Please help us make this page better by providing your feedback - click on the little "edit" link next to each heading to do so. You can also provide feedback on the page as a whole.

This page describes some of the lesser-known techniques you can use to get the most out of Vector Magic.

See also: How is Vector Magic different?, how to use Vector Magic, comparisons, and samples

We receive a fair number of scanned images - a lot of them are artwork that has been printed and need to be re-vectorized. There are several characteristic traits for these images:

• Shape edges are often blurry, with the edge transition happening over several pixels.
• 

  Color separation from scans. Limit the number of colors to get the most out of the scan.
  The color separation in the printing process leads to colors that do not perfectly overlap. This often causes the pixels around the edges to have "false colors". For example, the lower side of an edge may be a little more cyan, while the upper side is a little more yellow (many combinations are possible).

The key to successful results with scanned artwork is to limit the colors Vector Magic uses by picking the few colors option in the wizard and to use the right resolution when scanning. You normally want to use the logo with blending option, as the scanner almost always produces a result with blended edges. The main exception is if you've forced the scanner to just black-and-white, which we don't recommend. Grayscale is usually a better option if you want to get rid of the colors.

Scans usually have some salt-and-pepper noise in addition to being blurry, and the colors can vary quite a bit. You therefore want to use the option of limiting the colors used ("few colors" option), to minimize the impact of this. It also helps limit the effect of any remaining blurriness.

You want the resolution used when scanning to be such that the edge transitions happen over roughly one pixel. This makes for a sharp image that normally has all the essential information present in the original. If you don't know where to start, try scanning at 150 DPI and go up or down from there depending on how wide your edges are.

You can also scan at a higher resolution (e.g. 300 DPI) and then scale down the image in a bitmap editor. Be sure to use cubic interpolation when doing so to preserve the integrity of the image.

Using these simple techniques usually yields nice results from scanned artwork.

A lot of artists prefer to use traditional non-digital tools for their art creation, and scan-and-vectorize is one way to digitize this artwork.

Drawings are challenging from a vectorization perspective because they usually have very varying colors and the shapes are usually not fully connected, but instead separated by white between the strokes. They are then scanned, with the scanner adding its own artifacts to the end bitmap result.

When vectorizing drawings, it's important to be aware that the output can only be as good as the input. If you want something to be a line, but draw that as two lines with a clearly visible gap in between, then chances are it will come out as two lines.

That said, there are several simple ways in which you can improve the vector result:

• When vectorizing a scanned pencil or charcoal sketch on white paper, use the logo with blending option and limit the colors to white and a weak gray.

  

  Limiting the colors to gray and white can help preserve detail

  You're probably looking to get the sketch in black-and-white, but telling Vector Magic to use a weak gray allows it to interpret the more faint lines as lines and not background, thereby recovering more detail. You can then easily change the color in your preferred vector editor.

• Use a lower quality setting. This allows Vector Magic to smooth out the edges more, and reduces the amount of noise in the output due to unintentionally varying colors in the drawing.
• Filling in the sketch with e.g. ink before scanning it makes the colors much more solid and uniform, which makes for easier vectorization. Be careful with your original art though, and only do this if you're comfortable with it!
• Manipulating the bitmap image after scanning it is usually not as effective as limiting the colors used to white and weak gray, and is time probably better spent on the vector result.

Different levels of blurriness. Notice how the edge transition happens over more than one pixel in the blurry cases. Ideally you want your input images to be as sharp as possible

A lot of original bitmap art is quite blurry when inspected in detail. Basically, an image that has edge transitions wider than one pixel is effectively blurry. Almost all JPEG images get blurry as a consequence of how the image compression algorithm works, which is part of why we generally recommend PNG for your bitmap artwork.

There are two primary defenses against blurriness having a negative impact on your vector results:

• Limit the number of colors. The gray in between a white and a black region is not intended to be a shape in and of itself, but with blurry images this mistake can sometimes happen. If you limit the colors used to black and white, then these gray shapes will be removed.

• For very blurry images (typically happens with scans, or with extremely poorly JPEG-ed images) you can shrink the image a little (using cubic interpolation to preserve the image's integrity). This effectively makes it sharper and can give better vectorization results.

  This should only be done when the details you want to recover from the original are not so small as to disappear when the image is shrunk.

Turn a photo into stylized art! Process a photo using the logo with blending option, and use just two or three colors. This produces a cool image that you can use as graphic art such as the background of a poster.

Check this tool out! In the lower right corner of the troubleshooting guide is perhaps the most powerful editing feature ever included in a tracing tool: the Segment Editor. Segments are the coarsely divided regions of the image that are then smoothed together to produce the shapes in the completed vector.

The Segment Editor makes it very easy to edit a few pixels in the segmentation (repairing broken lines, removing pieces of noise, changing a color). Using it is much quicker than changing them in the completed vector.