How about allowing voice recordings to be associated with point locations in Collector??.
In many situations, it is easier (and the only option!) to simply record a voice file at a location where a point location is recorded. This is possible in ArcGIS 10.1, but not in Collector. A link to the .wav file in ArcGIS is then linked to the attribute table and the file itself is saved. The user then transcribes the data when they are back in the office. We use this for wildlife aerial surveys, when it is not practical to take your eyes over a survey transect and look down at a tablet to record data that include many attributes associated with a single point (habitat, group size, behavior, etc).