Kerning is more of an art than a science. While I would say the ’oc’ pair in the Metrics sample is too loose, there are many pairs in the Optical sample that are too tight, and the ’ry’ is too loose. Remember also that text fonts are kerned for text sizes (check the samples at 11pt) and Optical spacing is best left for headlines and used with care.

hmmm... interesting. in this sample, i still prefer the optical overall in this sample.

in the serif - the ’ry’ seems fine in both to me. and around the ’o’ seems better in the optical.

in the sans - your’re right about thr ’ry’, but check out the ’om’! the optical is heaps better! even is the ’sub-head’ the ’o’ seems to cause probs for metrics.

thanks stephen!

Not only is kerning more an art than a science, opinions differ, with few completely “right” or “wrong” answers. With really bad kerning, you might get 98 out of 100 people to agree it is bad.

As Stephen Coles pointed out, text fonts are kerned at text sizes. Moreover, foundry kerning has to assume any mix of letters. WA has to be kerned for a following “L” or a following “V”; only one value permitted in the font.

The rule at our shop (composition) is that display type must always be hand kerned. No font-based kerning can cover the all the sizes, the variations in letters used, the variations in the number of words in the title. These are all at play.

thanks charles. it’s clear i have a lot to learn!