Sharpening Negative Scans-What is the Best Approach

[JPlomley]

I'm interested in hearing from the experts the approaches they use to sharpen their scanned negs with printing as the endpoint. Positives are no problem, but I find that grain emphasis is much more of a problem with negatives using techniques such as high pass sharpening, edge sharpening, or even luminance sharpening. Does anyone have a magic bullet?

[Ed Richards]

Color or black and white?

[JPlomley]

Ed,

Both.

Thanks,
Jeff

[David Luttmann]

I'm interested in hearing from the experts the approaches they use to sharpen their scanned negs with printing as the endpoint. Positives are no problem, but I find that grain emphasis is much more of a problem with negatives using techniques such as high pass sharpening, edge sharpening, or even luminance sharpening. Does anyone have a magic bullet?

The best you'll be able to do is to decrease your USM strength, Inrease the pixel width, and start with a higher threshold like 2-5 as opposed to 0 or 1. Depending on your print size though, grain from a neg shouldn't be a problem other than at 100% on your screen.

[Kirk Gittings]

Are we talking 4x5? How is it scanned? How large are you printing? Glossy or mat. All these things matter. Generally, what I do is....If you start with a small USM as David suggests followed by a High Pass Sharpen at the end of the workflow (In a separate layer where you can control its final application for different size prints) you can control grain amplificationa nd still get a good sharpening. With grainy films like Tri-x in HC1110, I can get a sharper appearing print (vs. traditional silver) with no more grain than is apparent than in a silver print at the same size.

[JPlomley]

Kirk,

Apologies for not clarifying. Film dimensions are 4x5 and 6x7. I print on an Epson 7800 and my portfolio dimensions are 20" on the short side. I have access to an Epson 9800 when I wish to print larger. Scans are performed on either an Imacon 949 or Cezanne Elite (am hoping to try the Creo IQ3 soon). Current paper choices are Harman Gloss FB Al for B&W and Hahnemuhle Fine Art Pearl for color work (although that may change in the near future). I will also print on 100% cotton canvas if the image warrants that artistic effect. In my experience the natural texture/weave of the canvas reduces the amount of sharpening required.

[JPlomley]

I might also add that I am scanning XP2 for the 6x7 work.

[Gary Beasley]

I find when scanning B&W the unsharp mask needs the radius set large enough to exceed the grain size a few times and then run the percent slider up till the masking effect just shows and back off a quarter of the way. I do this with the image at 100% size so I can see the effects easier.

[bob carnie]

For colour work and rgb black/white I sharpen in the L channel of Lab, seems to work real nice.
I use a very large setting and use the opacity slider to taste.

[Bob McCarthy]

If I'm working on scans, I use the PK Sharpener plugin in CS3 which is fairly automatic.

If I'm sharpening digital I sharpen using the raw converter (Raw Developer). I sharpen using Hybrid or DoG.

Bob

[David Luttmann]

For colour work and rgb black/white I sharpen in the L channel of Lab, seems to work real nice.
I use a very large setting and use the opacity slider to taste.


Bob,

You can speed things up a bit by working in RGB and then using the Edit/Fade/Luminosity after the sharpen. This gives pretty much the same benefit as working in Lab, without the destructive conversion process of going rgb > Lab > rgb.

[bob carnie]

David

I am not seeing any destructive issues going back and forth.
I am doing some tests images where no conversion is done, and then multiple conversions . 2- 5 -10. to see the effects.

I will keep your suggestion in mind and give it a go. With your method can you affect detail only without affecting the colours?

Doing it in L is much like how I print in B&W with split contrast printing.. ie using the 5 filter for contrast and sharpness. As well dodging and burning is very intuitive for me in a duplicate L channel. I like the way I can affect the file this way.

I recieve files daily to print and I find that skys and open colour areas are generally the areas that are so often oversharpened ,This seems to be the biggest problem area I see when files are submitted to me to make large prints, large patterns of texture.
I know there are many ways to not have this happen, but unfortunately/fortunately I am recieving flattened tiffs and what has gone on before the file reaching me is beyond my capabilities of predicting and or correcting.

Most experiened photographers are not willing to submit unflattened psd files as all their magic is revealed, others are willing to just submit a raw file or neg/pos and let me work on their file.* we have decided to have weekend workshops at our lab where each participant brings in their computer , we profile the moniters to a paper, and then the group will go through step by step methods, of working on files and printing on the fly as we progress during the day, the goal is not to force our workflow on anyone but rather share some of our methods and as well learn from our clients different approaches to digital workflow. We see this as a win win situation for our Lab and our clients. The only prerequsites will be, CS3 on their computer, some skills and a willingness to share ideas*
I can see this happening in the spring and continuing as long as there is interest with our clients.
I would love to do this as a master class and open our doors for this kind of exchange, but all egos would have to be hung at the back of the room. This would be in my opinion a great learning tool for anyone willing to take the plunge, Imagine all the different approaches/styles/myths that would be revealed over a few days.
I will see how these meetings go with my group of clients and maybe propose a more detailed event that people here could plan on attending.


