to do with
optics and transpating analog optical algorithms to digital image
processing.

Regards
Jan Esmann

Jan,

My sentiments, exactly. Transferring linearly or non-linearly, has no effect on final output. The different gamuts only correlate to different levels of saturation in cetain colors, but do not make "non quality" or bad output. But, editing in 16 bit, USM or other edits, does make a difference, both theoretically and visually.(as opposed to 8 bit) But, again, even this is dependant on the image and "how much" editing is done.
Which was my point,.... not that the "theory" applied to test charts is wrong.

I will give PowerRetouche a try, and see how it looks on final output of 'real world' imaging, both inkjet and 150lpi CMYK output.

Thanks
Mastrianni

Timo Autiokari · Nov 24, 2002

Jan Esmann said:
Jan Esmann said:

Editing in non-linear working-space embeds various kind of errors
to the images, e.g from USM you get the white halo.

Click to expand...

Jan Esmann said:
This is not true.

Is true and accurate, 100%.

Jan Esmann said:
Regular USM constructs a more blurred image

That is correct. And blurring is the same as averaging neighboring pixels. Summing, divisions etc only give you correct answer in the linear domain.

Jan Esmann said:
and subtracts that from the original you want to sharpen -
and this is irrespective of wether the space is linear or not.

Not true.

Jan Esmann said:
The halos are created because the constructed unsharps
mask (the more unsharp image to subtract from the original)
is just that - more unsharp and therefore subtracts to wide
blurs from the original and creates negative edgelines.

No, in the non-linear working space the sharpened edges are visible as white halos because the white border of the edge gets way too much enhancement where the dark border of the edge gets only a little, due to the distortion by the gamma. In the linear space both the sides get equal amount of enhancement so we only see prefectly sharpened image with no side-effects.

Jan Esmann said:
Now the math involved in this - and avoiding it - has
nothing to do with the linearity of the working space
what so ever.

Sorry but you are totally wrong in this. Very simple test will demonstrate this easily no matter if it is a photograph of a calibration target or any other real life scene.

Timo Autiokari http://www.aim-dtp.net

Pete Stone · Nov 24, 2002

The best thing to do is download the demo, and try it yourself!
Pete Stone

Rylee Isitt said:
Is post-production sharpenning that important? I've seen some
reviews, of the 1ds for example, which show that the full frame
image lacks sharpness (aparently focused properly). Is it the
interpolation the camera does to get a full R, G, and B value for
each pixel what causes this slight blur in the images?

I don't have a digital camera, but am in the process of learning as
much as I can before I purchase one. I want to make an educated
buy, not base my choice off of sales pitches and the number of
megapixels (at least not JUST the number of megapixels).

Pete Stone said:

Of all the USM methodss out there I've seen ( and I think I've
tried them all! ) hands down the Power Retouche Sharpness filter
looks the best. It's the only sharpening filter that gives
increased sharpness WITHOUT any edge halo effect ( using their "fix
edges" option set to 80% ). You can download a trial version from
their website....let me know what you think!
--
pete stone
http://www.petestone.com

Click to expand...

--
pete stone

Robert Karpa · Nov 24, 2002

I have to vouch for this program. I'm very critical when it comes to the quality of my digital files. When I tested this I was blown away by the results. Far bettter than in camera sharpening or unsharp mask in Photoshop. I've never seen anything like it. The rest of the filters supplied with the program are real time savers and they have a lot of interesting creative applications as well.
Robert Karpa

Jan Esmann said:
I couldent help following these threads and i would like to
comment, that the images you use for testing USM's are not images
that reveal the powers or flaws of sharpeners. Grashoppergreen
against leafgreen wont reveal anything. You have to use an edge of
high contrast like a black hair or eyelashes against pale skin.
Black lines against white wont do either. You need black or dark
coloured lines against a or pale non-complementary colour. The
halos are always negative with respect to density and colour.

I have tested all the sharpeners discussed here with such an image
and i herewith challenge any one of them that they all produce
halos along high contrast edges and that my Powerretouche sharpness
Editor does not. Or, if the others avaoid halos, they dont sharpen
enough. This includes Vtie's otherwise nice software.

