Can I submit an audio role play instead of video?

No.

Given the nature of online training, the assessors do not have face to face contact with you and therefore cannot collect evidence of your eye contact, body language etc in an audio file.  We also cannot evidence a student simply by their voice.   Hence the need for a video recording.

Note:  Even if an assessment task requires you to role play a telephone conversation, you still record this as a video.