Delay in audio - any ideas? #18
-
Hi there! When using the Audio sample, I experience a 4-5 second delay before I receive an answer. I have 500mb+ down and up speeds so that shouldn't be the issue. I left all settings on default, if maybe you can help me what I could tweak? Also what is a normal delay that I should be experiencing? I want to create NPC-s and would be great if I could get closer to real time responses. My idea was maybe to use the same event that is sending the audio to the engine to play an audio file from an array of "time filler" audio, that can indicate the NPC "listening", maybe nodding or other animation as well. Please do share your thoughts! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
Hi @WolfhunterHUN, the delay usually is a couple of seconds and it might increase if the input audio and as well as the response is long. I will check it out on my side and see if there is anything I can improve. I did not bring real time audio support just yet since it is quite expensive on my side. Currently I have a simple "Thinking Indicator" implemented which is also available as a UI Element, extra animations etc can be introduced on characters. Such as you can play that nodding animation, when OnTranscriptionReceived (players voice turned into text and is displayed) event is received. Giving users a list of one liners in the selected voice also makes sense, i can add that feature to the SDK! |
Beta Was this translation helpful? Give feedback.
Hi @WolfhunterHUN, the delay usually is a couple of seconds and it might increase if the input audio and as well as the response is long. I will check it out on my side and see if there is anything I can improve.
I did not bring real time audio support just yet since it is quite expensive on my side.
Currently I have a simple "Thinking Indicator" implemented which is also available as a UI Element, extra animations etc can be introduced on characters. Such as you can play that nodding animation, when OnTranscriptionReceived (players voice turned into text and is displayed) event is received.
Giving users a list of one liners in the selected voice also makes sense, i can add that feature t…