The PowerRetouche sharpener has two unique sharpening algorithms
that emmulate optics and of these one is an adaption of regular USM
with a halo correction function, the other is a completely new
concept that has an inbuilt intelligent self-altering algorithm,
that adjust the sharpening algorithm according to several optical
factors culled from the image at large and from the particular area
being sharpened. How and what is for obvious reasons secret, but it
involves artificial intelligence.

For this reason the filtering is timeconsuming (processing times
are not secret and there are clear tables of processing time on the
PR sharpener website), but i assure you it can not be done faster
if you want sharpening of that quality. Of course you can
compromise and go for less than perfect. Personally i prefer
perfection and hope the PowerRetouches are nothing less.

I wrote them because i was # !¤# fed up with the necessety of
tweaking and twisting PS in order to minimize inevitable halos,
highlight overburning, shadow noise, etc. etc when i retouched
scans of my paintings (I am an artist by profession and i have been
retouching with PS for professional printing for at least ten years
now).

Try the sharpeners with this kind of image and your will see clear
pros and cons.

Regards
Jan Esmann
. developer of PowerRetouche

--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com
--------------------------------------------------------------

Jan Esmann · Nov 24, 2002

Yous asked for a comparison of the various sharpeners. Here's the research i did while developing the Sharpness Editor.

http://www.powerretouche.com/SharpenersTested.htm

Not that i want to denigrade others, but i want to back my claims that the Powerretouche sharpener is the only non destructive sharpener.

Regards
Jan Esmann
--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com

Jan Esmann · Nov 24, 2002

Timo...

Dear me...

Well. The proof of the pudding lies in the eating. Check out the comparisons of your AIM "Better" USM and other sharpeners compared to PowerRetouche Sharpness Editor.
http://www.powerretouche.com/SharpenersTested.htm

You say ". And blurring is the same as averaging neighboring

Timo Autiokari said:
pixels". In my understanding thats a smoothing operation, not a blur.

I agree with your comments on the gamma problems regarding linearity vs nonlinearity.

I also partly agree on your comments on how halos are created. I mere did not want to be technical.

Also I wont go into more, because i dont want to tell how to get rid of the halos, but check out above link and see the pudding for yourself.

Thanks for the corrections

--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com

Rich Gibson · Nov 24, 2002

For what it's worth that expression has become trite and overused to an extreme. Having lurked and participated in Steves-Digicams and DPreview since 1998 I haven't seen any case where competing products "blow away" anything. I've been following all these sharpening and own five of them: Photoshop, Nik Pro, FredMiranda, UltraShrapen, and VTie. None blows away any. At best they are marginally better under certain circumstances. The same with the comparisons of S2, D60 and D100.

Once a week someone "discovers" one of them and posts a miraculous new sharpener which, guess what? "BLOWS AWAY" everything else.

Please!

Rich

Robert Karpa said:
Jan Esmann said:

I couldent help following these threads and i would like to
comment, that the images you use for testing USM's are not images
that reveal the powers or flaws of sharpeners. Grashoppergreen
against leafgreen wont reveal anything. You have to use an edge of
high contrast like a black hair or eyelashes against pale skin.
Black lines against white wont do either. You need black or dark
coloured lines against a or pale non-complementary colour. The
halos are always negative with respect to density and colour.

I have tested all the sharpeners discussed here with such an image
and i herewith challenge any one of them that they all produce
halos along high contrast edges and that my Powerretouche sharpness
Editor does not. Or, if the others avaoid halos, they dont sharpen
enough. This includes Vtie's otherwise nice software.

The PowerRetouche sharpener has two unique sharpening algorithms
that emmulate optics and of these one is an adaption of regular USM
with a halo correction function, the other is a completely new
concept that has an inbuilt intelligent self-altering algorithm,
that adjust the sharpening algorithm according to several optical
factors culled from the image at large and from the particular area
being sharpened. How and what is for obvious reasons secret, but it
involves artificial intelligence.

