AFAIK that's just a case of choosing how many scans to use at encode time