Ad
Ad
Ad
Pages: « 1 [2] 3 »   Bottom of Page
Print
Author Topic: Lightroom print sharpening, why so layperson?  (Read 8803 times)
BartvanderWolf
Sr. Member
****
Online Online

Posts: 3884


« Reply #20 on: July 12, 2013, 06:33:30 AM »
ReplyReply

Do you understand the implications of parametric adjustments?

Yes Jeff, but thanks for your concern.

Quote
If you did you would realize the complexity of changing local control channel parameters...would I like different local adjustment parameters?

Of course it's complex, that's why the engineers get paid good salaries, to try and solve it. But using the complexity as an excuse also suggests that its functionality is currently sub-optimal, which is indeed what I said earlier.

Quote
Yes...do you have explicit use case examples where there is a clear advantage? If you do, bring them forward...Eric and Thomas are always willing to listen if you know how to speak to them.

I'm a bit surprised that you'd have to ask me for examples if/where Creative sharpening requires even different settings than Capture sharpening. It's one of the main topics of Bruce Fraser's Real World Image Sharpening book, which you co-authored. If Capture sharpening didn't require different settings, then why separate them in the first place.

I fully agree that they require a different approach, in fact they require different methods/algorithms to tackle the issue.

Capture sharpening requires deconvolution sharpening, because the Capture blur is characterized by a somewhat predictable hardware induced Point Spread Function (PSF) that lends itself for deconvolution very well. The common challenge with that approach is in separating signal from noise, and in avoiding ringing/halo artifacts.

Creative 'sharpening' on the other hand has more to do with boosting/attenuating specific spatial frequencies, and adjusting local contrast. That would be more logically approached by a simple USM or High-pass type of spatial domain adjustment, although the modern approaches use more computationally efficient frequency domain adjustments based on Wavelet conversions to address different feature sizes, and Adaptive Bi-lateral filtering to avoid edge halos. Here, also a lot of care must be taken to avoid perceptually unconvincing shifts in color/saturation which can occur when the local brightness levels are changed significantly.

Suggesting that Creative sharpening can be done well by only boosting the local amount of Capture sharpening, is silly. Yes, it's more complex to do in a parametric editor but, no, it is not good enough as it is currently implemented.

Cheers,
Bart

P.S. This also reminded me of a quote from a famous inspirational speech: "We do these things not because they are easy, but because they are hard"
-John F. Kennedy, Rice University, Sept. 12, 1962
« Last Edit: July 12, 2013, 07:02:21 AM by BartvanderWolf » Logged
Robert-Peter Westphal
Sr. Member
****
Offline Offline

Posts: 281



WWW
« Reply #21 on: July 12, 2013, 06:54:40 AM »
ReplyReply

Hello,

I wonder if we use the finest detail in the image to determine the radius of the sharpening in the detail-panel, and then, after applying it properly, use the brush with a higher amount of sharpening, does the result come near to the same result as when sharpening the same image in PS with complete different radius / amount using layers. Additionally we could use some halo-suppression by the details-slider.

@Jeff : I remember in one of the Lr videos you did the roundtrip to PS exactly becasue you did the sharpening of the iceberg with different radius.

Robert
Logged

'visit my completly renewed gallery at http://www.naturfotografie-westphal.com '
digitaldog
Sr. Member
****
Offline Offline

Posts: 9230



WWW
« Reply #22 on: July 12, 2013, 08:39:26 AM »
ReplyReply

Which then begs the question, why can't the Creative sharpening Radius and Method be adjusted when using the adjustment brush ...?

LR has creative sharpening? I know it has localized sharpening, is that the same (as creative sharpening in PKS)?
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
BartvanderWolf
Sr. Member
****
Online Online

Posts: 3884


« Reply #23 on: July 12, 2013, 09:54:26 AM »
ReplyReply

LR has creative sharpening? I know it has localized sharpening, is that the same (as creative sharpening in PKS)?

Hi Andrew,

Actually, we did...everything in LR's sharpening both capture, creative and output sharpening was implemented in accordance to Bruce's workflow.

