Generate speech audio from super long text through machine, via Google TTS, ffmpeg.

Do you want to:

• Share your ideas anonymously on YouTube?
• Pretend on phone that you are not a kid (Home Alone)?
• Learn good english pronunciation for a speech?
• Make your computer read out your school notes?
• Experiment with various voices from around the globe?
• Or, Upload Wikipedia TTS videos on YouTube?

Sample: "I want to order a stuffed crust pizza".

Setup

Install

1. Run npm install -g extra-googletts in console.
2. To install this as a package use npm install extra-googletts.
3. Make sure ffmpeg is installed (present in PATH).

Get service account key

1. Create an account on Google Cloud Platform.
2. Create a new project, and select it.
3. Enable Cloud Text-to-Speech API for the project.
4. Add credentials to your project.
5. Which API are you using? Cloud Text-to-Speech API.
6. Are you planning to use this API with App Engine or Compute Engine? No, I’m not using them.
7. Select What credentials do I need?.
8. Create a service account.
9. Service account name: googletts (your choice).
10. Role: Project -> Owner.
11. Service account ID: googletts (same as name).
12. Key type: JSON.
13. Select Continue.
14. Copy downloaded file to a directory.
15. Rename the file to account_id.json.

Set environment variable

1. Copy path of account_id.json.
2. Set environment variable GOOGLE_APPLICATION_CREDENTIALS to it.

On Windows, use RapidEE to set environment variable.

1# on linux or macos console
2export GOOGLE_APPLICATION_CREDENTIALS="[PATH OF account_id.json]"
3
4# on windows powershell
5$env:GOOGLE_APPLICATION_CREDENTIALS="[PATH OF account_id.json]"

Console

1googletts "I want to order a stuffed crust pizza"
2# out.mp3 created (yay!)
3
4googletts -t speech.txt -o speech.mp3
5# speech.mp3 created from text in speech.txt
6
7googletts "Hello 911, my husband is in danger!" -vg female
8# out.mp3 created with female voice
9
10echo "Dead man walking." | googletts --log -vn en-US-Wavenet-B
11# out.mp3 created with different male voice (log enabled)

Available TTS voices?

Reference

1googletts [options] <text>
2# text: input text
3
4# Options:
5# --help:        show this help
6# -l, --log:     enable log
7# -o, --output:  set output audio file (out.mp3)
8# -t, --text:    set input text file
9# -r, --retries: set speech synthesis retries (8)
10# -a, --acodec:  set acodec (copy)
11# -ae, --audio_encoding:  set audio encoding (MP3)
12# -af, --audio_frequency: set audio frequency/sample rate
13# -lc, --language_code:   set language code (en-US)
14# -vn, --voice_name:      set voice name
15# -vg, --voice_gender:    set voice gender (neutral)
16# -vp, --voice_pitch:     set voice pitch change (0.0)
17# -vr, --voice_rate:      set voice speaking rate (1.0)
18# -vv, --voice_volume:    set voice volume gain in dB (0.0)
19# -qb, --quote_break:        set quoted text break time (250)
20# -qe, --quote_emphasis:     set quoted text emphasis level (moderate)
21# -hb, --heading_break:      set heading text break time (4000)
22# -hd, --heading_difference: set heading text break difference (250)
23# -he, --heading_emphasis:   set heading text emphasis level (strong)
24# -eb, --ellipsis_break:     set ellipsis break time (1500)
25# -db, --dash_break:         set dash break time (500)
26# -nb, --newline_break:      set newline break time (1000)
27# -bs, --block_separator:    set block separator (.)
28# -bl, --block_length:       set block length (5000)
29# -ccf, --config_credentials_file: set google credentials path
30
31# Environment variables:
32$TTS_LOG     # enable log (0)
33$TTS_OUTPUT  # set output audio file (out.mp3)
34$TTS_TEXT    # set input text file
35$TTS_RETRIES # set speech synthesis retries (8)
36$TTS_ACODEC  # set acodec (copy)
37$TTS_AUDIO_ENCODING     # set audio encoding (MP3)
38$TTS_AUDIO_FREQUENCY    # set audio frequency/sample rate
39$TTS_LANGUAGE_CODE      # set language code (en-US)
40$TTS_VOICE_NAME         # set voice name
41$TTS_VOICE_GENDER       # set voice gender (neutral)
42$TTS_VOICE_PITCH        # set voice pitch change (0.0)
43$TTS_VOICE_RATE         # set voice speaking rate (1.0)
44$TTS_VOICE_VOLUME       # set voice volume gain in dB (0.0)
45$TTS_QUOTE_BREAK        # set quoted text break time (250)
46$TTS_QUOTE_EMPHASIS     # set quoted text emphasis level (moderate)
47$TTS_HEADING_BREAK      # set heading text break time (4000)
48$TTS_HEADING_DIFFERENCE # set heading text break difference (250)
49$TTS_HEADING_EMPHASIS   # set heading text emphasis level (strong)
50$TTS_ELLIPSIS_BREAK     # set ellipsis break time (1500)
51$TTS_DASH_BREAK         # set dash break time (500)
52$TTS_NEWLINE_BREAK      # set newline break time (1000)
53$TTS_BLOCK_SEPARATOR    # set block separator (.)
54$TTS_BLOCK_LENGTH       # set block length (5000)
55$GOOGLE_APPLICATION_CREDENTIALS # set google credentials path

