Using reciprocal approximation instead of full 32-bit division in alg_quant()
[opus.git] / NEWS