New – Long-Form voices for Amazon Polly

We’re launching three new voices for Polly. Powered by a brand new long-form engine, the voices are pure and expressive, with applicable pauses, emphasis, and tone.

New Voices
The brand new long-form voices are excellent for weblog posts, information articles, coaching movies, and advertising content material. The underlying Machine Studying mannequin extracts which means from the textual content, studying about speech segments, prosody (the sample of rhythm and pauses), intonation, and different points of expressive speech, permitting the synthesized audio to specific feelings, particularly in dialogs. The brand new long-form engine makes use of a deep studying text-to-speech (TTS) mannequin educated to amass a contextual understanding of the textual content that enables it to specific prosody in an applicable method. This permits the intention of the story to drive the vocal efficiency and create the right emphasis, pauses, and tones of a sensible human voice.

Listed here are the brand new voices:

Utilizing the New Voices
You’ll be able to entry the brand new voices utilizing the AWS Administration Console, AWS Command Line Interface (AWS CLI), or the AWS SDKs. Utilizing the CLI, I begin by itemizing the voices that use the brand new long-form engine:

$ aws --region us-east-1 polly describe-voices --output json 
  | jq -r '.Voices[] | choose(.SupportedEngines | index("long-form")) | .Title'
Danielle
Gregory
Ruth

I can decide one, or I can strive all of them:

for v in `aws polly describe-voices --output json 
          | jq -r '.Voices[] | choose(.SupportedEngines | index("long-form")) | .Title'`; do
    Textual content="Hi there my identify is $v and I can learn weblog posts, articles, 
and different long-form content material for you. I'm the very best!"
    aws polly synthesize-speech --output-format 'mp3' 
    --text "$Textual content" --voice-id $v $v.mp3 --engine long-form; 
    aws s3 cp $v.mp3 s3://jbarr-voices; 
performed

My shell script had a small quoting bug, however the ensuing audio was too humorous to not embrace!

Programmatically, you’ll be able to reproduce my instance by writing code that calls the DescribeVoices and SynthesizeSpeech features.

Issues to Know
Listed here are some fascinating issues that you need to know concerning the new voices:

Pricing – Lengthy-form voices are priced at $100 per million characters or Speech Marks requests. Take a look at the Amazon Polly Pricing web page to be taught extra.

Engines & Voices – A number of the voices that I listed above can be utilized with multiple engine. For instance, the Danielle voice can be utilized with the brand new long-form engine and the present neural engine.

Areas – The brand new engine and voices can be found within the US East (N. Virginia) Area.

Take a look at the brand new voices, construct one thing superior, and let me know what you assume!

— Jeff;