FAQ

Where does AEC fit in the voice processing pipeline?

What security recommendations does Krisp offer for utilizing its libraries/models, and does it provide an embedded authentication mechanism?

What is the recommended way to handle stereo audio input in Krisp SDK?

Does Krisp SDK establish a network connection or send data over the network?

• Can I load new models from a CDN in JS SDK?
• Is there public Krisp CDN available for model deployment?

What are the factors and conditions that can adversely affect the quality?

What modes does Krisp offer to configure Noise Cancellation aggressiveness depending on user preference/environment such as Low, Medium and High noise scenarios?

When choosing a Krisp model (wide-band, narrow-band, or full-band), what factors should be considered in relation to the audio sample rate and bandwidth used?

What is the suggested approach for utilizing WebRTC constraints (AGC, AEC, NC) in conjunction with Krisp Noise Cancellation (Krisp NC)?