Many of us have hundreds if not thousands of snaps taken on a variety of cameras from the one-use throwaway jobs to full blown professional SLR's. Some of which we may wish to publish on the web. Converting images from analogue transparencies or negatives can be tricky. This short tutorial aims to help you understand what to do. And perhaps as importantly why.
As an example I shall use the picture of myself lying on a Cornish beach, tin of beer in hand and with my dog asleep on my shoulder. This is in fact the photo at the head of the home page of this site.
First off, what equipment should you use? Well if you have a lot of images as slides (aka transparencies or dia-positives) or negatives then investing in a dedicated slide or film scanner makes a lot of sense. Flatbed scanners are really for documents, prints and such like. It is true that you can get film adaptors for such scanners but often these are more trouble than they are worth. You are doing a specialised job so get a specialised tool. Such scanners range in price from UKP140 to UKP500 and upwards. Obviously you get what you pay for and often the more expensive models permit you to load a batch of slides or negatives and scan them all in one operation rather than having to handle each item individually. For home sites though the cheaper end of the market is more than adequate.
Other tools needed are a slide viewer or a light box to select appropriate slides for scanning. Most non professionals would be hard pressed to determine the quality of a negative without looking at the print so for these scanning may be the better option. You may also want to acauire a CD writer as scanning at the sort of resolutions and scalings presented here will fill your hard drive rather quickly! Depending on the settings disk, usage can be as few as 100 images per Gigabyte. If you do not fancy opening up your computer to fit such a device then a USB model will fit the bill, prices start at around UKP50 for the internal models and UKP120 for USB models.
As an aside, the camera used to capture the image in the first place is also of importance. I have two: a Canon T90 SLR and a compact by Pentax. Both take good crisp images but those from the Canon are much better. And so they should, after all it cost almost 6 times as much! You get what you pay for.
On the software side you really need two programs, the first should come with the scanner. This is the software written by the manufacturer to drive the scanner, this may have some editing facilities but it is better to use the second program for this. You will also need a photo manipulation package. There are several available on the market, most are bundled with some other package so that you may not need to lay out any cash whatsoever. If you are on Linux then you have Gimp, otherwise on Windows there are Adobe PhotoShop, Ability PhotoPaint amongst many others.
Now to actual grabbing an image. The first thing to ensure is that the negative or slide is clean. It is much easier to remove specks of dirt than to touch up the image later in the process. In fact it is a general fact that the earlier you can fix a mistake the cheaper both in time and money it is to fix it. If the source is dirty then a simple soft brush as sold in a photographic shop should be enough for most dust etc. If you want to go the whole hog then there are also aerosol based brushes. These used compressed dried air to blow dust and dirt away. This avoids the possibility of scratching the surface during brushing.
Place the source in the scanner in the manner recommended by the manufacturer. (This assumes that you have the device turned on and connected to the computer of course!).
The scanning process may be initiated in two ways: the first way involves starting up the scanner driver program and using that to begin the process. However most image handling programs have an "acquire" menu option. This calls the scanner driver program for you. This way has the advantage that the digital image is placed directly in your image manipulation program rather than you having to save it then swap between programs.
The scanner driver will allow you to determine how you perform the scan. Since most computer screens have a finite resolution of either 72 DPI (Dots Per Inch) or 80 DPI it makes little sense to scan at a higher resolution than this. Other factors in the resolution to choose include the scanning time and the resultant file size. Since you are scanning an area, doubling the resolution means a file size that is four times as big! The file size will be big enough anyway so settle for around 150DPI. The second setting is scaling. Since you are scanning a small image, you want the result to be as large as possible. Also it is easier to shrink an image and loose information than to try and enlarge one and make the existing information stretch. So set the scaling to 4x or 5x or 500 - 600% depending on the package you are using. Finally if you are scanning a negative, choose the film type: most scanner drivers have filters that are optimised for particular films.