So, unless Capture and Creative sharpening are folded into one operation (contrary to Bruce's methods), the adjustment brush and Clarity are what it must be.

Cheers,
Bart
« Last Edit: July 12, 2013, 09:57:15 AM by BartvanderWolf » Logged
BartvanderWolf
Sr. Member
****
Online Online

Posts: 3884


« Reply #24 on: July 12, 2013, 10:11:28 AM »
ReplyReply

I wonder if we use the finest detail in the image to determine the radius of the sharpening in the detail-panel, and then, after applying it properly, use the brush with a higher amount of sharpening, does the result come near to the same result as when sharpening the same image in PS with complete different radius / amount using layers.

Hi Robert,

That's the issue, it would not have the same effect. We should indeed use the finest image detail to determine that optimum radius and amount for Capture sharpening. Which leaves the Creative sharpening, for which we only have an adjustment brush available with an amount between -100 and +100. Nothing there indicates that it also somehow changes the radius. That only leaves Clarity to influence the Creative 'sharpening', the impression of sharpness by changing local contrast.

Quote
Additionally we could use some halo-suppression by the details-slider.

That's partially addressed by the Masking control, but that probably would mean that we are using too large a radius already. With the correct Radius for Capture sharpening, and the correct amount, we would not have any halos, only sharpness (which requires deconvolution, which means that the Detail slider must be pulled as far to the right as acceptable).

Cheers,
Bart
Logged
tommm
Jr. Member
**
Offline Offline

Posts: 77


« Reply #25 on: July 12, 2013, 10:40:29 AM »
ReplyReply

Hi Jeff (and Bart),

I do understand a little about sharpening. And yes I also know that you took over writing Real World Image Sharpening after Bruce passed away. I've read Bruce's stance on sharpening and yours and many others. And stemming from Bruce's work most people agree that capture sharpening and creative sharpening should be separate stages in the workflow, with capture sharpening aiming to reverse the effects of hardwear (camera, lens, scanner, etc), whilst creative is content dependent.

You clearly agree with this and have helped implement good sharpening tools in lightroom. However, I (and clearly Bart and probably many others) feel that a truly three stage sharpening workflow would allow even better sharpening. I don't think we're trying to start an arguement but merely point out what we as uses would like to see in future lightoom upgrades.

One example of why more flexible output sharpening would be useful is that as stated by Bruce and others, there is vast difference in how screens currently show an image in terms of sharpness. So for a given image if I sharpen it in lightroom to look good on screen and then pop it open on a different screen it may look over or under sharpened. This means that each individuals on screen "to taste" sharpening, even ignoring there taste, will be different depending on what screen they are using. Therefore having three sizes fits all output sharpening doesn't allow fine tuning for different screens (or tastes or eyesight or viewing conditions, etc).

In an ideal world I would like to be able to do:

a) Capture sharpening based on capture hardwear alone. (with profiles as a starting points ideally)

b) Creative sharpening both globally and locally based on the image "looking correct" on screen for screen based uses.

c) Output sharpening based on output media and size. (with profiles as we have but with a slider for fine tuning, as screens and tastes vary)

I think Bart has made the case more strongly and more technically than I but hopefully you get the idea that we're just trying to advance things rather than just moan!

Cheers,

Tom

Tom
Logged
Robert-Peter Westphal
Sr. Member
****
Offline Offline

Posts: 281



WWW
« Reply #26 on: July 12, 2013, 11:00:15 AM »
ReplyReply

[...]
That's partially addressed by the Masking control, but that probably would mean that we are using too large a radius already. With the correct Radius for Capture sharpening, and the correct amount, we would not have any halos, only sharpness (which requires deconvolution, which means that the Detail slider must be pulled as far to the right as acceptable).

Cheers,
Bart

Hallo Bart,

as far as I understood Jeff right, the details slider has two different meanings. When moved to the left side, it works like a halo-supression and when moved to the left, it switches the complete sharpening to deconvolutionmsharpening, which means it doesn't simly increase the contrast, but calculates the funtion which is responsible for the 'unsharpness' and tries to find another function which inverts this process. In easy words, it tries to recreate the sharpness of the object by calculation the things responsible for the unsharpness.

