An Imaging Workflow

1) The Scan  2) Colour Correction  3) Spots  4) Choices  5) Cropping  

6) Preparing for Output  7) Re-Sizing  8) Archiving  9) Last Words

If you have done much with digital image editors, like Photoshop, you’ve come to appreciate the ease with which you can colour correct, de-spot, and even drastically alter your pictures. You can be fairly experimental in Photoshop secure in the knowledge a click on the History Pallet will return you to an earlier state if you wish. Problems can arise if you’ve gone too far though, so I’ve written a brief description of the way I work in Photoshop, as this method allows easy retreats, and the ability to re-purpose an image for many different uses.

Some of this is in other articles and I’ve described much of it on the rec.photo.equipment usenet forums. I don’t claim that this is the Only way to work, or that I invented this workflow. This is merely a starting place for you to develop your own workflow.

The Scan

Always scan at the optical resolution of your scanner. There is no reason to use any other setting. More than the resolution will just create gigantic files with extra pixels. Less will create files that cannot be fully exploited in the future. You may currently only need a small file for the web, but next week you might decide to make a small print, and the week after perhaps a large print, or you might decide to use an element from this file in another picture. By scanning at optical resolution, you will create a file that can serve all these purposes. See my previous articles for further information on scanning.

Colour Correction

If you’ve read the previous articles your image has been scanned into high bit colour - 16 bits per channel in photoshop PSD. You have to start with colour correction as there is precious little else the program can do until you reduce the colour depth.

Start by saving the file with a unique name, followed by a letter that informs you that this image is a raw scan. I use the letter x. KSinkX is the name of the raw scan for the picture Kitchen Sink.

With the image safely tucked away where we can find it later, we can now perform magic with Levels and Curves to either correct the colour to what it always Should have been, or change it to what it could never have been.

Since we are working in high bit colour, we can afford to make several small simple moves instead of one complex move that might not work as well. With 4096 levels of each colour, we will have shades to spare.

Start with the most major corrections, such as an overall blue cast. Move the sliders until it’s gone and click okay. This is now a discreet step, and you can return to it later.

Now look for other problems. Skin tones a bit too magenta? Open Levels again and try to improve them - one way to do this is to add green - be careful how much though. Once you have the skin colour pretty close to where you want it click okay again.

Now go to the history pallet. You will have 3 states there. Import, Levels, Levels. Click back to Import. Have you improved the overall look of the image with your changes, or have you simply changed the look? Go back and forth a few times to be sure. All okay? Save the file to a new name: KsinkY1 will do nicely.

You are now ready to do the smaller colour corrections. There may be areas you have to mask for them. Every couple steps save the file to a variant on your new name KSinkY2, KSinkY3, etc. until you have the colours as you want them.

Now open KSinkX and tile it with your current version. Quite a difference. Overall the picture looks better, but have some areas gotten worse? Are those areas ones you specifically changed to improve them? If something has gone wrong you have all the versions at your fingertips, and can return to one that looks better, without losing Any of the versions you have made.

Curves and Levels settings can be saved too. If you made a couple large complex corrections on one image - say from a whole roll shot under the wrong light - you can save the corrections and apply them to every image. Think ahead, and you can save many hours of work.

Once the colours look right convert the file to 8 bit and save it as KSinkA1.

Spots

Now is it time to get rid of all those annoying dust spots on the picture. You know they are there. No negative or slide comes out perfect.

Start by creating a new layer above your image. Name this layer SPOTS, and make it active. Use the Rubberstamp tool to clone out small spots. Adjust your brush to cover the smallest dots in one click. Somewhat larger ones should be decreased from all sides to blend them better. The largest spots may need to be removed with the masking tool.

Make the background layer active. Create a selection around part or all of the area you wish to hide. Move the selection to another area with the same texture and colour. Feather your selection by one or two pixels. Hit Control/J. This creates a new layer with nothing on it but the copied area. Use the Move tool to place this over the area you need hidden. You may have to do some colour correction to the new piece, and you may have to blend in parts with the rubberstamp tool.

If you don’t have much RAM you may want to combine the new layer with your Spots Layer. If you have several large areas that need work, you could end up with a huge file, if you don’t combine. Do not merge them with the background (original image) however. We are keeping this un-altered.

Save the file to KSinkA1 frequently. We are not making any changes we cannot easily remove with the eraser tool, so there is no need to archive different versions.

Choices

Once your picture is spotted out you have some choices to make. You probably have several versions of the picture, and you need to decide how many you are going to keep. Obviously our original KSinkX must be kept and archived, as must KSinkA1, our current work copy. But what about all those little variations in colour correction? You’re on your own.

If you have a CD writer you can save them all easily enough. These high bit files take up a lot of room, so it might be better to be selective. I do not keep any of the intermediate colour correction files once I have the colours right. I have a perhaps un-justified confidence in my skill, but I also have a small office and CDs take up space.

In a way there is no going back anyway, not without losing all the spotting you’ve done. Any colour changes from here must also include the spot layer or the spots will return.

I do Archive the A1 copy as a two layer file - the Background and Spots. This archive I re-name KsinkA1K . The last K stands for Keep.

Crop

Now is the time to decide how much of this image you want the world to see. First flatten the image. The extra layer is safely archived if you need it, and files work faster the fewer layers there are.

