Bruce McGuffin wrote:
Based on the coefficients, we can guarantee that the final sum will
have a magnitude which fits in 20 bits, including the sign bit. We
can just drop those extra sign bits, since they'll all be the same.
Matt
So I'm curious. If the final sum fit into 20 bits, why were the
accumulation and
following sum carried out with 24 or 25 bits?
Bruce
If we changed the coefficients we might need more bits.
Matt
_______________________________________________
Discuss-gnuradio mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
