search

watson-developer-cloud / speech-to-text-websockets-ruby

Ruby client that interacts with the IBM Watson Speech to Text service through its WebSockets interface
Apache License 2.0
19 stars 13 forks source link

[Speech to Text] This ruby tool for STT does not get any messages from the server until all chunked files are uploaded. #9

Open alpha-netzilla opened 6 years ago

alpha-netzilla commented 6 years ago

I had felt that it takes a long time to get a first message of transcription even though this ruby client tool is implemented on event driven architecture.

I could see that the client did not receive any messages until it uploads all chunked audio files. This behavior is easy to understand as compared with the node sdk for STT WebSocket.

Can you think of anything that might have caused it?

daniel-bolanos commented 6 years ago

Do you mean that it takes a while to return the first interim results? Do you have interims enabled? Latency should be comparable to what you see here https://speech-to-text-demo.ng.bluemix.net/ when speaking live.

alpha-netzilla commented 6 years ago

It takes a while to return the first interim results even though interim_results parameter is true.

Latency is not comparable to the sample web app and Node.js SDK. This behavior is noticeable when using a large audio file.

It looks like that once the ruby client tool start processing "read_chunk" method, it will not receive any message until all chuncked files are uploaded.

Note: I run this latest STT ruby client on ruby 2.3.4.