I think I read a post by K Gittings, a couple of years back where he said that he took basically one whole year off to study PS , maybe I'm wrong and it wasn't Kirk, but his comment then seemed overstated, but now I think he shoud of said taken a couple of years. All the different nuances of this PS are mind numbing but at the same time , I seem to understand PS and its different approaches to creating an image but I wish my fingers would just do what my brain is telling them to do.

I really feel like a kid again back at photoschool, the only difference is I have a few dollars in my pocket and there seems to be a rather large empty spot on the back of my head.



Bob

[Donald Miller]

http://www.pixelgenius.com/tips/schewe-sharpening.pdf

[Dan Schwartz]

Bob, where are you located? If you're in NYC or Philly, sign me up... And I'll use your lab, too!

By the way, Do you use a Chromira, Lambda, or Epsilon for your photo output?



Doing it in L is much like how I print in B&W with split contrast printing.. ie using the 5 filter for contrast and sharpness. As well dodging and burning is very intuitive for me in a duplicate L channel. I like the way I can affect the file this way.

I receive files daily to print and I find that skys and open colour areas are generally the areas that are so often oversharpened ,This seems to be the biggest problem area I see when files are submitted to me to make large prints, large patterns of texture.
I know there are many ways to not have this happen, but unfortunately/fortunately I am recieving flattened tiffs and what has gone on before the file reaching me is beyond my capabilities of predicting and or correcting.

Most experiened photographers are not willing to submit unflattened psd files as all their magic is revealed, others are willing to just submit a raw file or neg/pos and let me work on their file.* we have decided to have weekend workshops at our lab where each participant brings in their computer , we profile the moniters to a paper, and then the group will go through step by step methods, of working on files and printing on the fly as we progress during the day, the goal is not to force our workflow on anyone but rather share some of our methods and as well learn from our clients different approaches to digital workflow. We see this as a win win situation for our Lab and our clients. The only prerequsites will be, CS3 on their computer, some skills and a willingness to share ideas*
I can see this happening in the spring and continuing as long as there is interest with our clients.
Bob

[Brian Ellis]

"I think I read a post by K Gittings, a couple of years back where he said that he took basically one whole year off to study PS , maybe I'm wrong and it wasn't Kirk, but his comment then seemed overstated, but now I think he shoud of said taken a couple of years."

That may have been me or maybe several of us have said that. But I've said that it took me a year to get to a point where I thought I knew what I was doing with Photoshop, though I didn't spend a solid year doing nothing else. It's usually said when people tell me that printing digitally is just a matter of pushing a button, that it doesn't reflect "the hand of the artist," or that digital prints aren't any good because they saw one done by their next door neighbor who just got his P&S digital camera and they didn't like it. And of course Photoshop is a continuing learning process. I've been working through Katrin Eismann's book "Photoshop Masking and Compositing" lately. It's a terrific book, the best book I've found for someone who already knows a fair amount about Photoshop.

I still sharpen with USM in Photoshop CS2. I view at 100%, leave threshold at zero, and vary the percentage. I seldom apply the same percentage to the entire photograph. If there is sky, a body of placid water, or some other similar large, smooth area I select it and leave it out of the area to be sharpened. I then select other areas depending on the degree of detail in them and the importance of detail to the image and apply differing percentages to them. When viewing at 100% I usually sharpen just to a point that looks like slightly more than I would want in the print and then back down a little. To me the single worst offense that can be committed when printing digitally is oversharpening, it just screams out "I made this print digitally and I didn't know what I was doing" so if I err I'd rather it be on the side of slightly under-sharpening.

[bob carnie]

Hi Dan

If you sign up at www.elevatordigital.ca on our mailing list you will get all our propaganda.

I have thought seriously about this for a couple of years now, We are now close to being able to make a low cost*supported by a few key suppliers of materials* high energy or better said creative learning long weekend that brings workers from different backgrounds together and share workflow ideas.

Onsite we have many output digital devices, wet and inkjet, and there is a good interest to start these weekends.

Toronto is an amazing city and we have brought large groups to our Lab in the past and I think that this idea is a good one for hard core PS workflow learning using up to date output devices networked to your personal computers.
I can see each worker bringing some files and as a group working on each file in a step by step method that is then recorded and each person taking away workprints showing multiple methods of getting there.