So, when pushing the slider nearly all the way to the left, it suppresses effective halos. In the case you do your creative sharpening with a large amount, you could finetune it this way.

Best wishes

Robert

Logged

'visit my completly renewed gallery at http://www.naturfotografie-westphal.com '
BartvanderWolf
Sr. Member
****
Online Online

Posts: 3884


« Reply #27 on: July 12, 2013, 11:46:36 AM »
ReplyReply

Hallo Bart,

as far as I understood Jeff right, the details slider has two different meanings. When moved to the left side, it works like a halo-supression and when moved to the left, it switches the complete sharpening to deconvolutionmsharpening, which means it doesn't simly increase the contrast, but calculates the funtion which is responsible for the 'unsharpness' and tries to find another function which inverts this process. In easy words, it tries to recreate the sharpness of the object by calculation the things responsible for the unsharpness.

Hi Robert,

That's not how Eric Chan explained the functionality earlier:"...the Detail slider in CR 6 & LR 3 is a blend of sharpening/deblur methods and if you want the deconv-based method then you crank up the Detail slider (even up to 100 if you want the pure deconv-based method). "

So the detail slider effectively sets a blend between different basic sharpening methods, with a more traditional USM type of sharpening at its minimum, and deconvolution sharpening at it's maximum. Both will produce halo/ringing artifacts when set to the wrong radius and/or to too high an Amount.

Cheers,
Bart

P.S. In another post by Eric he explains how the Detail panel settings carry over to the adjustment brush.
« Last Edit: July 12, 2013, 12:33:53 PM by BartvanderWolf » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7926


WWW
« Reply #28 on: July 12, 2013, 12:40:13 PM »
ReplyReply

Bart,

For once you are wrong. Detail to left invokes halo suppression. You can use a lot amount and you get no halos. I'm pretty sure that both Eric and Jeff has stated this, but you can simply try yourself.

Best regards
Erik

Hi Robert,

That's not how Eric Chan explained the functionality earlier:"...the Detail slider in CR 6 & LR 3 is a blend of sharpening/deblur methods and if you want the deconv-based method then you crank up the Detail slider (even up to 100 if you want the pure deconv-based method). "

So the detail slider effectively sets a blend between different basic sharpening methods, with a more traditional USM type of sharpening at its minimum, and deconvolution sharpening at it's maximum. Both will produce halo/ringing artifacts when set to the wrong radius and/or to too high an Amount.

Cheers,
Bart

P.S. In another post by Eric he explains how the Detail panel settings carry over to the adjustment brush.
Logged

Robert-Peter Westphal
Sr. Member
****
Offline Offline

Posts: 281



WWW
« Reply #29 on: July 12, 2013, 12:43:30 PM »
ReplyReply

Hi Bart,

when you take a look at Jeff's book 'the digital negative' at page 89, he explains that nearly the way I did. Actually I'm waiting for Jeff's ok to give you the excerpts from his book.
( I don't want to be killed by Jeff  Smiley )

Best wishes

Robert
Logged

'visit my completly renewed gallery at http://www.naturfotografie-westphal.com '
BartvanderWolf
Sr. Member
****
Online Online

Posts: 3884


« Reply #30 on: July 12, 2013, 01:26:54 PM »
ReplyReply

Bart,

For once you are wrong. Detail to left invokes halo suppression. You can use a lot amount and you get no halos. I'm pretty sure that both Eric and Jeff has stated this, but you can simply try yourself.

Hi Erik,

Thanks for that. However, already with the detail slider to 1 or 2, I can apparently produce halo (see attachment). Not that I would ever use the settings as shown, but I'm not so sure that active halo suppression is applied. If there is (maybe it is added  in a later version) then it's restricted to the extreme left of the range, because at 1 or 2 I can already create some. Edit: I've also added a Detail=0 version of the Raw file conversion screen capture.

