Pages: [1] 2   Go Down

Author Topic: Hey, Shewe the Sharpener . . .  (Read 15609 times)

dabreeze

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 54
Hey, Shewe the Sharpener . . .
« on: May 14, 2007, 06:38:44 pm »

Tell me, Jeff, am I crazy for preferring your general all-purpose capture sharpening action you outline in your workflow PDF (green channel/find edges, etc.) @ 68% opacity over the PKS capture sharpen for high res cameras?

Yes, 'tis a bit more aggressive, but for the landscapes I shoot with lots of fine detail it seems to define things more and leaves the really fine tonal edges within leaves, branches et al alone. PKS capture is less aggressive but finds many more edges and looks, for lack of a more technical term, more spackled.

Even with the more aggressive sharpening, I have been happy with prints (up to 20x30) with additional layers of output sharpening (PKS) and even the occasional localized creative sharpen brush (PKS) as well.

Camera used is 1DsII, RAWS, 7600 inkjet, 300 dpi, luster and semi-matte papers.

Thoughts?
Logged

MichaelEzra

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1146
    • https://www.michaelezra.com
Hey, Shewe the Sharpener . . .
« Reply #1 on: May 15, 2007, 09:33:32 am »

300 dpi is incorrect setting try 360 dpi and your prints will dramatically improve. (otherwise printer driver will do it's own very low quality interpolation to 360 dpi)
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Hey, Shewe the Sharpener . . .
« Reply #2 on: May 15, 2007, 09:38:59 am »

Quote
300 dpi is incorrect setting try 360 dpi and your prints will dramatically improve. (otherwise printer driver will do it's own very low quality interpolation to 360 dpi)
[a href=\"index.php?act=findpost&pid=117671\"][{POST_SNAPBACK}][/a]

What printer model are you talking about, and have you actually tested this?
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

dabreeze

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 54
Hey, Shewe the Sharpener . . .
« Reply #3 on: May 15, 2007, 10:34:03 am »

I should add that the image comparison I'm making of the two capture sharpenings is on screen at 50% and 100% (LaCie 321). I haven't done side by side comparisons of 11x17 work prints or larger yet, although I most definitely should.

I haven't yet because I have fine art gallery prints using both styles and they look great.

I'm just noticing the difference on screen and was wondering if Jeff (sorry about the mispelling of your name in the topic title! oops!) could give some insight as to the differences and the reasoning behind the less aggressive but seemingly more detailed, comprehensive sharpening of fine edges in the PKS version.
« Last Edit: May 15, 2007, 11:19:57 am by dabreeze »
Logged

Eli Burakian

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 62
    • Photogaboutit
Hey, Shewe the Sharpener . . .
« Reply #4 on: May 15, 2007, 10:58:56 am »

Quote
What printer model are you talking about, and have you actually tested this?
[a href=\"index.php?act=findpost&pid=117672\"][{POST_SNAPBACK}][/a]

Yea.  I thought the whole idea of a native resolution on a printer was debunked. Is this not so? And if the print really is interpolated up by the printer, how do we even know when it's doing this.  Yes, I try to print at 300 or 360 but I assume if I have resolution to print at the size I want at 312dpi, there's no reason to downsample (or upsample it).

Eli
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Hey, Shewe the Sharpener . . .
« Reply #5 on: May 15, 2007, 11:42:58 am »

Quote
Yea.  I thought the whole idea of a native resolution on a printer was debunked. Is this not so? And if the print really is interpolated up by the printer, how do we even know when it's doing this.  Yes, I try to print at 300 or 360 but I assume if I have resolution to print at the size I want at 312dpi, there's no reason to downsample (or upsample it).

Eli
[a href=\"index.php?act=findpost&pid=117696\"][{POST_SNAPBACK}][/a]

No, it's not "debunked", but we should be sure what we're talking about. The Epson X800 series printers, and the series of the generation before it, rasterize image file data at 360 PPI. That is the meaning of "native resolution". You know it is doing this is if you send a file to the printer that is not 360 PPI. There is one school of thought that we should have Photoshop resample the files to 360 before sending them to print, and another school that says it doesn't matter what you send to the printer between say 240 and 480, you won't see a quality difference on paper without a loupe. I have tested this on letter sized and A3 prints and my observations generally tend to support the latter position. But there is another factor at work here for those of us who use it: PK Sharpener. As we know, the Inkjet Output Sharpener has a number of discrete settings: 180, 240, 300, 360 and 480. Ideally, one's image PPI should be at one of those settings for the most accurate sharpening, because the calculations were done for those settings and file sizes. However, Pixelgenius has advised me in the past that fine results are obtained using the nearest setting to the file size within their ranges. Bottom line: in principle a case can be made that lining everything up from Output Sharpening to Printing at 360 may deliver the best print quality from these series of Epson printers, but not doing so will still deliver very high print quality - the differences don't jump out at you.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Hey, Shewe the Sharpener . . .
« Reply #6 on: May 15, 2007, 12:34:01 pm »

