Is there a settings or tip to improve the response time?

[jrichardsz]

First I want to thank you for your time to develop this client.

Your toolkit worked with buffer from sockets. I tried with another library without success

I just wonder if there is a setting or tip to decrease the response time. Currently is 4 or 5 seconds in localhost tests. I need a approximately response time of 1 or 2 seconds.

I tried to instance the recognizer (my local fork) before and the send it on each recognition but the time is the same :(

const recognizer = createRecognizer( model, {sampleRate, grammar, alternatives, words} )
const result1 = await transcriptFromBufferWithPreLoadRecognizer(data, recognizer)
const resultN = await transcriptFromBufferWithPreLoadRecognizer(data, recognizer)

Also according to my tests, the recognizer.acceptWaveform method is about 2 or 3 seconds.

Thanks