# Re: track kerning

```At 19.26 +0200 0-06-06, Walter Schmidt wrote:
>Hi,
>
>apologizing that my question is a bit off-topic -- it is not
>special for TeX.
>
>I found the following piece of information in an afm file:
>
>StartTrackKern 3
>TrackKern -1 6 0.10 144 -2.09
>TrackKern -2 6 0.05 144 -4.02
>TrackKern -3 6 0.00 144 -5.96
>
>I know that this means, that letterspacing should vary with
>font size.  However, what does this mean _quantitatively_?
>How do I have to read this information?

On p. 27-28, Adobe technical note 5004 (AFM specification) says:

Normally track kerning is provided in different degrees of tightness.
Within a
track (a degree of tightness), the amount to decrease (or possibly increase)
the amount of space between characters increases (or possibly decreases)
with the point size of the font (for example, for tight track kerning, the
amount to decrease the space between characters at 6 point might be 0.1
points and at 72 point it might be 3.78 points). These distances are measured
along the width of the characters: parallel to the x-axis for horizontal
writing directions and parallel to the y-axis for vertical writing
directions.

The data itself begins with the keyword TrackKern and is followed by the
track kerning information

TrackKern degree min-ptsize min-kern max-ptsize max-kern

The degree is an integer where increasingly negative degrees represent
tighter
track kerning and increasingly positive degrees represent looser track
kerning. min-ptsize, min-kern, max-ptsize, and max-kern are all numbers.
Since
the track kerning is a linear function, the minimum and maximum cut-off
values
(point sizes) are provided, along with the amount to track kern by, at the
point size.

The kerning amounts are given relative to the point size. From those four
values, the track kerning function can be derived. The track kerning
function
is a linear function. The equation for the line is determined from the data
provided and, therefore, the track kerning values for any point size can be
determined. The track kerning values for any point size below/above the
minimum/maximum point size are constant (the minimum kerning amount/maximum
kerning amount).

I must say it looks a bit confusing, though, since what follows appears to
claim that min-kern and max-kern are in absolute points. They do give
formulae for how the track kerning varies with at-size, but the asymptotic
behaviour of those formulae gets absurd. (Maybe a case of programmer not
thinking too hard of what the formula implemented really means.)

Lars Hellström

```

• References:
• track kerning
• From: "Walter Schmidt" <walter.schmidt@arcormail.de>