By Gerd Waloszek, User Productivity, SAP AG – 02/22/2005, updated 01/30/2007
Screenshots and the redesign of screens are the bread-and-butter business of many user interface designers. Nevertheless, I often find that there are "knowledge gaps" that need to be bridged. Therefore, in a series of two articles, I want to help you avoid the greatest obstacles in dealing with screenshots:
Please note that this short overview is by no means exhaustive – it just scratches the surface...
User interface designers often use screenshots in their daily work, for example, in:
They show what actual screens look like – or should look like, as is the case with specifications and guidelines.
Screenshots are also used for:
Redesigns and prototypes are a useful instrument for discussing designs with colleagues or decision makers. Ideally in this situation, you can create design alternatives from which a decision can be made (or a compromise found).
The basic procedure for creating and using a screenshot is to capture it and publish it in a document or an application without modification. In its simplest form, this is just a matter of:
See Where Do I Get Screenshots from? for more information and some caveats...
More advanced processing of screenshots might include:
I will address these topics in Part II of this series.
Most screenshots originate from one of the following three sources:
Windows offers two simple ways of capturing a window or screen:
Dedicated screenshot applications offer far more options, such as showing or hiding the mouse cursor, capturing open menus, creating sections, saving the capture in a certain image file format, and many more.
Update: Joshua Vaughn drew my attention to Size-O-Matic: "It's a great little tool for making sure screenshots you capture are representative of a specific resolution when you're using a larger monitor." The tool offers a set of fixed resolutions from 640*480 pixels up to 1600*1200 pixels. You can download Size-O-Matic for free from www.pythoness.com. (The company whould, however, appreciate a donation of $3 to $9.)
Most screenshots that I receive from colleagues are embedded in MS Office documents, such as Word and PowerPoint documents. In this case, however, you can run into problems:
Note: In PowerPoint, you can select a screenshot or any other bitmap graphic and save it as an image file (see below for image file formats). This is not possible in Word.
Note to Office 2003 users: Regrettably, Office 2003 no longer includes the MS Photo Editor application, which is used in the procedure below as an intermediate stage for retrieving images from Office documents (Word, PowerPoint). Read more below.
Obviously, the Windows clipboard preserves unwanted meta information about the screenshot. My "trick" (with thanks to Kai Willemborg!) is to use the Microsoft Photo Editor as an "interim storage" to get rid of this information. Here are the necessary steps:
Other programs might do the trick as well; for example, you can also use an older version of ThumbsPlus for this operation (read more).
Note Continued: Currently, I cannot offer a universal solution to this problem. I returned to an old strategy of mine, namely to paste images into an older version of ThumbsPlus (I use version 3.2) and retrieve the bitmap (Image -> Extract Bitmap). This works in nearly 100% of the cases. Read more about this procedure in an older article on mine.
Depending on your needs, you can:
When capturing screenshots, keep the following recommendations in mind:
In general, you should remove unnecessary elements from the screenshot, for example, by cropping it. This helps people to focus on the relevant features of the screen, instead of being distracted by irrelevant elements.
Without going into the details of image file formats too deeply, I would like to discourage the use of the JPEG format for screenshots for the following reasons (the same applies to the PNG-24 version that uses a lossy compression):
The drawbacks of the JPEG format lead me to make the following recommendations (which also apply to PNG-24 images that are saved with a lossy compression):
|
You Can Use the JPEG Format...
|
Do Not Use the JPEG Format...
|
|
|---|---|---|
|
... if the only purpose of an image is to provide a visual impression of the screen or screen prototype. In this case, you can use it in:
Be careful not to use a compression rate that is too high, because this may severely impair the visual impression of the screenshots. |
... if it is known or expected that the screenshot will be edited later for:
|
In some cases, artifact may be barely noticeable. In others they can be very disturbing, particularly if an image or parts of it are to be reused for designing new screen prototypes.
These are the most common Web file formats that can be used for screenshots:
In most cases, I use the GIF file format for screenshots; PNG-8 leads to slightly smaller files, though. The screens themselves rarely require more than 256 colors. Only window titles or banners usually introduce more colors. Dithering is acceptable here. There is a problem with the color palettes, though: In Adobe Photoshop it is hard to find out, which of the color palettes (perceptive, adaptive, selective) does not introduce subtle hue shifts and preserves the exact colors. (I opt for "selective.")
Comment from Martin Rieger: Based on my experience it is therefore the best to use PNG-24 for lossless compression of images (e.g. screenshots) or saving screenshots uncompressed (Windows Bitmap).
(Thanks, Martin for some corrections regarding the PNG format!)
The following example demonstrates why JPEG is an unsuitable file format for screenshots:
![]() |
![]() |
|
| File size for GIF: 1.3 KB | File size for JPEG: 4.7 KB |
Figure 1: Comparison of screenshot saved in GIF and JPEG
GIF image vs. JPEG image (30% compression rate). Note that even though the JPEG compression rate is fairly high in order to make compression artifacts more visible, the lossy JPEG compression leads to a larger file size than the lossless LZW compression in the GIF image.

Figure 2: Enlarged JPEG screenshot
The JPEG image was enlarged to demonstrate the compression artifacts. Most noticeable are the artifacts around text and at edges. But even the color values of uniform areas are not guaranteed to preserve the original values. (The compression can lead to subtle color shifts.)
See also the section of the screen partially filled with yellow color to demonstrate the hue artifacts (tolerance 1 – tolerance 4).
The following examples demonstrate artifacts around edges and text, hue shifts, and the creation of "block areas" in JPEG screenshots.
|
Preview (Click for Large Version)
|
Enlarged Section
|
|---|---|
|
See also section of the screen partially filled with yellow color to demonstrate hue artifacts (tolerance 1 – tolerance 4) |
Section showing artifacts around text and block artifacts and hue shifts in uniformly colored areas |
|
Screen saved in JPEG format with artifacts (See also enlarged section to the right) See also screen partially filled with yellow color to demonstrate hue artifacts |
Enlarged section showing artifacts |
|
Screen section saved in the JPEG format to demonstrate block areas
Section partially filled with yellow color to demonstrate hue artifacts |
Enlarged section showing color blocks
Ditto with yellow color |
This concludes Part I of my article series on screenshots. In Part II, I will briefly address simple editing tasks, comment on how to patch screens together and how to work with text, and finally provide some hints on batch processing and Web galleries.