Handling relevance response

The pronunciation and fluency interpretation of the key elements in the response of the spoken word or sentence remains the same as previously described. The parameter which we need to notice is the relevance check.

Relevance Check

If the response is relevant, then relevance.class will be TRUE:

If the response is not relevant to the context, then relevance.class will be FALSE and the error will be shown:

If the response is too similar to the question prompt, then the error will be as follows:

Other fields

Field
Description

transcript

The speech-to-text transcript of what the user has said.

speechace_score

An overall score on a scale of 0 to 100, in addition to subscores for: Fluency, Pronunciation, Grammar, Vocabulary, Coherence.

ielts_score

An overall score on an IELTS scale of 0 to 9.0, in addition to subscores for: Fluency, Pronunciation, Grammar, Vocabulary, Coherence.

pte_score

An overall score on a PTE scale of 10 to 90, in addition to subscores for: Fluency, Pronunciation, Grammar, Vocabulary, Coherence.

cefr_score

An overall score on CEFR scale of A0 to C2, in addition to subscores for: Fluency, Pronunciation, Grammar, Vocabulary, Coherence.

toeic_score

An overall score on an TOEIC scale of 0 to 200, in addition to subscores for: Fluency, Pronunciation, Grammar, Vocabulary, Coherence.

relevance.class

TRUE or FALSE indicating whether the response was relevant given the relevance_context passed as input to the API.

Protip: The relevance_context is an optional field and can be included only if the relevance of an audio response needs to be assessed by transcribing the response first and then comparing it to the provided question prompt.

Last updated