To sharpen or not to sharpen in Capture One

[robgo2]

I am familiar with the concept of multi-stage sharpening, and it is what I have typically used heretofore.  The C1-5 User's Manual states:

"When an image is destined for post-production or retouching in other
software, it is recommend processing the image without sharpening applied."

However, Capture One 5 also contains some Pre-sharpening defaults.  Somehow, this seems contradictory.  When, if ever, might one want to use these defaults?

Rob

[Jack Varney]

I suggest you ask your question on the PhaseOne site. There are many Capture One users and PhaseOne experts there.

[ErikKaffehr]

Hi,

In my view the multi stage sharpening workflow is optimal. The question is where capture-sharpneing should be done. If you are using some more advanced tools like Photokit Sharpener or Noise Ninja downward the processing line you would probably not use capture sharpening in C1. On the other hand, capture sharpening, noise reduction and elimination of chromatic aberration are probably done in connection with the demosaic step.

So I would suggest using capture sharpening in C1 but be careful with amount and radius so no haloes will be induced.

Best regards
Erik

I am familiar with the concept of multi-stage sharpening, and it is what I have typically used heretofore.  The C1-5 User's Manual states:

"When an image is destined for post-production or retouching in other
software, it is recommend processing the image without sharpening applied."

However, Capture One 5 also contains some Pre-sharpening defaults.  Somehow, this seems contradictory.  When, if ever, might one want to use these defaults?

Rob

[tho_mas]

However, Capture One 5 also contains some Pre-sharpening defaults.  Somehow, this seems contradictory.  When, if ever, might one want to use these defaults?

C1 is a quite well equipped RAW software. So you can edit the images in a way that no further editing in another software is needed. In this case you will likely apply sharpening in C1. And that's what the presets are for (the presets, of course, are just a set of choices).
Too, even if you disable sharpening on output, you might want to use sharpening for preview.

[robgo2]

C1 is a quite well equipped RAW software. So you can edit the images in a way that no further editing in another software is needed. In this case you will likely apply sharpening in C1. And that's what the presets are for (the presets, of course, are just a set of choices).
Too, even if you disable sharpening on output, you might want to use sharpening for preview.

My usual workflow with DxO/Lightroom has been to apply no sharpening in the RAW stage.  Upon converting to TIFF, I use Nik's Sharpener Pro 3 to apply what they call pre-sharpening and subsequently output sharpening as the final step.  If I use the C1 sharpening in RAW, I will skip Nik's pre-sharpening, but I still do not know what is the optimal process.  I must say that on the monitor, C1's default sharpening looks awfully good.

Rob

[tho_mas]

My usual workflow with DxO/Lightroom has been to apply no sharpening in the RAW stage.  Upon converting to TIFF, I use Nik's Sharpener Pro 3 to apply what they call pre-sharpening and subsequently output sharpening as the final step.  If I use the C1 sharpening in RAW, I will skip Nik's pre-sharpening, but I still do not know what is the optimal process.  I must say that on the monitor, C1's default sharpening looks awfully good.

I think nobody can really help you here as sharpening is totally subjective and it depends very much on the final target. If you do some uprezzing (or downrezzing) for the final print in my experience it's generally the best way to leave the capture without any sharpening prior to resizing. The same goes for distortion correction and/or for straightening an image. So I always apply sharpening with regard to the final purpose.
Too, I store my processed 16bit TIFs in my archive (together with the RAW files) and this is why I wouldn't apply any sharpening in the RAW software anway.
Finally - although the sharpening in C1 is generally very good since Version 4 - I miss an option to spare the whites and blacks from sharpening.
However I use the C1 preset "Version 3.7 soft look" as preview for editing my files.

[Jack Flesher]

What Thomas said.  What I can tell you, speaking only from my perspective, is I used to do light capture sharpening, then do post sharpening in CS.  However with C1 5, I soon found I no longer needed to bother with any post sharpening, so I now only do capture sharpening in C1 and call it a day.  (Yes, I think it is that good.)

[Henry Goh]

I fully agree with Jack.

[qwz]

i agree with Jack too
no need for sharpening after C1 5 in most cases
(i use 3.7 Soft Look preset)

[LynnNoah]

Jack:  Could you describe your C1 sharpening considerations in more detail (pre-sets, settings for variables)?  I'm now using PK for capture & print-output sharpening after processing P-40+ files in C1-5.1.

