Signed-off-by: Mans Rullgard <mans@mansr.com>

Shaves at least 3KB off code size on x86, should improve cache utilization.
This would probably be useful to do for other decoders/encoders as well.

prob[0] is the only prob array ever accessed, so prob[1] can serve as padding
for prob[0].

(cherry picked from commit bcf4568f)