IBM Watson Speech to Text Only Returning First Word With Java SDK

Question

I'm using the IBM Watson speech to text Java SDK, and when I upload the .wav file the response JSON only contains the first transcribed word. When I upload the same file to the web demo, I get the full response.

Very simple implementation with the SDK:

SpeechToText service = new SpeechToText();
service.setUsernameAndPassword("<username>", "<password>");

File audio = new File("src/test/resources/sample1.wav");

SpeechResults transcript = service.recognize(audio, HttpMediaType.AUDIO_WAV);
System.out.println(transcript);

Answer 1

The recognize() signature you are using will return after the first pause. To see all of the results do this:

RecognizeOptions options = new RecognizeOptions();
options = options.continuous(true)
          .contentType(HttpMediaType.AUDIO_WAV)
          .interimResults(false)
          .inactivityTimeout(10)
          .maxAlternatives(1)
          .wordConfidence(false)
          .timestamps(true)
          .model("en-US_BroadbandModel");
SpeechResults transcript = service.recognize(audio, options);

This works for me using the following maven dependency:

<dependency>
  <groupId>com.ibm.watson.developer_cloud</groupId>
  <artifactId>java-sdk</artifactId>
  <version>2.8.0</version>
</dependency>