Quote
Tell me, Jeff, am I crazy for preferring your general all-purpose capture sharpening action you outline in your workflow PDF (green channel/find edges, etc.) @ 68% opacity over the PKS capture sharpen for high res cameras?
[a href=\"index.php?act=findpost&pid=117556\"][{POST_SNAPBACK}][/a]

Pretty sure that PDF pre-dates PK Sharpener...and was a basis for Bruce and I to discuss sharpening workflow. And no, it doesn't surprise me that some people may prefer more aggressive vs less aggressive.

When Sharpener first came out, the Digital HighRez was the original 1Ds. We debated over adjusting or adding a new routine for the 1Ds MII-Bruce decided against it. I actually will often run SuperSharpener 1 before capture sharpening to add a touch of additional sharpening-then turn that layer off after. Now that a 1Ds MIII is looming, we'll need to revisit the capture sharpening routines, although Bruce won't be here for that-I've learned enough that I can muddle through.

Comparing the results on screen is a bit foolish-even at 50% it won't represent the final print. Closer to 25% is more accurate in terms of the screen dither. But even that is off due to the differences in sizes-particularly large prints. The only way of testing a sharpening workflow is to actually do the prints...

Quote
I thought the whole idea of a native resolution on a printer was debunked. Is this not so?

Depends on what your are asking...the "effective resolution" of Epson's 1440/720 printers is 360ppi. The effective resolution for Canon & HP is 300ppi. But unless you are doing substantial upsampling AND image processing on top, you really don't get much help with an image unless you are going 200-400% upsampled. You are a lot better off using the image's "native resolution" (which is also a bit of a misnomer) and sharpening for the actual resolution sent to the printer.

And just to be clear, the print drivers don't do interpolation so much as use an error diffusion algorithm to form dots...if you send it TOO much resolution, it just tends to drop it like a sieve...but too much can actually be a bad thing in some cases where you can encounter interference patterns.

It's been my experience that as long as your file has uninterpolated resolution between 180 and 480 ppi, you DON'T want to resample to print-just resize with no resampling and sharpen for the actual pixel density. If it falls between the PK Sharpening settings, use the one closest.

Bruce Fraser figured out the resolution required based on how much resolution the normal human eye can resolve at various viewing distances. Since human vision is based on 1 minute of 1 degree, depending on how close or far your viewing distance is the eye can resolve:
8" - 480
10" - 360
12" - 300
18" - 240
24" - 180

So, as long as your viewing distances fall in the 8"-24" range, 180-480 is all you need to worry about. Note, more resolution in the print isn't bad and Bruce liked to point out the intended viewing distance of photographers is limited only by the length of their nose...
« Last Edit: May 15, 2007, 12:35:30 pm by Schewe »
Logged

dabreeze

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 54
Hey, Shewe the Sharpener . . .
« Reply #7 on: May 15, 2007, 01:12:00 pm »

Actually I had always understood it that that was the proper viewing distance (nose length!) for photographers!! I just learned that by watching other photographers at exhibitions and they all seemed to be doing exactly the same thing.

Thanks for the feedback on capture sharpening. The interesting thing about the PKS capture setting is that while not as aggressive overall as your PDF method, it seems to seek and find many more edges within things like leaves, rocks, etc. and sharpens them, creating that "spackled" look I described in the original post.

I have compared web-prepped images using both methods and I think the PDF method looks a lot better after subsequent output sharpening for the web.

Both work fine in print, having compared 11x17 work prints. I have very large prints using both and I need to compare them. I think I'll probably prefer the initially more aggressive PDF method as I don't use the Super Sharpeners as you outlined in a Digital Photo Pro article a while back for large inkjet prints (that was your article, yes?).

Have you refined your large inkjet print approach since that article?