Thanks,

Lynn Noah

[Jack Flesher]

Jack:  Could you describe your C1 sharpening considerations in more detail (pre-sets, settings for variables)?  I'm now using PK for capture & print-output sharpening after processing P-40+ files in C1-5.1.

Thanks,

Lynn Noah

Hi Lynn,

As Thomas indicated, sharpening is somewhat a personal taste item, though there are practical considerations -- like artifacts.

That said, with a P40+, I would spend some time with one well-captured file with lots of good detail in various frequencies.  Then I would process it starting with the built-in Pre-Sharpen 1 (200/0.5/0..  From there, I would start dialing back on the amount to that which still holds detail without creating artifacts in the highest frequency detail.  Then I'd move up on the threshold a bit so noise* doesn't get exaggerated until you find a suitable value.  Then I would tweak radius up and down to benefit most of the various frequencies while keeping free of obvious artifacts.  *Note that sharpening by itself is incomplete without also having a defined NR setting, so a good idea to settle on that first. Here I look for a combination of Lum and Color settings that keep detail high without exaggerating noise, while at the same time keeping color smooth. (As a general rule, I find most of the time I prefer my color slider at around 2x the value of my Lum slider.)  You'll know you've hit sharpening and NR on the head when you pull the processed file into CS and even a slight sharpening like Smart Sharpen at 100%/0.4 generates visible and undesirable artifacts. These usually show up first in high-frequency areas with good contrast.  

Of course all of this is not only camera, but ISO dependent, so at the end of the day you may need to repeat this with several different files.  For myself, I have saved several various groups of these combinations of settings as "Styles" in C1, each with a specific name like "P65+ Daylight ISO 50-100."  So there is no "magic bullet" single sharpening and NR solution, and basically one needs to spend time with good files from each camera at various ISO to get it all sorted out as respects their personal tastes.

PS: In the latest version of C1 an "Advanced" NR tool has been added. For the purposes of dialing in regular sharpening and NR settings, this tool should be zeroed out.  The Fine Grain slider actually defines noise to make it look "better."  The Surface slider knocks back detail in smooth, even-toned areas, so can negatively affect the high frequency detail you're trying to optimize.

Cheers,

[tho_mas]

Of course all of this is not only camera, but ISO dependent

it also depends on the lens IMO ...

[LynnNoah]

Thanks Jack and Thomas.  I'm experimenting with the tests Jack suggested.  As with other tools in more recent versions of C1, I've found the learning curve a little confusing and frustrating but the superior results worth the effort, so I'm staying in C1 for more of the complete workflow.  A follow-up and perhaps dumb question on which the C1 User Guide still has me scratching my head:  I assume we've been talking only about capture sharpening, but when I get ready to print (from CS4) in possibly different sizes or different paper surfaces, would I still do the appropriate Photokit Sharpener inkjet output sharpening which I've been doing, or, as the C1 User Guide seems to suggest, go back into C1 for what they describe as a more aggressive second stage sharpening?  Or are you saying there would be just the one sharpening stage?  

Lynn

[ErikKaffehr]

Hi,

My opinion is that we need a multi step sharpening workflow. Capture sharpening tries to restore fine detail contrast (MTF) lost in capture. The losses arise trough lens aberrations, focusing errors, AA-filter and diffraction. It is not possible to compensate for all problems but for a well executed image capture sharpening really works.

The image that is capture sharpened can be used for any further processing.

Resizing the image causes aliasing and loss of sharpness, so resizing needs to be followed by some sharpening. Rescaling algorithms may or may not sharpen.

Once an image is output for some purpose it needs to be sharpened.

For printing it may be ideal to use Lightroom. Why? If the image is rescaled before printing the rescaling needs to taken into account and proper sharpening applied for output. Lots of experience is contained in the print routine. You can naturally do it yourself in PS, using Photokit Sharpener. The problem is that you need to scale and output sharpen for each image size. BTW, the capture and output sharpening in LR are based on Photokit Sharpener. Lots of experience in that product!

Lightroom does that with a workflow that is correct, tested and proven.

I really recommend the book by Bruce Fraser and Jeff Schewe on image sharpening.