For this reason the filtering is timeconsuming (processing times
are not secret and there are clear tables of processing time on the
PR sharpener website), but i assure you it can not be done faster
if you want sharpening of that quality. Of course you can
compromise and go for less than perfect. Personally i prefer
perfection and hope the PowerRetouches are nothing less.

I wrote them because i was # !¤# fed up with the necessety of
tweaking and twisting PS in order to minimize inevitable halos,
highlight overburning, shadow noise, etc. etc when i retouched
scans of my paintings (I am an artist by profession and i have been
retouching with PS for professional printing for at least ten years
now).

Try the sharpeners with this kind of image and your will see clear
pros and cons.

Regards
Jan Esmann
. developer of PowerRetouche

--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com
--------------------------------------------------------------

Click to expand...

Robert Karpa · Nov 24, 2002

Rich,

Thankyou for your kind remarks. I was specifically referring to PowerRetouche. Have you tried it?
Robert Karpa

Rich Gibson said:
Once a week someone "discovers" one of them and posts a miraculous
new sharpener which, guess what? "BLOWS AWAY" everything else.

Please!

Rich

Robert Karpa said:

I have to vouch for this program. I'm very critical when it comes
to the quality of my digital files. When I tested this I was blown
away by the results. Far bettter than in camera sharpening or
unsharp mask in Photoshop. I've never seen anything like it. The
rest of the filters supplied with the program are real time savers
and they have a lot of interesting creative applications as well.
Robert Karpa

Click to expand...

Rylee Isitt · Nov 24, 2002

None of those were terribly great. Avoiding blur in the first place seems the best thing to do.

Jan Esmann said:
Yous asked for a comparison of the various sharpeners. Here's the
research i did while developing the Sharpness Editor.

http://www.powerretouche.com/SharpenersTested.htm

Not that i want to denigrade others, but i want to back my claims
that the Powerretouche sharpener is the only non destructive
sharpener.

Regards
Jan Esmann
--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com
--------------------------------------------------------------

Rich Gibson · Nov 24, 2002

You're entirely welcome, but I think you were being sarcastic. No, I haven't tried it as well as numerous other products which are declared to be so outstanding as to "blow away" the other products. Oops, I sait "that" phrase again.

I'll try it.

Rich

Robert Karpa said:
Rich Gibson said:

Once a week someone "discovers" one of them and posts a miraculous
new sharpener which, guess what? "BLOWS AWAY" everything else.

Please!

Rich

Robert Karpa said:

I have to vouch for this program. I'm very critical when it comes
to the quality of my digital files. When I tested this I was blown
away by the results. Far bettter than in camera sharpening or
unsharp mask in Photoshop. I've never seen anything like it. The
rest of the filters supplied with the program are real time savers
and they have a lot of interesting creative applications as well.
Robert Karpa

Click to expand...

Click to expand...

Rich Gibson · Nov 24, 2002

Tried it. It is no better or worse than the others. Awfully expensive for what it does. Sorry.
Rich

Rich Gibson said:
I'll try it.

Rich

Robert Karpa said:

Rich Gibson said:

Once a week someone "discovers" one of them and posts a miraculous
new sharpener which, guess what? "BLOWS AWAY" everything else.

Please!

Rich

Robert Karpa said:

I have to vouch for this program. I'm very critical when it comes
to the quality of my digital files. When I tested this I was blown
away by the results. Far bettter than in camera sharpening or
unsharp mask in Photoshop. I've never seen anything like it. The
rest of the filters supplied with the program are real time savers
and they have a lot of interesting creative applications as well.
Robert Karpa

Click to expand...

Click to expand...

Click to expand...

MarkMcCombs · Nov 24, 2002

Rylee,

Since you mentioned that you are new to digital photography, I will be the one to tell you......unsharp masking's purpose is not to FOCUS an image that's not in focus. USM is used because there is digital softness inherent in acquiring a digital image, even when perfectly focused. It is not intended to take an image that's out of focus and make it focused. A very common misconception, however......

Mark

Rylee Isitt said:
Jan Esmann said:

Yous asked for a comparison of the various sharpeners. Here's the
research i did while developing the Sharpness Editor.

http://www.powerretouche.com/SharpenersTested.htm

Not that i want to denigrade others, but i want to back my claims
that the Powerretouche sharpener is the only non destructive
sharpener.

