In ACELP decoder, the pitch values should be scaled before they enter the NB post-filter at https://forge.3gpp.org/rep/sa4/audio/ivas-basop/-/blob/main/lib_dec/acelp_core_dec_ivas_fx.c#L1433: ``` #ifdef FIX Copy_Scale_sig( pitch_buf_fx, pitch_buf_tmp, st->nb_subfr, -Q6 ); #else Copy( pitch_buf_fx, pitch_buf_tmp, st->nb_subfr ); #endif ``` In addition, the scaling of pitch values is used at two other places in `acelp_core_dec_ivas_fx()`. The scaling should be called using the parameter `st->nb_subfr` instead of macro `NB_SUBFR16k` (in NB, 4 subframes are generally used).