Pages: 1 [2]   Go Down

Author Topic: Hey, Shewe the Sharpener . . .  (Read 15612 times)

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Hey, Shewe the Sharpener . . .
« Reply #20 on: May 16, 2007, 09:59:30 pm »

I would also highly recommend Bruce's book "Real World Image Sharpening with Adobe Photoshop CS2".
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Hey, Shewe the Sharpener . . .
« Reply #21 on: May 16, 2007, 10:49:36 pm »

Quote
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
[{POST_SNAPBACK}][/a]

"Native resolution" of Epson printers is often quoted as 720*2880 or 1440*5760 dots per inch, but these printers use error diffusion dithering and the relationship of dots to pixels is fairly indirect, but it takes many dots to make a pixel. The above resolution numbers describe the accuracy with which the individual dots making up the pixel are placed on the paper.

According to tests done by Bruce Fraser and reported in Image Sharpening with Adobe Photoshop CS2 there is a slight advantage to sending up to 480 ppi of suitably sharpened data to the printer, and above that value, there is very little benefit from a higher pixel resolution. Sending more than 720 ppi to the printer may actually degrade image quality.

If you print high contrast line pair targets at even multiples of the printer's native resolution (often said to be 360 ppi) there may be some slight advantage, but Bruce reported that the difference is minimal with most real world images and he simply sends the image to the printer at capture resolution in most cases.

If you do print high contrast line patterns at resolutions unequal to the native resolution of the printer, you often get an interference pattern from aliasing, much as you can get Moire when taking pictures of fabrics if your camera has a weak or no blur filter. [a href=\"http://www.rags-int-inc.com/]Rags Gardner[/url] has done some tests with the Epson 2200 printer and has determined that the native resolution of the  printer is actually 288 ppi rather than the oft quoted 360.

When an orchestra tunes up, it does so with an A tone from the oboe (440 Hz). If the instrument being tuned has a different frequency, a beat results and the pitch of the instrument is changed until there is no beat. The same principle applies in optics.

Bill
Logged

jbrembat

  • Full Member
  • ***
  • Offline Offline
  • Posts: 177
Hey, Shewe the Sharpener . . .
« Reply #22 on: May 17, 2007, 05:47:53 am »

Quote
"Native resolution" of Epson printers is often quoted as 720*2880 or 1440*5760 dots per inch, but these printers use error diffusion dithering and the relationship of dots to pixels is fairly indirect, but it takes many dots to make a pixel. The above resolution numbers describe the accuracy with which the individual dots making up the pixel are placed on the paper.

You are speaking about DPI. I'm speaking about PPI.
DPI=PPI (just as value, not as definition) only for contone printers or monitors.

Dithering is for DPI.


PrinterData is an application that connects to the printer driver and get the manufacturer settings. That's the way to know exactly at what PPI value the driver works.

I'm not so happy to define a "native resolution" as I experienced that PPI value may change depending on quality and on borderless settings.

Jacopo
Logged

jbrembat

  • Full Member
  • ***
  • Offline Offline
  • Posts: 177
Hey, Shewe the Sharpener . . .
« Reply #23 on: May 17, 2007, 05:54:48 am »

My general suggestion is:
 Set your printer preference
 Verify the driver PPI value
 Crop and Resample with best algorithm to the driver PPI
 Print

The benefits are sometime little? Who cares, I have to fix a workflow!

Jacopo
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Hey, Shewe the Sharpener . . .
« Reply #24 on: May 17, 2007, 10:33:51 am »

Quote
You are speaking about DPI. I'm speaking about PPI.
DPI=PPI (just as value, not as definition) only for contone printers or monitors.
[a href=\"index.php?act=findpost&pid=118142\"][{POST_SNAPBACK}][/a]

Please re-read my post and take some remedial reading lessons if necessary. I was clearly speaking about both DPI and PPI and their relationship. In a Postscript halftone, the relationship between the two is straight forward. With a color ink jet printer the details of the relationship are usually proprietary. I made no mention of contone printers or monitors, since they were not involved in the discussion.

