Hi All, We've noticed a high load on our transcoding servers when doing OPUS<->PCMA transcoding. All our client software is using opus@48khz, but inside the audio payload we have narrowband audio (forced with OPUS_SET_BANDWIDTH(OPUS_BANDWIDTH_NARROWBAND) ) . Questions : 1. How can we avoid resampling ? We've noticed this log line showing twice for each call: 29128b92-d842-11e4-877f-9bd23fac7445 2015-04-01 07:39:14.798759 [NOTICE] switch_core_io.c:1287 Activating write resampler If we modify the client software to use opus@8khz and leave freeswitch as it we have very bad audio (nothing intelligible). But If we modify the client software to use opus@8khz and change mod_opus.c in freeswitch so it would use only OPUS@8khz then we have proper audio. 2. We have noticed that there is a "samplerate" param in fmtp (which is not in this RFC: https://tools.ietf.org/html/draft-spittka-payload-rtp-opus-03 ) it's used only for configuring the decoder. Why only for the decoder ? If "samplerate" is set then it's passed to opus_encoder_create() , but not to opus_decoder_create(). 3. How about using sprop-maxcapturerate to set the sample rate of the decoder ? Regards, Dragos Oancea