by Todd L'Herrou
As you might guess, this is a common problem, and a common type of question. Unfortunately, small images or highly compressed images (including JPG format images in particular) provide very little to work with when trying to enlarge them. It simply isn't possible to display information which isn't there. The compression for JPG images also often introduces artifacts (squiggly lines, noise around edges, blurring, a smeared appearance, etc) which can be quite distracting when enlarged.
Whenever possible, it's best to go back to the original, but if the original is small to start with, or just isn't available, here are some approaches worth trying, and in some cases the results can be quite satisfactory.
For all of the following approaches, when you try to make your images larger, the software tries to guess, based on the information that is
there, what the missing information might be. There are a number of methods that can be used to make that "guess" about the missing information. Some of the steps you can take depend on the software you have. Most graphics software offer a range of different options for resizing images, with some options working better on some images or image types than others. What works best in general may not be best for your specific images.
In Photoshop, for example, there are some different settings for the software to use when re-sizing the images. If you look at the "Image Resize" box in Photoshop, you should see that there is a checkmark at the bottom of it, marked "Resample Image". Next to that is a pulldown box for the algorithm to be used to resample the image. You might try experimenting with those algorithms, to see what works best for you - bicubic is usually best for photos, but not necessarily for other images.
For the free software Irfanview
, you can choose, under the Resize / Resample menu item, to resize or resample the image. If you select "Resample", there are several alternatives which may be selected. I have shown below a series of images enlarged to 4 times the original size using Irfanview. The first of these is simply resized, creating a pixellated (or "jaggy") look. The other three were resampled using various options: Hermite, B-Spline, and Lanczos methods. As can be seen, the results differ - for example, the Lanczos method, while bringing out detail, also enhances the previously mentioned jpg artifacts - so ultimately you, the user, must decide which is best for YOUR image.
Another option that can be used is stepped enlargement. Some images seem to benefit from enlarging in steps rather than all at once. Try using about 5 steps of about 20% of the total enlargement you want, and see if that makes a difference. So, rather than going from 120pixels to 500pixels in a single step, you would go from 120 to 200, from 200 to 280, etc.
You can also slightly oversize the image by about 20%, and apply some light blurring, then resize back down to your final size. This usually works best when the size change is not extreme.
There is also specialty software for image resizing. This software is often used for commercial work when the original source file is already high-resolution, but must be enlarged further for large-scale printing. Be prepared to spend some $$$ if you go this route. The best known of these specialty programs is Genuine Fractals.
Finally, I have to point out AGAIN that all of these methods are extremely limited when working with small JPG originals. You are NOT likely to get results that are particularly satisfactory, but they may help. You may have better success with other formats (for example, the TIF and PNG formats are uncompressed, which can mean larger file sizes, but no data was lost when the file was originally saved).
Unfortunately for those trying to enlarge a reduced-size photo, there's no movie-style solution, where you can zoom in on some tiny detail, blow it up to the size of a computer monitor, and read the date on the newspaper to solve the mystery.