Voice recording

Project Scope To provide wake word utterances audio collection and relevant

transcriptions
Out of Scope NA
Deliverable To provide wake word audio collection and relevant transcriptions
> The required data set must contain voice recordings with wake word and
sample commands in US English along with transcription and metadata

Requirements Utterances:

> There are 4 categories to cover. C1, C2, C3 and C4.
Each category contains 3000 utterances.
>  Utterances per speaker: 
1. Category C1-C3 : Max. 5 utterances per speaker per category
 Utterances can be chosen randomly from each category for each
speaker. Overall, all
            utterances must get covered equally. 
 
1. Category C4 : All utterances per speaker 
Total Approx. participant required for the project is 600 (For C4, we will
reuse any 334 speakers from a pool of 600 speakers who participated in
C1, C2, C3)