I will put some more thought into this idea, I can tell you that I have already a core group of about 10 ready to roll in May on this, Dan if you want to be part of this group you can email me bob@elevatordigital.ca to confirm your intent.

Basic things

Costs--- minimal to none-- we may bring in expert leaders that the group will need to share a day rate.
Materials-- we will supply with our suppliers-- My lab is predominently Fuji FCa and Harmon product orientated.*there may be a small fee to spead the cost out that go beyond what our suppliers are willing to give us*
Hotels-- I can get a group rate around $119 per night at a nearby hotel.
Food/Refreshments--you are on your own , but we usually bring in stuff and everyone chips in, if weather is good we have a patio and bbq area in back.
CS3 on your computer with lots of RAM.
A keen willingness to learn,teach , critique with style , and no egos allowed.
I see these as being 12-14 hour days so we better learn to get along.

We will provide the facility, some coffee and buns , in the morning , profiling to our different machines and the lab rats to run around and make the output happen live and ready for review each day.

My goal would be to try to see as many methods of skinning the cat and not to try to champion one method over the other. One camera device over the other , but basically simplify workflows for exhibition print criterea.

Who will lead this ??
For now 3 of us here are taking extensive workshops on PS. All three of us have a vested interest in 42 Industrial *our location* and would probably like to start the ball rolling in the way we have been taught by others.
I am hoping within the first few groups a few king poobas will emerge that will take their hand at overseeing each learning group.They will probably emerge from the Toronto GTA but
I can visualize that there will be some talent emerge from other parts unknown at yet that we think that should be brought back for a fee, because of their unique teaching abilities. There are natural born teachers , and then there are those like myself that are better working alone.

For me I work 7days a week and take my time off in spurts, I would love to have something like this happening here on weekends.
Seeing an unfamiliar adjustment being done to a file and then printed in a no pressure enviournment would be awesome.

[bob carnie]

Hi Dan

sorry , We use Lambda at 400ppi, Epson 9600 , and Cannon 9000 which I think will print 16bit files *haven't tried this yet*. Fuji papers, Harmon Silver Gelatin for Lambda , and a variety of papers for the Cannon and Epson units.

Brian

Katrin is fantastic, I have booked a one on one day with her in Feb to help me understand a boatload of things that I cannot make my fingers do. She explains the workflow in a very wonderful way.
Dan Margulis is in my opinion a super teacher and his methods are what I am using now . I flunked his first workshop*and I am taking it again as well as the advanced*
Dan is much harder to understand but what he dosen't know about PS ,probably dosen't matter

[Kirk Gittings]

I think I read a post by K Gittings, a couple of years back where he said that he took basically one whole year off to study PS , maybe I'm wrong and it wasn't Kirk, but his comment then seemed overstated, but now I think he shoud of said taken a couple of years. All the different nuances of this PS are mind numbing but at the same time , I seem to understand PS and its different approaches to creating an image but I wish my fingers would just do what my brain is telling them to do.

That may have been me. I did do something like this preparing some images for my retrospective a couple of years ago. Of course I could not take a whole year off, but I scaled my business way back, took no holidays and worked about 12-18 hours a day, 7 days a week for 18 months. It almost cost me my marriage. It was more intense and demanding than my MFA. I learned a ton, but the education continues at a very high constant level. As an old analogue guy, it did not come naturally, like pulling wisdom teeth. I've had the great fortune to be tutored by friends who are experts in the field, one of whom informed a few months back that I had now surpassed his knowledge base and was now tutoring him (only because he had become more involved in administration at SAIC and had been away from image making/teaching for over a year and he had gotten behind).

Basically this technology demands constant continuing education.

[Lenny Eiger]

Kirk,
Film dimensions are 4x5 and 6x7. I print on an Epson 7800 and my portfolio dimensions are 20" on the short side. I have access to an Epson 9800 when I wish to print larger. Scans are performed on either an Imacon 949 or Cezanne Elite (am hoping to try the Creo IQ3 soon). Current paper choices are Harman Gloss FB Al for B&W and Hahnemuhle Fine Art Pearl for color work (although that may change in the near future). I will also print on 100% cotton canvas if the image warrants that artistic effect. In my experience the natural texture/weave of the canvas reduces the amount of sharpening required.

I'm sorry, but I am going to suggest the opposite technique. I prefer to sharpen very little, only enough to crisp things up. I usually leave off the threshold, but set the radius first, usually to .2 for my scanner (a Premier drum). For lesser scanners, like ccd's I might go to .3 or .4. Never more than .8 even for the worst. The I use the amount to whatever I think it needs, anywhere from 150 to 350.

I don't sharpen things I don't want to. I have used edge sharpening, but I usually just have a mask that blocks things like a sky, water, etc. from the sharpening.

Lenny
EigerStudios