I've not seen Eric (who should know better than any other poster's anecdotal interpretation) mention the active halo suppression, but I would of course welcome any links to where he might have done so. Maybe there is less of a tendency to the formation of halos due to the choice of algorithm (not USM-like but something else), so it would be interesting to know if it is specifically activated or a welcome by-product.

Cheers,
Bart
« Last Edit: July 12, 2013, 02:16:50 PM by BartvanderWolf » Logged
Robert-Peter Westphal
Sr. Member
****
Offline Offline

Posts: 281



WWW
« Reply #31 on: July 12, 2013, 01:37:51 PM »
ReplyReply

Hi Erik,

Thanks for that. However, already with the detail slider to 1 or 2, I can apparently produce halo (see attachment). Not that I would ever use the settings as shown, but I'm not so sure that active halo suppression is applied. If there is, maybe it is added  in a later version, then it's restricted to the extreme left of the range because at 1 or 2 I can already create some.

I've not seen Eric (who should know better than any other poster's anecdotal interpretation) mention the active halo suppression, but I would of course welcome any links to where he might have done so. Maybe there is less of a tendency to the formation of halos due to the choice of algorithm (not USM-like but something else), so it would be interesting to know if it is specifically activated or a welcome by-product.

Cheers,
Bart

Bart,

halo suppresssion does not mean that it removes all halos or prohibit the creation at all, but it suppresses them in a 'special range'.

Robert
Logged

'visit my completly renewed gallery at http://www.naturfotografie-westphal.com '
BartvanderWolf
Sr. Member
****
Online Online

Posts: 3884


« Reply #32 on: July 12, 2013, 02:14:47 PM »
ReplyReply

halo suppresssion does not mean that it removes all halos or prohibit the creation at all, but it suppresses them in a 'special range'.

Hi Robert,

The question then becomes; is it a welcome by-product or an intentional addition to the functionality of the underlying slider algorithms, and in what range does it happen? I would appreciate it if someone can find a reference to a post by Eric Chan that indicates such functionality. Eric is one of the few people who can know and understand what exactly happens under the hood.

That BTW still doesn't mean that a low Detail slider value is useful for Capture sharpening, because only Deconvolution really increases resolution (by restoring the blur components to their original position). Other methods only mimic resolution by boosting edge contrast, which is not the same as real resolution.

Cheers,
Bart
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #33 on: July 12, 2013, 02:28:19 PM »
ReplyReply

I would appreciate it if someone can find a reference to a post by Eric Chan that indicates such functionality. Eric is one of the few people who can know and understand what exactly happens under the hood.

The term "halo suppression" comes from Mark Hamburg, the original engineer that wrote the code for the ACR sharpening changes instituted in ACR 4.1. This was before Eric was even on the team. However, Eric has taken over the code from Mark and he's the one that modified the high detail setting moving towards deconvolution sharpening. I'll leave it to Eric to say whether the lower detail settings have been changed...
Logged
NikoJorj
Sr. Member
****
Offline Offline

Posts: 1063


WWW
« Reply #34 on: July 12, 2013, 03:29:49 PM »
ReplyReply

Wouldn't it be great (given that capture sharpening should be based solely on camera, lens and aperture, and not content, output size, etc) if a future version of Lightroom implemented capture sharpening along the lines of lens profiles? Where the appropriate capture sharpening had been calculated and was applied automatically (tweakable by user as with lens profiles), leaving us to get on with creative sharpening?
Some lens defects are already tackled in a full automatic way in LR (distorsion with µ4/3 cameras eg), and this could be another good use case.
I'm not even sure if this has to be lens-specific (as DxO seems to do with a blur mapping, giving more sharpening in the edges which can give a strange look if overdone), but it could at least take into account the sensor (AA filter) and aperture (diffraction, plus some generic optical defects?).
Logged

Nicolas from Grenoble
A small gallery
ButchM
Full Member
***
Offline Offline

Posts: 188


« Reply #35 on: July 12, 2013, 09:18:00 PM »
ReplyReply

I have had my problems, concerns and direction with Lr development. However, sharpening, both in the Develop module and upon Export ... is something that Adobe got right and is one aspect I have invested little worry over.

