# Basic info <!--- Add commit SHA used to reproduce--> - Float reference: - Encoder (float): db364042fba738d5d0f9823c047e75edea62bae7 - Decoder (float): db364042fba738d5d0f9823c047e75edea62bae7 - Fixed point: - Encoder (fixed): f14583620e84f03d718223da0ea4be6cc7c77c61 - Decoder (fixed): f14583620e84f03d718223da0ea4be6cc7c77c61 # Bug description (note, initial description was updated to reflect correct current status) Looking at LTV signal MASA to binaural processing, there is an outlier case with rate switching that shows some extra spikes in the paths using BASOP decoder. The input can be either from float encoder or BASOP encoder. See the figure below for BASOP spikes. This is the float enc to BASOP dec combo. As this is rate switching, it is obvious that some switching artifacts are expected but this seems a bit worse and also is probably causing higher MLD of 16.5.  # Ways to reproduce <!--Commandline or script--> The command line below and rate switching pattern here [generic_start_13k2_20frames.bin](/uploads/fd4ce084d58c5f6f7e6cecec98f7f2a8/generic_start_13k2_20frames.bin) ```bash IVAS_cod -masa 2 ltv48_MASA2TC.met -max_band wb -dtx generic_start_13k2_20frames.bin 32 ltv32_MASA2TC.wav test_0168.bs IVAS_dec BINAURAL 48 test_0168.bs 0168_ltv32_MASA2TC.wav_generic_start_13k2_20frames_32-48___BINAURAL.wav ``` <!--- Below are labels that will be added but are not shown in description. This is a template to help fill them. Add further information to the first row and remove and add labels as necessary.-->