Also, it's nice to know PG is anticipating the new 1Ds3. Have you done any shooting with the 1D3 yet? Was wondering if the 14 bit processing and increased tonal gradation are making that much difference. Certainly others (Michael, Rob Galbaith) are praising the image files pretty highly and I would think this would mean great things for the new 1Ds!
« Last Edit: May 15, 2007, 01:21:51 pm by dabreeze »
Logged

jbrembat

  • Full Member
  • ***
  • Offline Offline
  • Posts: 177
Hey, Shewe the Sharpener . . .
« Reply #8 on: May 15, 2007, 01:15:53 pm »

Quote
The Epson X800 series printers, and the series of the generation before it, rasterize image file data at 360 PPI. That is the meaning of "native resolution"
If you want to know the PPI value expected from printer driver, go to http://www.photoresampling.com/index_eng.php and download for  free PrinterData. For quality printing Epsone PPI value is 720.
Quote
And just to be clear, the print drivers don't do interpolation so much as use an error diffusion algorithm to form dots...
This is a general misconception.
PPI are different from DPI. Error diffusion is to simulate to human eyes a color. Dithering is a process performed at DPI level on paper as soon as a pixel color is sampled. Image sampling is performed at PPI level.

Jacopo
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Hey, Shewe the Sharpener . . .
« Reply #9 on: May 15, 2007, 04:33:26 pm »

Quote
For quality printing Epsone PPI value is 720.

Jacopo
[a href=\"index.php?act=findpost&pid=117724\"][{POST_SNAPBACK}][/a]

This is incorrect - at least for the Epson 4800. Epson ProGraphics has confirmed to me that 360 PPI is the value at which image file data is rasterized. If more than this is sent to the printer the printer software has an algorithm which "smartly" discards excess information.

(This of course has nothing to do with whether one selects 720, 1440 or 2880 DPI as the ink coverage setting.)
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

jbrembat

  • Full Member
  • ***
  • Offline Offline
  • Posts: 177
Hey, Shewe the Sharpener . . .
« Reply #10 on: May 16, 2007, 03:34:37 am »

Quote
This is incorrect - at least for the Epson 4800. Epson ProGraphics has confirmed to me that 360 PPI is the value at which image file data is rasterized. If more than this is sent to the printer the printer software has an algorithm which "smartly" discards excess information.

(This of course has nothing to do with whether one selects 720, 1440 or 2880 DPI as the ink coverage setting.)
[a href=\"index.php?act=findpost&pid=117756\"][{POST_SNAPBACK}][/a]

I think you are right. I know that Epson fixed 720 PPI for best quality on desktop printer and 360 PPI for large format printer.

Jacopo
Logged

Chris_T

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 541
Hey, Shewe the Sharpener . . .
« Reply #11 on: May 16, 2007, 08:52:13 am »

From my unscientific testing, I could not conclude that it makes any observable difference. To claim that "prints will dramatically improve" would suggest that in a blind test, either the printer or a panel of viewers can EASILY tell the difference. I doubt very much either would be the case.

Quote
No, it's not "debunked", but we should be sure what we're talking about. The Epson X800 series printers, and the series of the generation before it, rasterize image file data at 360 PPI. That is the meaning of "native resolution". You know it is doing this is if you send a file to the printer that is not 360 PPI. There is one school of thought that we should have Photoshop resample the files to 360 before sending them to print, and another school that says it doesn't matter what you send to the printer between say 240 and 480, you won't see a quality difference on paper without a loupe. I have tested this on letter sized and A3 prints and my observations generally tend to support the latter position. But there is another factor at work here for those of us who use it: PK Sharpener. As we know, the Inkjet Output Sharpener has a number of discrete settings: 180, 240, 300, 360 and 480. Ideally, one's image PPI should be at one of those settings for the most accurate sharpening, because the calculations were done for those settings and file sizes. However, Pixelgenius has advised me in the past that fine results are obtained using the nearest setting to the file size within their ranges. Bottom line: in principle a case can be made that lining everything up from Output Sharpening to Printing at 360 may deliver the best print quality from these series of Epson printers, but not doing so will still deliver very high print quality - the differences don't jump out at you.
[a href=\"index.php?act=findpost&pid=117704\"][{POST_SNAPBACK}][/a]
Logged

jbrembat

  • Full Member
  • ***
  • Offline Offline
  • Posts: 177
Hey, Shewe the Sharpener . . .
« Reply #12 on: May 16, 2007, 02:37:15 pm »

Quote
From my unscientific testing, I could not conclude that it makes any observable difference. To claim that "prints will dramatically improve" would suggest that in a blind test, either the printer or a panel of viewers can EASILY tell the difference. I doubt very much either would be the case.
[a href=\"index.php?act=findpost&pid=117857\"][{POST_SNAPBACK}][/a]
The prints can be better not dramatically better, but some detail may be better viewed.
Of course the improvement depends on the scene and on the resampling algorithm too.
PhotoShop cannot be considered as top level resampler.