Now you can scan your image. Depending on the scanner this may take up to a minute. On completion the image will be placed in your image editor ready for you to work on it. Save the resultant image so that you have something to work with. The size of this image could be big. The file size of my photo is 1611Kb. If I increase the resolution to 500% the the resulting file may be as much as 9Mb! The best format to save the image as is probably TIFF.
The first thing to check are the balance levels. Most photographs have some colour tint and you now have to get rid of this as best you can. There may be an option for "auto levels", try this and see if it results in a pleasant image, if it does then confirm the changes and move on. If it doesn't then look for a menu option named along the lines of "Levels", this allows you to manipulate the colours either as a whole (luminosity) or via red, green or blue elements. Typically you will be presented with a profile of the colour and three associated handles, one at the bottom of the scale, one in the middle and one at the top. By moving these with the mouse you alter the colour balance of the image. The main image will update in real time so you can see what your changes will do. If you make a mess then click the cancel button and the image will be returned to its original state.
The next step appears somewhat counter-intuitive, namely "unsharpen mask". This refers to a photographic technique whereby a non-sharp mask is created from the original image and placed over the original image. By some magic that I don't understand this actually sharpens up the original image! So find the option in your software and apply it.
Artefacts
The next thing to do is to get rid of artefacts. These are blemishes caused by dust and dirt on the original image. If you had cleaned the source properly you wouldn't have to do this bit.... For this operation zoom in on the image using the appropriate menu option. There is one gotcha to look out for when doing this, if there is dirt on the screen then you may find yourself attempting to clean up a speck of dirt that is not where you think! Not that I speak from personal experience of course.
Locate the blemish. Now out of the range of tools choose the "color picker", this is normally indicated by a pipette symbol. Click on an area of the image as close to the blemish as possible, this chooses your colour for the next step. Now choose the paintbrush tool and click on the blemish. You may have the option to choose the size of the spot of "paint" that you apply, if so choose a small size. If the blemish is large you should work in from the sides choosing a local colour for each side. Repeat this task for every blemish on the photo.
Remember to save regularly, in the same format as the original.
The next thing to do is to crop the image to remove areas at the edges where the scanner went past the edge of the image or similar. Basically you are selecting the best bit of the raw image. Use the select tool and drag out a rectangle of interest. Copy this to a new file. If you get it wrong, undo or throw the new image away and try again.
Nearly there! The image that you have created is of some indeterminate size, you need it in a particular size to fit on your web page. There are two things to consider here: a thumbnail is a very small version of the image that is usually used as a link that your users can click on to see a bigger version. You need to save the image twice, once in its full size and once as a thumbnail.
Save the image in a format suitable for the web, usually "jpg" or "giff". As an example choose "JPG" from the list of types in the "save as" menu. Ensure that the resolution of the image is 72DPI. Using your computers file display mechanism find the new file and check the size, in my instance it is now 75Kb which is a massive decrease in size! This reduction comes at a cost, as an experiment open the JPG file and save it as a TIFF and look at it. It will be of much poorer quality than the original TIFF. This is because JPEG is a lossy compression scheme, once you have saved it in this format you lose information never to retrieve it. Now is the time to note the size of the image in width and height terms. This is usually measured in pixels. You need this information for the width and height attributes when placing the image in a page.
Go back to the last image saved as a TIFF. NOT the JPEG you have just saved. Choose image size and choose some suitable size for one of the dimensions. If you have the option enabled to keep proportions the other dimension will resize automatically after you enter your dimension. Again ensure that the resolution is 72DPI as any more is just being wasted on a computer screen. Save as JPEG again, using the same name as before but add "_t" just before the dot to indicate that this is a small or thumbnail image, i.e. "home.jpg" and "home_t.jpg". In my instance I end up with a 200 pixel by 138 pixel image a mere 7Kb in size!
The above method, while appearing convoluted, attempts to make the greatest changes to the image whilst you have most information at hand. By the time you save the image for use and lose a large amount of unneccessary detail you have made your changes. If you keep the original scan image after tidying up then you can play around with it later.
All the above will take between 5 and 20 minutes depending on how much work is needed to be done on the image.