Software solutions should simplify our workflow, not complicate it. I think that for the vast majority of images I work with, Lr sharpening gets it right ... even though the options and individual control over the end process may not be as configurable as what we have become accustomed with when using other options. Lightroom sharpening gets it right far more often that it gets it wrong. For me, that's a huge plus in simplifying my duties without sacrificing quality. IMHO that is a huge bonus.

Life as a photographer shouldn't be so complicated. I know we try to take on a structured process where we have infinite possibilities to hone our processing to the "Nth" degree ... but seriously ... does it really need to be that intricate and convoluted in order to achieve our goals? ...
Logged
Robert-Peter Westphal
Sr. Member
****
Offline Offline

Posts: 281



WWW
« Reply #36 on: July 13, 2013, 01:56:42 AM »
ReplyReply

Hi Bart,

when you take a look at Jeff's book 'the digital negative' at page 89, he explains that nearly the way I did. Actually I'm waiting for Jeff's ok to give you the excerpts from his book.
( I don't want to be killed by Jeff  Smiley )

Best wishes

Robert



Ok, Jeff allowed :
The digital Negative - page 89
[...] When adjusted towards 0, the Detail slider kicks in a halo supression algorithm,[...].
[...] Moving toward 100, the Detail kicks in a deconvolution based sharpening [...] )

As far as I know Jeff, he has talked over hours with Eric and Thomas prior to writing this.
Logged

'visit my completly renewed gallery at http://www.naturfotografie-westphal.com '
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #37 on: July 13, 2013, 02:05:33 AM »
ReplyReply

As far as I know Jeff, he has talked over hours with Eric and Thomas prior to writing this.

Talked? Not so much...emailed? Yep...(well, ok, I've talked to Eric a lot...Thomas doesn't "talk" much).
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7926


WWW
« Reply #38 on: July 13, 2013, 04:24:22 AM »
ReplyReply

Hi,

I would say that I am satisfied with what LR has to offer. Now, would I shoot people I perhaps would need more selective tools.

If I need heavy creative sharpening I would go to Photoshop anyway.

Best regards
Erik


Talked? Not so much...emailed? Yep...(well, ok, I've talked to Eric a lot...Thomas doesn't "talk" much).
Logged

BartvanderWolf
Sr. Member
****
Online Online

Posts: 3884


« Reply #39 on: July 13, 2013, 06:41:08 AM »
ReplyReply

The digital Negative - page 89
[...] When adjusted towards 0, the Detail slider kicks in a halo supression algorithm,[...].

Hi Robert,

I understand what has been written, but based on what we can actually see happening, I remain skeptical.

Quote
As far as I know Jeff, he has talked over hours with Eric and Thomas prior to writing this.

I don't doubt that a lot of info has been exchanged but, as it often happens, that doesn't mean that something couldn't have gotten lost in translation between sender and receiver. It's only human, we tend to filter the incoming info (also based on what we understood of it), and pass on something that's slightly modified (unless we quote verbatim, in context).

Because we didn't partake in the conversations, all we can do is observe the results, and the observation is that halos still exist, hence the skepticism. So, either the halo suppression algorithm isn't 100% effective, or lower amplitude halos are simply the by-product of the chosen algorithm.

Anyway, this potentially distracts from the fact that with the detail slider towards the left, one is basically using a kind of Creative sharpening method (modifying feature and edge contrast), and with the Detail slider towards the right one is using a type of Capture sharpening method (Deconvolution, restoration of resolution).

Also established is that the Detail panel parameters are carried over to the adjustment brush, which only allows to change the amount (also in the negative direction, thus allowing to remove either contrast or sharpness, depending on the detail slider position). It is not possible to use different types of sharpening for Capture or Creative sharpening, we can only use a given blend between them for both. Only changing the amount is not an accurate implementation of Bruce Fraser's intended separation between Capture and Creative sharpening.

Cheers,
Bart
Logged
Pages: « 1 [2] 3 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad