diff libaf/af_resample_template.c @ 33736:a5a54c7a15ce

Fix the precision loss in float -> 32bit conversion case, introduced by my earlier commit. Instead use method proposed by Reimar. Also, avoid using ldexp, it is slower than multiply with constant.
author iive
date Wed, 06 Jul 2011 23:35:39 +0000
parents 0f1b5b68af32
children
line wrap: on
line diff