Jan Buethe authored 2 years ago and Jean-Marc Valin committed 2 years ago

Model now stores LPC gamma, look-ahead, and end-to-end.
Parameters aren't quite reliable yet, YMMV

Helps forget concealment features faster

Variable cannot be uninitialized

Bias the energy for vowels, but not for consonants

Causal vs non-causal, DC filter, and blending are now configurable

Makes the offset signal more continuous

Defer some of the extra resync work to the next packet

Remove DC, conceal, add it back at the end

Instead of t from t