Quote
Dithering is for DPI.
[a href=\"index.php?act=findpost&pid=118142\"][{POST_SNAPBACK}][/a]

I did not state if the dithering is applied to dots or pixels, so I don't know why you brought up this topic.

Quote
PrinterData is an application that connects to the printer driver and get the manufacturer settings. That's the way to know exactly at what PPI value the driver works.
[a href=\"index.php?act=findpost&pid=118142\"][{POST_SNAPBACK}][/a]

I did download PrinterData. It gives the PPI value for the default printer, in my case a HP Color Laserjet. It gives a value of 600 PPI for the printer, but does not tell me if the printer uses some type of error diffusion dither or half toning.

For my Epson 2200, it gives me a value of 360 PPI, which agrees with Bruce Fraser's figure in the sharpening book. However, this value is of no use to me, since I follow Bruce's advice (and also that of Jeff Schewe) and send the digital file, suitably sharpened, directly to the printer without any resampling.  There is no need to resample the image to 360 PPI-- Jeff has stated that this can do more harm than good

Quote
I'm not so happy to define a "native resolution" as I experienced that PPI value may change depending on quality and on borderless settings.
[a href=\"index.php?act=findpost&pid=118142\"][{POST_SNAPBACK}][/a]

Native resolution of what? The capture? The printer?

Bill
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Hey, Shewe the Sharpener . . .
« Reply #25 on: May 17, 2007, 11:29:11 am »

Quote
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 .
[a href=\"index.php?act=findpost&pid=118050\"][{POST_SNAPBACK}][/a]

Dabreeze,

A great picture, but without any before and after comparison, I do not know if this is due to your photographic skills or to the magic of PKSharpener.  

According to Bruce's sharpening book, Real World Image Sharpening with Adobe Phtotoshop CS2, the haze cutting brush is an adaption of his Depth of Field Brush. The latter uses an unsharp mask with amount = 500%, radius = 4 pixels and threshold = 0 followed by use of the highpass filter with a radius of 25 pixels. This is done on a layer whose blending mode is set to Overlay, Opacity = 50%, and with blend-if sliders set to give protection to the shadows and highlights. He considers it to be a combination between conventional sharpening and a mid-tone contrast boost. The haze brush adds a warming filter. That's what he described in the book, but he might not divulge any trade secrets used in the commercial product.

PKSharpener is a great product and time saver, but I think that it is worthwhile to buy the sharpening book to gain an understanding of the principles of Bruce's sharpening workflow even if you are a happy user of the commercial product.

To boost midtone contrast Bruce used the High Pass filter with a high radius setting (30-50 pixels) in an layer with a blending mode of Overlay. Others use the conventional unsharp mask with a high radius and low amount. Dan Margulis calls this "hiraloam". For your Grand Canyon shots you might want to take a look at his book Photoshop Lab Color.

Bill
Logged

jbrembat

  • Full Member
  • ***
  • Offline Offline
  • Posts: 177
Hey, Shewe the Sharpener . . .
« Reply #26 on: May 17, 2007, 12:13:57 pm »

Quote
Please re-read my post and take some remedial reading lessons if necessary.
Bill
[a href=\"index.php?act=findpost&pid=118184\"][{POST_SNAPBACK}][/a]

When people start with offences....he is in deep difficulty.

You quoted:
Quote
I know that Epson fixed 720 PPI
and then you start to speak in DPI terms with some magic black to PPI.....

Quote
but does not tell me if the printer uses some type of error diffusion dither or half toning.
This information is not exposed by the printer driver, sorry. You may ask to Microsoft and printer manufacturer to implement it, but untill now I'm not able to pick up such info.

Quote
"Native resolution" of Epson printers is often quoted as 720*2880 or 1440*5760 dots per inch
Quote
If you print high contrast line pair targets at even multiples of the printer's native resolution (often said to be 360 ppi)
You use the same term for different things.

I'm intersted only in PPI value used by the the driver. I can't have any influence in PPI to DPI process.  