Regards
Jan Esmann
--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com
--------------------------------------------------------------

Click to expand...

Jan Esmann · Nov 25, 2002

Rich, look at the tests, please. The difference is glaringly obvious.

http://www.powerretouche.com/SharpenersTested.htm

Jan

Rich Gibson said:
Rich Gibson said:

I'll try it.

Rich

Robert Karpa said:

Rich Gibson said:

Once a week someone "discovers" one of them and posts a miraculous
new sharpener which, guess what? "BLOWS AWAY" everything else.

Please!

Rich

Robert Karpa said:

I have to vouch for this program. I'm very critical when it comes
to the quality of my digital files. When I tested this I was blown
away by the results. Far bettter than in camera sharpening or
unsharp mask in Photoshop. I've never seen anything like it. The
rest of the filters supplied with the program are real time savers
and they have a lot of interesting creative applications as well.
Robert Karpa

Click to expand...

Click to expand...

Click to expand...

Click to expand...

--
--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com

Vtie · Nov 25, 2002

Jan Esmann said:
Rich, look at the tests, please. The difference is glaringly obvious.

http://www.powerretouche.com/SharpenersTested.htm

Jan

Why are you using a test image that does not contain any
noise at all? As you definetely know, the real challenge of a
sharpening algorithm is to sharpen the information content while
minimising the noise amplification.

If you want, I can easily write you a small piece of software that
perfectly "sharpens" your image to the original one. It is just a matter
of inverting the Gaussian blur. However, although it would do a
perfect job on your test image, it would be useless for real images.

I always enjoy a good challenge, but please use a data set that
makes sense. Never judge a sharpening algorithm on its behaviour
in the absence of noise.

Vtie

Vtie · Nov 25, 2002

Jan Esmann said:
Yous asked for a comparison of the various sharpeners. Here's the
research i did while developing the Sharpness Editor.
http://www.powerretouche.com/SharpenersTested.htm
Not that i want to denigrade others, but i want to back my claims
that the Powerretouche sharpener is the only non destructive
sharpener.

Why are you using a test image that does not contain any
noise at all? As you definetely know, the real challenge of a
sharpening algorithm is to sharpen the information content while
minimising the noise amplification.

If you want, I can easily write you a small piece of software that
perfectly "sharpens" your image to the original one. It is just a matter
of inverting the Gaussian blur. However, although it would do a
perfect job on your test image, it would be useless for real images.

I always enjoy a good challenge, but please use a data set that
makes sense. Never judge a sharpening algorithm on its behaviour
in the absence of noise.

Vtie

phil wigglesworth · Nov 25, 2002

Jan Esmann said:
The PowerRetouche sharpener has two unique sharpening algorithms
that emmulate optics

Traditional USM, or transfer function stuff like this was done via film chemistry or creation of an actual mask: these are what USM emulates. In what way does this "emulate optics"? I don't understand that.

Jan Esmann said:
with a halo correction function, the other is a completely new
concept that has an inbuilt intelligent self-altering algorithm,
that adjust the sharpening algorithm according to several optical
factors culled from the image at large and from the particular area
being sharpened. How and what is for obvious reasons secret, but it
involves artificial intelligence.

Sounds like a heuristic. I'm not sure that makes it artificialyl intelligent. Does it do anything else?

Jan Esmann said:
Personally i prefer
perfection and hope the PowerRetouches are nothing less.

I guess it depends how you define perfection. The samples are apparently good, although they're hardly perfect.

The samples are very specific. I would be surprised if any Photoshop add-in supplier would not have a set of samples which demonstrate their product in a good light.

As you say, the best thing is to download it and try it out, but I'm pretty happy with PS7 already. I've failed to see improvements over basic PS with most of these thiings, so I think I'll wait until everyone else gets blown away before I worry.

phil

Andrew Rodney · Nov 25, 2002

I agree. Jan's test IS informative to a certain degree. But what we are sharpening here are IMAGES which behave a lot differently then the test image of Jan's. Yes, his image does show some interesting things about how the software products deal with this kind of file.

