# Bug description In lib_dec/ivas_omasa_dec_fx.c, there are a few floating-point leftovers in the definition of the constants: ``` #define EXT_RENDER_IIR_FAC 0.95f #define MULT_17_DIV_20_Q15 ( Word16 )( ( 17.0 / 20.0f ) * pow( 2, 15 ) + 0.5f ) #define MULT_1_DIV_CLDFB_NO_COL_MAX_Q15 ( Word16 )( ( 1.0 / CLDFB_NO_COL_MAX ) * pow( 2, 15 ) + 0.5f ) #define ONEMINUX_EXT_RENDER_IIR_FAC_Q31 ( Word32 )( ( 1.0f - EXT_RENDER_IIR_FAC ) * pow( 2, 31 ) + 0.5f ) #define EXT_RENDER_IIR_FAC_Q31 ( Word32 )( EXT_RENDER_IIR_FAC * pow( 2, 31 ) + 0.5f ) ``` This floating-point code is used during runtime, and it shouldn't be too difficult to convert that into fixed-point constants.