On Wed, 2013-02-13 at 17:26 +0100, Peter Meerwald wrote: > From: Peter Meerwald <p.meerwald at bct-electronic.com> > > at least valgrind complains about it > > Signed-off-by: Peter Meerwald <p.meerwald at bct-electronic.com> > --- > src/modules/echo-cancel/adrian-aec.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/src/modules/echo-cancel/adrian-aec.c b/src/modules/echo-cancel/adrian-aec.c > index 05a5d51..fcaf1d2 100644 > --- a/src/modules/echo-cancel/adrian-aec.c > +++ b/src/modules/echo-cancel/adrian-aec.c > @@ -86,6 +86,7 @@ AEC* AEC_init(int RATE, int have_vector) > a->cutoff = FIR_HP_300Hz_init(); > a->acMic = IIR_HP_init(); > a->acSpk = IIR_HP_init(); > + a->dotp_xf_xf = 0.0; > > a->aes_y2 = M0dB; I'd rather use pa_xnew0() when allocating the AEC struct. -- Tanu