Package

1const googletts = require('extra-googletts');
2
3await googletts('out.mp3', 'I want to order a stuffed crust pizza');
4// out.mp3 created (yay!)
5
6const fs = require('fs');
7var speech = fs.readFileSync('speech.txt', 'utf8');
8await googletts('speech.mp3', speech)
9// speech.mp3 created from text in speech.txt
10
11await googletts('out.mp3', 'Hello 911, my husband is in danger!', {
12  voice: {gender: 'FEMALE'}
13});
14// out.mp3 created with female voice
15
16await googletts('out.mp3', 'Dead man walking.', {
17  voice: {name: 'en-US-Wavenet-B'}, log: true
18});
19// out.mp3 created with different male voice (log enabled)

Reference

1const googletts = require('extra-googletts');
2
3googletts(output, text, options={})
4// output:  output audio file
5// text:    input text
6// options: given below
7// -> Promise <table of contents>
8
9// Default options:
10options = {
11  stdio: [0, 1, 2], // set child process stdio
12  log: false,       // enable log
13  retries: 8,       // set speech synthesis retries
14  acodec: 'copy',   // set audio acodec
15  audio: {
16    encoding: 'MP3',     // set audio encoding
17    frequency: 0,        // set audio frequency/sample rate
18  },
19  language: {
20    code: 'en-US'        // set language code
21  },
22  voice: {
23    name: null,          // set voice name
24    gender: 'NEUTRAL'    // set voice SSML gender
25    pitch: 0.0,          // set voice pitch change
26    rate: 1.0,           // set voice speaking rate
27    volume: 0.0,         // set voice volume gain in dB
28  }
29  quote: {
30    break: 250,          // set quoted text break time
31    emphasis: 'moderate' // set quoted text emphasis level
32  },
33  heading: {
34    break: 4000,         // set heading text break time
35    difference: 250,     // set heading text break difference
36    emphasis: 'strong',  // set heading text emphasis level
37  },
38  ellipsis: {
39    break: 1500          // set ellipsis break time
40  },
41  dash: {
42    break: 500           // set dash break time
43  },
44  newline: {
45    break: 1000          // set newline break time
46  },
47  block: {
48    separator: '.'       // set block separator
49    length: 5000,        // set block length
50  },
51  config: {
52    credentialsFile: null // set path to credentials
53  },
54  params: null            // set synthesize speech parameters "directly"
55}

Similar

Do you need anything similar?

• extra-youtubeuploader can upload videos with caption to YouTube.
• extra-stillvideo can generate video from audio and image.

Suggestions are welcome. Please create an issue.

References: SSML, TTS voices, TTS client docs.