Railfan Technology > Learning Photoshop

I'm stumbling through the Photoshop mine field, trying to learn the "Image->Adjustments" menu. These two images are two examples.

1) The initial scan, using my Minolta Dimage Scan Dual IV and Vuescan. Obviously under exposed, more than the original slide.

2) After tweaking with Auto Levels and Auto Color, in different order. I've tried Levels, and Curves, and Exposure, and even Color Balance. This one is Auto Color, then Auto levels, then Gamma in Exposure.

What would you do, given the scan and Photoshop?

I have hundreds to do for a book project, so I'm trying to develop a workflow that can be automated to some degree.

Don Strack

---

---

One thing I have found that in scanning old slides is that it is extremely difficult to do batch edits and make things come out looking good. I would rather take a couple extra minutes per slide and tweak each one individually.

I have read people recommend and I have used shadows/highlights as the simplest and quickest way for many images. Use the advanced settings and you can tweak it so it only affects the darkest areas. However, there is no adjustment layer so that is a negative. Watch the highlight command as it seem to tend to "sharpen" the image also.????? doubt you can automate that but it works pretty quickly.

Bob

Since it's for publication, it appears that converting it to grayscale improves the image quite a bit.

Don Strack

---

This is one of those instances that Lightroom would be a better tool than Photoshop.

It's just a thought, but you can download a 30 day trial of Lightroom from Adobe's website. I find the workflow is much easier to navigate, and editing, or rather "developing" goes alot quicker.

donstrack Wrote:
-------------------------------------------------------
> I'm stumbling through the Photoshop mine field,
> trying to learn the "Image->Adjustments" menu.
> These two images are two examples.

I think you are starting in the right place. I usually adjust the levels first and then go for color balance and then anything else that needs a tweak.
>
>
>
> 2) After tweaking with Auto Levels and Auto Color,
> in different order. I've tried Levels, and Curves,
> and Exposure, and even Color Balance. This one is
> Auto Color, then Auto levels, then Gamma in
> Exposure.

I don't put alot of faith in the Auto Levels and Auto Color. I'll try them but most of the time I end up making some additional adjustments. Sometimes if you have a difficult situation they can get you in the ballpark real quick. I hardly ever hit them and end up happy with the results unless I'm doing something that I don't really care much about.

With my camera I shoot raw and I can often do some batch corrections that come out pretty good but I still look at every one of them and usually do a slight tweak. With slide I would imagine that it would be even more difficult to batch correct unless the batch was all shot with the same lighting and about the same time of day and with the same type of film, maybe even the same batch of film. With all of the possibilities and variations, I think it might be pretty hard to automate the work flow.
>
>
I'm glad you ask the question. I'm looking forward to reading the suggestions from some of the "great ones" we have on this forum.

John

Don,
are you doing the "prepress" on the files, or is the publisher going to be doing this? I've been scanning a lot of slides as well, and sought the advice of Dale Sanders, who is a slide-scanning machine. He advised scanning everything in as large a file size as you can (though 4000 ppi should be fine for virtually everything, as 5400ppi is going beyond the resolution of the slide itself), and instead of making it look "perfect" on the screen, adjust the values to give you a very flat looking image--no blown-out highlights, no dark dark shadows. Then you give the production person a file they can adjust to meet their pre-press needs. . otherwise, if it ain't there and they need it when making the separations, it can't be created from nothing.

Don,

I think the above results you got were very good especially considering the axial light on the image. It's impossible to make a shot like that look perfect. There will always be a compromise between the highlight and shadow detail.

I also concur with silagi that trying to find batch processes for correcting scans of older slides is going to be almost impossible. I've found that hand tweaking them on a case-by-case basis is the only way to get decent results.

Above all though, make sure you are saving your files out of the scanner and after editing in some sort of uncompressed or lossless compression format like PSD (native Photoshop) or TIF files. If you save them in JPEG format and have to make more adjustments later (which you most likely will), you will lose some of the image integrity just through opening and re-saving it. Surely your pre-press person will want them saved in the CMYK color space eventually which, if you don't do it now, means they will all be opened, changed to CMYK and saved again. You can minimize the destruction of data by using an uncompressed or lossless format.