Quote
However, this value is of no use to me, since I follow Bruce's advice (and also that of Jeff Schewe) and send the digital file, suitably sharpened, directly to the printer without any resampling
You are free to do every thing you like!

Quote
There is no need to resample the image to 360 PPI-- Jeff has stated that this can do more harm than good.
I don't agree at all. Jeff statment han no any scientific basis: he is stating that nearest neighbor is better than a smart resampling algorithm!

Jacopo
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Hey, Shewe the Sharpener . . .
« Reply #27 on: May 17, 2007, 02:35:06 pm »

Quote
There is no need to resample the image to 360 PPI-- Jeff has stated that this can do more harm than good
Native resolution of what? The capture? The printer?

Bill
[a href=\"index.php?act=findpost&pid=118184\"][{POST_SNAPBACK}][/a]

Bill, There is no "need" to resample to 360, but depending on when and how done it could be marginally helpful if only because it alligns the PK Output Sharpener setting with the corresponding PPI in the image file. I've tested this issue with and without PK Sharpener with and without an exact 360 going to the printer - without a loupe the differences don't hit you in the face; staring very hard at the comparison photos sometimes one sees a slight advantage to lining everything up at 360 for an Epson wide-format printer. The kind of detail in the image probably affects these comparisons one way or another.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

dabreeze

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 54
Hey, Shewe the Sharpener . . .
« Reply #28 on: May 18, 2007, 12:05:56 am »

Thanks Bill for the kind comments and constructive criticism. Good info on Bruce's books and for sure i should get it and read it. I've read all of Ron Bigelow's fine series on sharpening as well as jeff's and PKS' PDFs. So I understand the basics fairly well.

I used both the haze cutter in certain areas of the middle ground on this and as I do on almost every image, made a visible stamp of the underlying layers and applied Michael's Local Contrast Enhancement (USM @ 20, 50, 0) and vary the opacity to taste. With this I'll occasionally use the blend if mode as well to prevent too much additional black from being introduced.

Basically, this is almost interchangeable with the Mac Holbert high pass method of mid-tone boost which I think is almost the same as the Bruce method you described.

You're very definitely right though about getting Bruce's book eventually and gaining even more knowledge of this most crucial of post-processing skills.

As for the question of photographic skill vs. PP knowledge, you guessed it, it's PKS alone that made the magic!! LOL!! Even more reason to rush out and get it!

really, i think both are part and parcel of the same ability these days as a digital photographer. ansel was a genius in the field and moreso in the darkroom. without a good starting point (composition, light, exposure, focus, subject) you're sunk. Good PP is just surmounting  the limitations of the media and optimizing the beauty that you found before you.

I'm a traditionalist in that I believe the true art of the landscape photographer is finding the moment (and the huge effort and skill set that entails) and capturing it well. Nature and the universe are the true master artists and too much creative license, for my taste, is just too self-absorbed. creative license has its place in almost all aspects of photography and i love and applaud people's art. it's just that when it comes to the landscape, i think we should be attentive to the beauty that exists.

Sure, I'm guilty as any in sometimes over-romanticizing nature. but we've all been trained over the last three decades to expect our landscapes to have punch, immediacy, vibrancy, and saleability (thank you velvia!) and it may be some time before the infinitely more real and documentary (variable and more accurate WB is just one example), wider gamut, and arguably less saturated potential of digital comes to be accepted as the standard.

Until then, large format film photographers and their film stocks will continue to set the standard and we will romanticize to the point of velvia and no further (!!) lest we be accused of overcooking in PS!! Some day we may return to the less saturated, wider gamut film color style of say Elliot Porter whose seminal work in the 40s-60s seems these days much flatter and less immediate. I saw a wonderful retrospective exhibition in LA last year and that was my first impression. After that, I began to appreciate the incredible subtlety and nuanced color palette but once trained to velvia, it's hard to untrain the mind's and eye's expectations!!

We could probably start a whole thread on this topic alone . . . But this one's been hijacked enough from my original post what the heck!! LOL
« Last Edit: May 18, 2007, 12:07:33 am by dabreeze »
Logged
Pages: 1 [2]   Go Up