By: Bobby Sudekum
We upgraded our voice guidance in both our iOS and Android Navigation SDKs to Amazon Polly, which generates audio guidance from the turn-by-turn text instructions streaming from our APIs. Developers don’t need to provide any extra access tokens or pay for accessing Polly services — it’s all included as part of our SDKs.
Poor voice guidance and incorrect pronunciation can make it hard to understand driving instructions. According to JD Power, the leading automotive research group, “In-vehicle voice recognition is failing.” Of all issues in new cars, navigation, voice, and Bluetooth receive the highest frequency of complaints.
Polly lets us fine-tune instructions so they sound the way a person would naturally speak. First, listen to this comparison between the default speech synthesizer on iOS and our new navigation SDK voice guidance:
Both voices read: “In 400 feet turn left onto Coeur d’Alene Avenue and then turn left onto I-280”.
AV speech synthesizer:
Amazon Polly guidance:
Our new guidance isn’t just read by a more natural sounding voice; we also now correctly pronounce both the street name “East Coeur d’Alene Avenue” and the route number (as “eye two-eighty”).
Polly can even pronounce street names that have unintuitive pronunciations, like Houston Street in New York City:
New releases for both our Navigation Android and iOS SDKs are ready for download and include the upgrade to Polly.
Turn-by-turn voice guidance upgraded to Amazon Polly was originally published in Points of interest on Medium, where people are continuing the conversation by highlighting and responding to this story.