I keep a pair of cropping "Ls" on top of my monitor and use them to determine the crop from the screen image. I made these out of a pre-cut matte bought at the local hobby shop. At times I cannot decide on a crop using the screen, and end up printing the image full frame. Once I have a hard copy I can frame it with my "Ls" and step back to where I can get a better idea of how it will look. For full frame printing follow the Sizing instructions below

After cropping a file I always save it to a new name -- KSinkC1. I then bring up KSinkA1 and decide if I made the right move. If I think there is a better way I flatten the image, make my new crop and compare it to C1. I also save this file as KSinkC2. There can end up being a lot of files this way but keep them all on disk and only the relevant ones on the screen. When you are happy with your crop name it KSinkCK and save it. Go through the earlier versions and delete the bad, while keeping the good. All of these files are fully colour corrected and spotted out. The only thing you need do is re-size and they are print ready.

Preparing for Output

You will notice that up till now I have said nothing about changing either the size or resolution of your original. This was quite intentional. By staying at the original size and resolution, we have lost no information, and still have All our options for output.

You are now prepared to make sub-files for you specific purposes. If you want to make a group of 4x6s, and an 8x10, and put the picture on you website, all you need do is open KSinkCK, and set it to the specific size and resolution you need, then save it with an appropriate name. Ksink8x10, KSinkWeb1, Ksink16x20, etc. These files you can either keep or delete after they complete their task. You have only to go to KSinkCK to make a new copy that meets your exact needs.

Re-Sizing & Re-Sampling

This is one of the most confusing tools in Photoshop. The Re-Sizing dialog box actually contains two tools that, while they can be used together should probably always used separately. Attempting to re-size and re-sample an image together will only lead to a long session with the dreaded Hourglass, and the need to do it over anyway.

We start with re-sizing for most files. Uncheck the Re-sample boxes. Now it is simply a matter of setting one of the two dimensions to your preferred printing size - eg, setting the long side to 10 inches will allow you to print most pictures to an 8x10 (the actual printable size of most 8.5x11 sheets) paper. Click okay.

Now, with size set, re-open the re-sizing dialog box and put a check mark in both the Re-sample boxes. There will be a secondary box asking for re-sample type. Bucubic is the only one worth doing.

Set the dpi to the best you can get for your printer. Most printers cannot use more than 240 dpi currently - this will change in future - one of the reasons why you have archived KsinkCK at full resolution. You can actually get along on quite a bit less than 240 dpi if needed. If you find the file is well under 200 dpi for the size you will print you may want to Up-sample. Up sample is quite controversial. I do it. Some condemn it. I feel it will cause the printer to lay down more ink and give richer colour. It will not add detail that is not there, it will not cure "jaggies" caused by too low a resolution. It will only add a bit of ink to the print. Obviously, there is no other time I would even consider an Up-Sample to be advisable. Save this too under the name KSinkUpPrn or KSinkDownPrn. Make sure you have changed the name, so you don’t overwrite your CK version. Print and hang your new picture.

If you want a web version too, don’t work from your PRN version. Return to the CK copy. It will work faster. For this you do not need to set both size and sample. For web graphics all that really matters is the actual number of pixels in your file. The average screen your files will be seen on is still 480x640 pixels. Anything larger will not fit on a single screen. You could, if you think your main audience has higher resolution monitors, decide to shoot for 600x800 resolution, but any higher would be asking a lot of your potential viewers.

To set size for the web go to the very top of the re-size dialog box. The actual image size in pixels will be listed there. With the Re-Sample boxes checked, change one of the two pixel dimensions to your final screen size - eg 500 pixels will fit on a 640 wide pixel screen with comfortable margins. After you click okay the image will be ready to save.

Saving for the web is a little more complicated. You want to save in Jpeg format, which will take up the least amount of time when downloading. There is a joker in the deck though. If you save the image with a thumbnail, it will be almost twice as large as an image without a thumbnail, and may take several times as long to load.

If your program does not ask you if you want a thumbnail, you may have it set in the preferences to not ask. In Photoshop: go to Preferences/Saving Files and click on the box for Always Ask.

Go back to your file and save in jpeg format - medium quality (4-6) will do nicely. Just as a test save it again to a different name, With a thumbnail. Now compare the size of the two files. Usually when you go to a site that takes forever to load, it’s because the builder either re-sampled large files to 72 dpi (known as the Worst method of prepping images for the web) or he included a thumbnail (known as the Also Worst method).

Archiving

This method takes a bit of storage. There is really no alternative to the Cd for the moment although Writable DVDs are coming. Zip, Jazz and similar products either do not have the storage capacity you will need, or are too slow – or both! By writing all the files you are keeping to one (or more) CD Rom(s) you will have them together, and ready to serve another purpose later. You will also have most, if not all of them off your hard drive, giving you room to start another project.

As the sun sinks slowly in the west . . .

This concludes, at least for the time being my series on Digital Imaging. I am a photographer and don’t do X Files lettering or multi-picture compositing. I really only use a few of the Photoshop tools, but I use those constantly.

There are several books (see the Photographer’s Photoshop Library) and many websites that can help you with Graphic Design aspects of image editing but few for the photographer. I have attempted to fill the gap a bit. If there is a technique you would like to know, by all means write me, and I will do my best to either explain it or point you to someone who can. If I get enough requests for something that I know how to do, I will publish another article.

All of the articles are copyright 2000 by Tony Spadaro. You are free to print them out for personal use, and even make copies to give to others. Republishing, either in print or electronic media can be arranged with the author - ME - or my estate or surviving agents - whatever that means. Do not re-publish this work without obtaining that permission First.