Sorry for talking about Lightroom, I'm aware of you using C1, but LR is the tool I'm familiar with. Anyway, the short answer is that sharpening for output needs to be done just before printing. An image sharpened for printing is not really useful for anything else than printing it in a single predefined size.


Best regards
Erik

Thanks Jack and Thomas.  I'm experimenting with the tests Jack suggested.  As with other tools in more recent versions of C1, I've found the learning curve a little confusing and frustrating but the superior results worth the effort, so I'm staying in C1 for more of the complete workflow.  A follow-up and perhaps dumb question on which the C1 User Guide still has me scratching my head:  I assume we've been talking only about capture sharpening, but when I get ready to print (from CS4) in possibly different sizes or different paper surfaces, would I still do the appropriate Photokit Sharpener inkjet output sharpening which I've been doing, or, as the C1 User Guide seems to suggest, go back into C1 for what they describe as a more aggressive second stage sharpening?  Or are you saying there would be just the one sharpening stage?  

Lynn

[Jack Flesher]

I assume we've been talking only about capture sharpening, but when I get ready to print (from CS4) in possibly different sizes or different paper surfaces, would I still do the appropriate Photokit Sharpener inkjet output sharpening which I've been doing, or, as the C1 User Guide seems to suggest, go back into C1 for what they describe as a more aggressive second stage sharpening?  Or are you saying there would be just the one sharpening stage?  

Lynn

Hi Lynn,

I think the best way to state it is this: If the native file is close to perfectly sharpened, then any follow-on sharpening is likely to only add detrimental artifacts and not improve the image in any meaningful fashion, so best to not do it.

It used to be I did a capture sharpening in raw, then in post I ran a detail extraction routine to maximize high-frequency detail in the native sized file. Essentially, these two series went after different frequencies of detail.  In my current workflow it is this second detail enhancement step that I'm finding is no longer needed.  When I downsize for web or small prints, I perform that in CS via "Bicubic Sharper," which in itself contains sharpening as you downsize -- so in that fashion I am sharpening for output, but I do not use any other sharpening tool beyond that.  If I have to upsize, my routine uses Bicubic Smoother to 20% over desired size, then Bicubic Sharper to get back down to desired print size. This is sort of a quasi fractal sharpening method,  and again is my typical sharpening routine for print and usually no other output sharpening is required.  With very large prints, I sometimes invoke a targeted sharpening routine for edges and a second masked set for detail, however, since C1 ver 5 I have yet to need to do this for any of my captures.  Finally, if you print via a RIP or printing program like Q-Image, they apply their own proprietary output size-based sharpening algorithms that are usually quite good. (Personally, I send a pre-sized and optimized file to the printer as per above, so I dial this option down to the lower settings.)  

As for 3rd party sharpeners like Nik, I honestly have never needed them.  Not saying they're bad -- in fact they are probably great for folks that can't sharpen using CS, or who want to paint-in their sharpening. But IMO they're definitely not necessary, and even less so now with the improvements in C1 --  but then I like a simple, streamlined workflow, so the fewer programs I use, the better.

Again, this is just my basic workflow and YMMV...

Cheers,

[LynnNoah]

Jack, I appreciate your sharing your detailed insights into what I've found to be one of the fuzzier (pun intended) aspects of my C1 learning curve.  I'm working on some comparison testing, including with & without PK & LR pre-print sharpening as Erik suggested, and BTW have signed up for Glacier, so I'll let you know how it goes for me.

Lynn

[Jack Flesher]

BTW have signed up for Glacier, so I'll let you know how it goes for me.

I am planning to be there too .  I'll be happy to spend some time with you if you don't already have C1 nailed by then.

Cheers,

[Nill Toulme]

I sometimes like to use variations on the 20,50,0 "haze reduction" USM approach.  Is there a good way to efficiently accomplish something similar in C1?

Nill

[Jack Flesher]

I sometimes like to use variations on the 20,50,0 "haze reduction" USM approach.  Is there a good way to efficiently accomplish something similar in C1?

Nill

Hi Nill,

C1 only lets you take the radius up to 2.5 in their sharpening tool,  so you can't get wide enough.  I think the best option within C1 is probably the clarity slider.

[robgo2]

Jack,

I really appreciate your expert input on this subject.  Can you explain why the Capture One User's Guide recommends against sharpening in C1, if sharpening will be done in downstream editors?

Rob