# Bug description The initialization of `hItd->E_band_n_fx` and `hItd->E_band_n_exp` in `stereo_enc_itd_init_fx()` should be reviewed. In Float, the initialization is done as follows: ``` set_f( hItd->E_band_n, (float) ITD_VAD_E_BAND_N_INIT, STEREO_DFT_ITD_VAD_BAND_NUM ); ``` where `ITD_VAD_E_BAND_N_INIT` == 200000. In BASOP, the initialization is done as follows: ``` set32_fx( hItd->E_band_n_fx, ITD_VAD_E_BAND_N_INIT, STEREO_DFT_ITD_VAD_BAND_NUM set16_fx( hItd->E_band_n_exp, Q31, STEREO_DFT_ITD_VAD_BAND_NUM ); ``` where `ITD_VAD_E_BAND_N_INIT` == 200000, but the exponent is not matching. However, I have my doubts that this initialization is actually used at all - `E_band_n` is set to 0 in `stereo_dft_enc_itd_vad_fx()` --\> `itd_vad_ms_snr_calc_fx()` at the beginning of the loop.