Mike

Not doing any prepress. I'm just scanning the negs and slides, hoping to add them to the package of my glorious words, to get a publisher excited about the project. Most of the stuff is coming across just fine, especially the B&W, but I'm trying to figure out what to do with a bunch of underexposed images, since I've never really been that great of a photographer. That, plus the fact that most cameras tend to underexpose both K25 and K64 by a slim bit. I've had two editors tell me that they always have to lighten almost every slide they prepare for publication.

Here is a nice video tutorial I found at the Adobe site:

http://www.adobe.com/designcenter/video_workshop/?id=vid0011

This video, and several of the others at the same location make me start thinking that I might be using the wrong tool here with my iMac computer. I'm not an artist or a graphic designer. All I want to do is scan color and B&W images, fix the colors and tones for pleasant viewing, and move on. Nothing creative is required. I'll give Lightroom a try. The description at Adobe shows me that it is aimed at photographers, and fixing any problems between the camera (or scanned image) and the final image.

Don Strack

trainjunkie Wrote:
-------------------------------------------------------

> Above all though, make sure you are saving your
> files out of the scanner and after editing in some
> sort of uncompressed or lossless compression
> format like PSD (native Photoshop) or TIF files.

I only save the images as uncompressed TIF. My Minolta slide scanner scans at 4800ppi, and the Epson 4990 flatbed scans at 3200ppi, so I think my resolution is good. I'm getting files in the 35-45 meg range. The B&W are coming in at about 11 meg. I only convert to JPG, at 800px wide and 300ppi, for the shots I share here at TO.

Don Strack

Here's an example of what I'm talking about when discussing "flat" scans for the benefit of the prepress artist. The first photo is the scan as it came out of photoshop as a TIFF file--tremendously flat and lifeless. But, if you want a "raw file" the prepress person can work with, this one is far better to submit than the second photo, which is the flat scan punched up in photoshop for presentation on-line. If you were to submit the second one to a publisher, it'd be tough for them to pull the required information from the file, as you'd already tweaked the highlights and shadows to the point of being unable to pull them back out.

So, scan flat, save it, and then do your tweaks and such on a file that you'll post on line or send to your personal printer, etc. When it come to submitting the image to a publisher, give em the flat one.

---

---

Don,

Since you aren't doing the prepress work, I completely agree with Jackpot. Beware of using photographer's tools to produce images for press. A good production person will be able to make sure the image yields the maximum amount of information when the ink hits the paper in the relatively low dynamic range environment of printing presses. If you tweak the image to look good on a monitor or even in a photographic print, it may not translate well on a printing press. The images from digital cameras and high quality scans (and film for that matter) have a much broader dynamic range than your monitor can display or your printer can print. But the offset press has an even lower dynamic range than that so it's possible to tweak an image that looks great on screen or from an inkjet photo printer but looks like mud from a printing press.

As Jackpot said, it's probably best to give the publisher or production person an unedited or minimally edited scan and leave the rest to them. I would suggest making copies of all your image scans, tweak the copies using whatever tools you find easiest to use, and use those copies to pitch the project to a publisher while letting them know that you have a complete set of "raw" scans available for production. I've done a lot of prepress production work and I can assure you he is right about it better to have a rough image with a lot of data to work with than one that looks good on screen but is missing vital data to make it useful in the offset process.

Of course, we still haven't answered you on what tool would be the best for tweaking the images you want to use to pitch the project. I only use Photoshop so that's all I know. If you have a group of scans that need similar corrections (like Levels, Unsharp Mask, Image Size, etc.), it is possible to create an automated batch process in Photoshop then execute that series of instructions (Actions) on that batch of images. You record the actions using the "Action" palette under the "Window" menu item. Then once you have an Action recorded, you "play" it using the "Automate" > "Batch" command under the File menu. Google "photoshop automate batch" (without the quotes) and you'll get some instructions on how to create automated batch processes. Hope this helps.

Mike

You might want to straighten the image just a bit. Rotate it about 1 degree clockwise for a start.

JMM

RS12394 Wrote:
-------------------------------------------------------
> You might want to straighten the image just a bit.
> Rotate it about 1 degree clockwise for a start.
>
> JMM

it's on a .3 percent grade, so supposed to look that way!!

I'm in the process of scanning thousands of my slides, and here's the workflow I'm using:

1) Scan big (I have a Minolta Dimage 5400) as a TIF.

2) Rotate if necessary (a degree or two), then trim away excess black borders. Sometimes if I've rotated a shot and discover that trimming all the black will cut out an important part of the image, I'll leave some of the black border next to the sky or next to some non-descript ground and fill in the black using the cloning tool.