Jacopo
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Hey, Shewe the Sharpener . . .
« Reply #13 on: May 16, 2007, 04:35:09 pm »

Quote
PhotoShop cannot be considered as top level resampler.

Jacopo
[a href=\"index.php?act=findpost&pid=117962\"][{POST_SNAPBACK}][/a]

If you are talking about the time before Adobe developed Bi-Cubic Sharper and Bi-Cubic Smoother that may have been true, but since these algorithms were developed I think there is a broad spectrum of professional opinion that Photoshop resamples at least as well as anything else on the market, provided the operator uses those tools correctly.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Jack Varney

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 413
    • http://
Hey, Shewe the Sharpener . . .
« Reply #14 on: May 16, 2007, 06:19:36 pm »

Quote from: Schewe,May 15 2007, 04:34 PM
You are a lot better off using the image's "native resolution" (which is also a bit of a misnomer) and sharpening for the actual resolution sent to the printer.


How do you compute the "native resolution" of an image?
Logged
Jack Varney

Kenneth Sky

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 463
    • http://
Hey, Shewe the Sharpener . . .
« Reply #15 on: May 16, 2007, 06:19:38 pm »

Mark :
Are you saying there is no need for another sharpener plug-in in CS 3? I've just installed CS 3 on my iMac G5 (PPC) and am in a quandry about purchasing a sharpener plug-in for it. BTW the HP B9180 plug-in for CS2 is not transportable(?) to CS3. And with the extended version, it would appear I can get around video noise at high ISOs by merging several exposures without the need of a noise plug-in.
Ken
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Hey, Shewe the Sharpener . . .
« Reply #16 on: May 16, 2007, 07:25:41 pm »

Kenneth,

The remark I made three posts back is not about sharpening, it is about resampling images when "Resample" is checked in the Image Size Dialogue box - that is, changing the image dimensions at a fixed resolution (resolution defined as PPI) or fixing the dimensions and changing the PPI, or selecting any combination of dimensions and PPI which differs from that obtained without checking the resample box, but sticking with the "image native resolution" - i.e. whatever Photoshop calculates it to be as you finalize an image's dimensions. The term "BiCubic Sharper" is the name given to a resampling process for "down-rezzing" and BiCubic Smoother for "up-rezzing". I was saying that Photoshop's algorithms for resample are said to be as good as any.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

dabreeze

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 54
Hey, Shewe the Sharpener . . .
« Reply #17 on: May 16, 2007, 08:06:44 pm »

Well this topic started off originally as a discussion of a couple of different methods of capture sharpening both pioneered by Jeff Schewe and Bruce Frazier.

As to the need for sharpening plug-ins, I can't imagine being without Photokit SHarpener and its innovative three-pass sharpening methodology. I try and keep track of what the pro's pros are using and it would appear there's a very good consensus for some form of multiple pass sharpening that takes into account capture and output devices.

Jeff has an invaluable article/tutorial on three-pass sharpening that I had access to as an Epson Print Academy participant. I have it downloaded as a PDF and pass it on to my digital darkroom students religiously. I'm not sure where it can be accessed without a Print Academy participant's user name and password but maybe Jeff will make this informative article available to LL forum members?

An alternate would be to download the free trial version of PKS (www.pixelgenius.com) and with it the detailed 38 pg. PDF that covers much of the same ground.
Logged

dabreeze

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 54
Hey, Shewe the Sharpener . . .
« Reply #18 on: May 16, 2007, 08:26:58 pm »

Here's a recent image using all three sharpening passes (capture; creative; output). Especially useful on this hazy, late afternoon shot at Desert View in the Grand Canyon was the creative sharpening brush called Hazecutter. From what I understand it's a fairly aggressive edge sharper that also compensates for the increase in blue/cyan light from particulate matter in the air by adding a warming color correction.

Just one of the many tools as your disposal in PKS! I agree with Michael in that until something better comes along, they'll have to pry it out of my cold, dead hands!! Enjoy . . .

Logged

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Hey, Shewe the Sharpener . . .
« Reply #19 on: May 16, 2007, 08:33:15 pm »

Quote
Jeff has an invaluable article/tutorial on three-pass sharpening that I had access to as an Epson Print Academy participant.
[{POST_SNAPBACK}][/a]


Well, that particular PDF is still kinda locked to the Academy...but you can get the gist of my PDF by reading Bruce's article on Creative Pro. See: [a href=\"http://www.creativepro.com:80/story/feature/20357-1.html]Out of Gamut: Thoughts on a Sharpening Workflow[/url]
Logged
Pages: [1] 2   Go Up