Skip to content

Latest commit

 

History

History
45 lines (27 loc) · 2.08 KB

README.md

File metadata and controls

45 lines (27 loc) · 2.08 KB
Raw

node-red-contrib-wavenet

A Node-RED custom node to convert text to speech using Google Wavenet voice.

Update: Now optional override of all parameters like voice, speaking rate, pitch, volume, sample rate and audio encoding. No need to wait for our updates to try new voices from Google!

Wavenet voice represents a new way of creating synthetic speech, using a WaveNet model, the same technology used to produce speech for Google Assistant, Google Search, and Google Translate.

Takes a string as an input. Returns a base64 encoded array buffer. Attach a file node to save it as an audio file on your disk.

main flow

Sign in to Google Developer Console to get a Speech API key.

Inputs

payload (string)

  • Use SSML elements in your text to give it more personality. You can use all elements found here.

The following can be used to override the config set in the node:

voice (object)

  • Choose a specific voice or random

rate (number | 0.25 to 4.0)

  • In the range of 0.25 to 4.0. 1.0 is the normal native speed. 2.0 is twice as fast, and 0.5 is half as fast.

pitch (number | -20 to +20)

  • In the range -20.0 to 20.0. 20 means increase 20 semitones from the original pitch. -20 means decrease 20 semitones from the original pitch.

volume (number | -6 to +6)

  • A value of -6.0 (dB) will play at approximately half the amplitude and a value of +6.0 (dB) will play at approximately twice the amplitude of the normal native signal.

sample (number)

  • The synthesis sample rate (in hertz). When this is specified, the synthesizer will honor this request by converting to the desired sample rate. Default is 24000.

encoding (string)

  • The encoding format of the audio byte stream can be MP3, LINEAR16 or OGG_OPUS

Output

payload (string | base64)

  • A base64-encoded string that is wrapped in containers according to the encoding specified.