17.0.0 nVoq Dictation Release Notes

17.0.0 nVoq Dictation & Dictation API Release
December 1, 2021

nVoq Dictation

• Dictation with Long Pauses Not Fully Transcribed on ReDS - Dictations with very long pauses between words are not always fully transcribed by the new dictation server (the ReDS). A silence weight has been added to help resolve this issue. As a reminder, we do not recommend dictating with long pauses between words; rather dictations should ideally be spoken in the same way as you would normally speak. (D-12782, 11/8/2021)
• "Stop Dictation" Command Appears in Transcripts from ReDS - The "Stop Dictation" command for SayIt Hands Free mode is appearing at the end of Direct Text (a.k.a. STABLETEXT) dictations processed by the ReDS. It is working fine for non-Direct Text (HYPOTHESISTEXT). In 17.0.0 the Direct Text (STABLETEXT) code was modified to remove the “stop dictation” command. (D-12828, 11/10/2021)
• nVoq.Mobile Voice Transcript Not Finalized/Replaced by Next Dictation when Alternate Words Include Extra Space - With the 16.2 version of the ReDS dictation server, Alternate Words in the nVoq.Mobile Voice keyboard could sometimes include an extra space before the word. This could cause the transcript to never finalize. When the transcript was not finalized, the user would see an elipsis (...) remain at the end of it. Because it was not finalized, the transcript would be removed and replaced with a new one when the next dictation was started. This has been fixed with the 17.0.0 version of the ReDS dictation server. (D-12840, 11/18/2021)
• Logging Added to Diagnose Combined Words Forming Gibberish Text - It was reported that some users were occasionally seeing some words in their dictation combined into gibberish words in their dictated transcripts. Since we are unable to reproduce this issue internally, logging has been added to the ReDS in hopes of learning more about what may be causing this. (D-12612, 9/30/2021)  

nVoq Dictation API

• New HEARTBEAT Message for WebSocket Dictation - It was reported that some users were occasionally receiving no transcript after recording a short dictation. These dictations would display in Review & Correct with a "No audio data available" error message. In all instances these dictations were less than five seconds long and received a STARTDICTATION response from the server, but the dictation client received no transcript and closed after ten seconds. As part of our work in trying to resolve this issue, a new HEARTBEATmethod has been added to the WebSocket dictation API. The client may optionally send this message to the server to find out if it is still connected to the server or to obtain round-trip time metrics. (D-12751, 10/14/2021)
  • https://test.nvoq.com/apidoc/dictation#operation/wsMessageHeartbeat
• method:BOUNDARYREQUEST Now Working for ReDS - The BOUNDARYREQUEST method for WebSockets was working for the EDS, but not for the ReDS. This has been rectified.
  No changes were made to the documentation, only functionality.(D-12830, 11/9/2021)
  • https://test.nvoq.com/apidoc/dictation#operation/wsMessageBoundaryRequest