Voice

Transcription

The speech recognition engine, the call language, and the noise suppression that cleans the caller's audio before the agent hears it.

Transcription configures speech recognition: the engine that transcribes the caller, the language the call runs in, and how their background noise is cleaned before it arrives. The engine picker is at the top of the drawer. Language and noise, turn-taking, recognition hints and vocabulary sit under Advanced.

Two parts of it have pages of their own:

  • Turn taking, for how patiently the agent waits before deciding the caller has finished.
  • Recognition hints, for the words the agent must not garble.

The settings on this page are versioned. They belong to the draft, reach live calls when you publish, and are read-only on a published version. Recognition hints and Vocabulary are the exceptions: they apply across every version and reach live calls as soon as they save.

Speech to text

The engine that turns the caller's audio into text. By default speech recognition runs in Australia. The picker lists each engine by name; this table describes the six by what they do.

EngineLanguagesTurn detectionRecognition hints
The defaultEnglish onlyOn the engineYes
MultilingualMultilingualPlatformYes
English, turn-nativeEnglish onlyOn the engineYes
Multilingual, turn-nativeTen languages, detected each turnOn the engineYes
Multilingual, previous generationMultilingualPlatformYes
Multilingual, no hintsMultilingualPlatformNo

The default engine

Engines are grouped by provider in the picker, and the platform default is marked Default.

The platform default decides turn boundaries on the engine itself, which removes a hop and makes conversations feel tighter. It is English only and accepts recognition hints. An agent that has never picked an engine runs it.

The multilingual engine is the right alternative when you need multilingual support.

Who decides the caller has stopped

Three engines transcribe only, so the platform decides turn boundaries from the audio. The other three, the default and the two turn-native engines, decide on the engine itself and report the end of a turn directly.

That removes a hop, and it also takes several of the turn-taking controls out of your hands. Turn taking covers which ones.

The other engines

The turn-native multilingual engine detects the caller's language each turn, so a caller can switch language mid-call and the agent's voice follows. With it selected, a Languages callers speak setting appears. Pick at least two for the voice to switch.

The previous-generation multilingual engine is kept for agents that were tuned against it. The multilingual engine without hints is the only one that ignores recognition hints, so an agent with a list of proper nouns should not use it.

Language

The language of the call, on both sides: what the agent listens in and what it speaks.

English, Spanish, French, German, Italian, Portuguese, Japanese, Korean, Chinese, Hindi and Arabic.

There is no separate setting for the spoken language, because a call cannot listen in one language and answer in another.

When the engine limits the choice

Not every engine transcribes every language. When the selected engine is English only, the other options are disabled and the card names what would unlock them.

Multilingual engine   every language selectable
Default engine        English selectable, the rest disabled

Publishing a version whose engine is English only, with another language selected, is blocked outright, so this cannot ship half-configured.

Two more things a non-English line needs

The language reaches a call when the version is published, so a change sitting in a draft is not yet answering anyone.

Check the voice as well: pick one on the Voice page that speaks the language you selected, because the setting tells the engine which language to speak, not which accent to speak it in.

The system prompt is worth a line too, since the model writes the words the voice reads.

Noise suppression

Two separate controls, because a phone call and a browser call clean audio in different places.

On a phone call

Noise suppression (phone) runs at the telephone carrier, on the caller's inbound audio, before it reaches the agent at all. It applies to real calls. Five options, each a different suppression engine:

OptionBehaviour
OffNo carrier-side cleaning.
StandardGeneral-purpose noise reduction, no tuning.
Steady noiseStronger separation of speech from steady background noise.
BroadBroad suppression, including other voices in the room.
Speech clarityThe default. Tuned for speech clarity on a phone line.

Suppression strength

Suppression strength appears for every engine that can be tuned, on a 0 to 100 scale in steps of 5, with a default of 70. Off has nothing to tune and Standard exposes no tuning, so the control is hidden for those two rather than shown dead.

Lower keeps the audio more natural. Higher removes more noise but can clip quiet speech, which costs you the ends of words on a hesitant caller.

On a browser call

Noise suppression (web calls) is a single switch that cleans the caller's microphone in the browser. It applies to calls that run in a browser, such as the in-app test call and the shared demo link, and has no effect on a phone call.

Turn taking and barge in

How patiently the agent waits, and how readily it stops talking when the caller cuts in, is tuned in this section too. Turn taking covers turn-start detection, the waiting controls, and what a turn-native engine decides for itself.

On this page