I have a Python script which uses an audio file and Watson speech to text service, and prints the recognized transcript and also the confidence. Is there a value for "latency" calculated by Watson service that I can print out?
I already have a Python script to calculate the latency myself, just wanted to know if Watson provides this number so I can compare it with the latency from other speech to text services.
Thanks!
The latency depends on the network access you have and your geographic location (how far/well connected are you from the datacenters).
An easy way to measure end-to-end latency is via timestamps on the client's end. You get a timestamp right before making the request and then another one right after getting the first interim result or first final result, depending on how you define latency. Then you calculate latency as the difference between the two timestamps.