Some of the more involved techniques in Photoshop use the image itself in building a very complex mask for applying a USM over. In such a case, the tests seen so far don't fly. Using the actual image to insure that a mask is generated so that sharpening is only applied to certain tonal levels and not in smooth areas can produce very, very good results.

Bottom line would be to see the various software products applied on a REAL image (or set of images ganged up in a file) and OUTPUT to a KNOWN device whereby we could see the output! Sharpening IS output and resolution dependant (as well as subject dependant). So doing a test to a 4 color press and then something like a Lightjet will require different degrees of sharpening even if output the same size with the same subject.

I do think most of the demo filters of Jan's are way cool! And I think Adobe needs to update USM which is (with exception of an upgraded GUI preview) 12 years old!
--
Andrew Rodney
http://www.digitaldog.net

Jan Esmann · Nov 25, 2002

This noise test, you say i have not performed, has been available on the PowerRetouche website for a whole year now!

http://www.powerretouche.com/SharpnessUserguide.htm
Scroll to the bottom of the page

???

Jan

Vtie said:
Why are you using a test image that does not contain any
noise at all? As you definetely know, the real challenge of a
sharpening algorithm is to sharpen the information content while
minimising the noise amplification.

If you want, I can easily write you a small piece of software that
perfectly "sharpens" your image to the original one. It is just a
matter
of inverting the Gaussian blur. However, although it would do a
perfect job on your test image, it would be useless for real images.

I always enjoy a good challenge, but please use a data set that
makes sense. Never judge a sharpening algorithm on its behaviour
in the absence of noise.

Vtie

--
--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com

Jan Esmann · Nov 25, 2002

This noise test, you say i have not performed, has been available on the PowerRetouche website for a whole year now!

http://www.powerretouche.com/SharpnessUserguide.htm
Scroll to the bottom of the page

???

Jan

Vtie said:
Why are you using a test image that does not contain any
noise at all? As you definetely know, the real challenge of a
sharpening algorithm is to sharpen the information content while
minimising the noise amplification.

If you want, I can easily write you a small piece of software that
perfectly "sharpens" your image to the original one. It is just a
matter
of inverting the Gaussian blur. However, although it would do a
perfect job on your test image, it would be useless for real images.

I always enjoy a good challenge, but please use a data set that
makes sense. Never judge a sharpening algorithm on its behaviour
in the absence of noise.

Vtie

--
--------------------------------------------------------------
Photoediting plugins: http://powerretouche.com
Paintings and writings: http://janesmann.com

Vtie · Nov 25, 2002

Jan Esmann said:
This noise test, you say i have not performed, has been available
on the PowerRetouche website for a whole year now!

http://www.powerretouche.com/SharpnessUserguide.htm
Scroll to the bottom of the page

I definetely did not say that you never performed any noise test!
How could I possibly know what you have done in your life...

The only thing I said is that, if you publish a comparison between
sharpening algorithms, you should absolutely include the noise
aspect in it.

As far as I can see, the web page that compares PowerRetouche
to some other methods does not compare how they act on noise,
although this is a crucial issue.

Please don't change my words... Its a cheap demagogical trick and
definetely below your level :->

And just to make sure: I do not say that powerretouche is not
a superior package. It may very well be better than anything else!
I am definetely not going to compare Powerretouche to SharpControl.
As you can tell from the UI, the latter is a quick-and-dirty thing I just
wrote for fun in 2 evenings. I am not sure whether I ever will put
some serious effort in it to make a mature piece of software.
Anyway, the algorithm was designed to treat very very bad data
(maybe out of focus) with lots of noise. For "nice" data, it is not even
better than PS's straight USM...

Vtie

Vtie

Best USM Tool I've seen

Leading Member

Leading Member

Active member

Member

Member

Member

Rich Gibson

Guest

Member

Senior Member

Rich Gibson

Guest

Rich Gibson

Guest

Leading Member

Member

Senior Member

Senior Member

Leading Member

Senior Member

Member

Member

Senior Member

Similar threads

Keyboard shortcuts