3) Hit auto levels, just for kicks. Sometimes when I hit auto levels I'll notice that there's a lot more magenta in the photo than what I originally thought there was -- the comparison helps. Auto levels will often give you a "brassy" look, like a roll of bad Kodachrome. Auto levels gives me an idea of what to shoot for when I do adjustments, but I use it only as a guide.

4) Undo autolevels, then go into levels. I find that I normally give the right slider a tweak to the left to open the blacks a bit, and move the middle slider to the left to brighten the overall image -- take care that your blacks don't go gray. This takes care of almost all color issues.

5) The scanner does scan a little towards magenta sometimes, so I'll only choose the "R" channel (red) and give the left slider the slightest of movement to the right. *Use carefully* or things will get green.

6) Happy with color? Next go to the unsharp mask. About 100% works for full-resolution images.

7) Save with a new name.

8) I have an action set up to make four copies of each image and put each copy in a separate folder. The first part of the action converts the full-size image to a full-size JPEG with no compression (for uploading to on-line print processors like Shutterfly or on-demand book publishers like Blurb); the next thing the action does is scale the photo to 1024 pixels across and 72 d.p.i. and saves it as an uncompressed JPEG (for projection on a 1024x768 projector) (verticals get scaled at 768 pixels high); next the action takes that projector image and saves it as a JPEG with medium compression with no change to the dimensions (for posting on-line); and finally the action scales it down to 3 inches across at 72 d.p.i. for a thumbnail (should I need one -- I'm finding out more and more that I don't need this, but haven't gotten around to eliminating this step from the action). All the images created by the action wind up with the same name, but the action saves them to different folders.

This is a quick way to get 95% quality out of an image. You can spend a lot more time per image getting that final 5%, but you need to judge for yourself what time/quality ratio is best for you. If I discover that my 95% isn't good enough somewhere down the road, I still have the original scan to back to and I can start over.

Steve Barry
Railfan & Railroad

Thanks Steve. I have figured out the Autolevels thing as a reference, just to see the the robot thinks the photo should look like. I then do the real work in Levels. I really like the way black&white responds to adjustments using Levels. Wayne Fulton's "A Few Scanning Tips" has been a great help with my basic understanding: http://ww.scantips.com.

I've tried doing an action to change each image to 800px wide for my web albums, and for use here at TO. But the problem is that I have to tweak the folder location each time I use the action. I need to learn how to set up an action that will work no matter what folder I start with, like a wildcard in Windows, which OSX apparently is immune to.

Why 72dpi for digital viewing? Wayne Fulton says "Bah Humbug" to 72dpi. I've noticed a real difference in clarity and resolution of images here at TO, with digital originals appearing to be much better. Mine are a little grainy, and they are 300dpi.

I'll start a new thread to answer this question.

Don Strack
http://utahrails.net

It's a matter of math. The d.p.i.isn't the issue as much as the pixel dimensions. Here's why:

PC monitors display at 72 d.p.i. (Macs at 96, I think, but we'll stick with PC's for now). A digital image sized at 10 inches across at 72 d.p.i. will appear 10 inches across on your PC monitor. However, an image sized at 10 inches across at 300 d.p.i. (3000 pixels across -- 10x300) will *print* at 10 inches across but display on a monitor at 3000/72 inches across (42 inches across -- you better have big monitor). A pixel is a pixel when it comes to display. Now a 3x5 print at 300 d.p.i. will print at 3x5 but display on a monitor at 21 inches across (now we're into the range where high-end monitors can display the entire image). So, getting back to math, a 3x5 300 d.p.i. image will display exactly the same on a monitor as a 21-inch 72 d.p.i. inmage -- pixel-wise, they're exactly the same (1500 pixels across).

For projection, it's best to match your image to the projector's resolution. My projector is a 1024x768. It doesn't matter what resolution you use as long as the horizontal dimension is 1024 pixels across (verticals should be 768 pixels high). Anything over these sizes will be interpolated by the projector and the projector is still going to put 1024 pixels across on the screen -- it';s just a matter of if you want to get the best looking image at 1024 or if you trust the projector to pick its favorite pixels. I use uncompressed JPEG's for projection because you lose virtually none of the image to JPEG artifacts, but teh image is still small enough that it won't bog down the computer processor (especially in multi-media slide shows where slide changes happen frequently).

Does this help?

Steve Barry
Railfan & Railroad