Get speech mark from Amazon Polly using NodeJs

问题

I am working on an animation project to add subtitle to what my character says. I can get the mp3 file from AWS Polly with no issue.

However, when I want to get each part of the word separately, it doesn't work. I checked inspector tab, and I can see some params are passing to request to polly.aws. Any idea how I get json/mark-up file to know the start and end of each word & sentence?

const AWS = require('aws-sdk')
const Fs = require('fs')

const Polly = new AWS.Polly({
    signatureVersion: 'v4',
    region: 'us-east-1'
})

// # this part works fine
let params = {
    'Text': 'Hi, my name is Soley. We are building something amazing!',
    'OutputFormat': 'mp3',
    'VoiceId': 'Matthew'
}

// # from chrome's network tab:
// # and is there a way to get mp3 and mark-up text at the same time?
// "text": "Hi, my name is Soley. We are building something amazing!",
//     "textContentType": "text",
//     "voiceId": "Matthew",
//     "languageCode": "en-US",
//     "engine": "standard",
//     "outputFormat": "json-8000",
//     "lexiconNames": [],
//     "speechMarksTypes": [
//         "word",
//         "sentence"
//     ]

Polly.synthesizeSpeech(params, (err, data) => {
    if (err) {
        console.log(err)
    } else if (data) {
        console.log(data)
        if (data.AudioStream instanceof Buffer) {
            Fs.writeFile("speech."+params.OutputFormat, data.AudioStream, function (err) {
                if (err) {
                    return console.log(err)
                }
                console.log("The file was saved!")
            })
        }
    }
})

some useful links to check: https://aws.amazon.com/blogs/aws/new-amazon-polly-speech-marks/

using cli also works file: https://docs.aws.amazon.com/polly/latest/dg/speechmarkexamples.html but I want it in NodeJs

回答1:

Oh, I think I found something:

let params = {
    'Text': 'Hi, my name is Soley. We are building something amazing!',
    'OutputFormat': 'json',
    'VoiceId': 'Matthew',
    'SpeechMarkTypes': ['word', 'sentence']
}

Thanks to java: https://docs.aws.amazon.com/polly/latest/dg/SynthesizeSpeechMarksSample.html

来源：https://stackoverflow.com/questions/57022394/get-speech-mark-from-amazon-polly-using-nodejs

标签

amazon-web-services

text-to-speech

amazon-polly