New Acoustic Model
Acoustic models account for the way in which end users will interact with your app.
Previously, our Speech to Text contexts were based on a “smartphone” acoustic model, meaning that our contexts worked best for apps used by customers with a smartphone—speaking directly into the microphone.
We recently updated all of our Speech to Text contexts to also support the recognition of speech at a limited distance—so our Speech API now supports use cases where the customer is near a microphone, but not directly in front of it (within about 10 feet). For example, interacting with a tablet or a device that is in relative proximity to a customer—like a voice-enabled appliance.
A new X-Arg parameter called ‘Microphone’ was introduced in V3 of the Speech API in which you can specify which acoustic model best meets the needs of your app:
- ‘smartphone’ or