eden-emu/opus
18
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fixed-point: MULT16_32_Q15 for TI DSP (not entirely happy with it)

Browse source
This commit is contained in:
Jean-Marc Valin 2008-03-21 10:53:09 +11:00
parent f5b0587bd2
commit 948dabc7da
1 changed files with 6 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

6
libcelt/fixed_c5x.h
View file

@ -59,6 +59,12 @@
#undef MULT_16_16
#define MULT_16_16(a,b) _lmpy(a,b)
/* FIXME: This is technically incorrect and is bound to cause problems. Is there any cleaner solution? */
#undef MULT16_32_Q15
#define MULT16_32_Q15(a,b) ADD32(SHL(MULT16_16((a),SHR((b),16)),1), SHR(MULT16_16SU((a),(b)),15))
#define celt_ilog2(x) (30 - _lnorm(x))
#define OVERRIDE_CELT_ILOG2