In this thread http://typophile.com/node/121385 OP put a scanned page from a 1961 book. Used types have somewhat twitched appearance. As I made earlier an OT feature which does that effect I modified a font (in that case Plantin) to recreate that look.
1. FEA file:
@class_alt1 contains UC and lc letters, space, comma, etc
@class_alt2 contains as above, slightly displaced, names with .alt2
etc up to @class_alt10 ( .alt10)
2. I made new glyphs - and copied the references of original glyphs and add the .alt1 to their names
3. Again the same with .alt2 names
etc up to .alt10
4. I moved glyphs of every class slightly up or down
feature calt { # contextual Alternates
sub @class_alt1 @class_alt1' by @class_alt2;
sub @class_alt2 @class_alt1' by @class_alt3;
sub @class_alt3 @class_alt1' by @class_alt4;
sub @class_alt4 @class_alt1' by @class_alt5;
sub @class_alt5 @class_alt1' by @class_alt6;
sub @class_alt6 @class_alt1' by @class_alt7;
sub @class_alt7 @class_alt1' by @class_alt8;
sub @class_alt8 @class_alt1' by @class_alt9;
sub @class_alt9 @class_alt1' by @class_alt10;
sub @class_alt10 @class_alt1' by @class_alt1;
} calt;
As feature calt is applied, letters (and spaces, commas etc) are substituted and in effect their sequence is:
x1, x2, x3, (etc), x10, x1, x2, x3 ...
so every next letter belongs to the group moved in slightly different way.
Because font is slightly thinner (and more contrasty) than lead types used I added some black stroke and - to compensate inter letter space - some tracking.
Advantage: -calt works almost everywhere
Drawback:
- in InD tracking about 40 switches off calt (why?!)
- when you repeat lines the letters are moved in the same way (doesn't happen frequently in books)
