In order to "...automatically reference each frame to start time of the file..." you have to know the start time of that file (well, duh!). Fortunately, sound files commonly include that information as part of their filename.
For example, I was recently asked to analyze a sound file with the following filename:
SWIFT #2_96K_20200621_172433.wav
Note the bolded portion, which contains the date_time that the recording started. This recording started on June 21, 2020 at 24 minutes and 33 seconds after 5 pm. The field recorder automatically appended that information to the filename as yyyymmdd_hhmmss, where:
yyyy = year
mm = month
dd = day
hh = hour (24)
mm = minute
ss = second
If this file contained a sound that had a start time of 14 seconds, then that time is simply added to the filename's date_time. In this case that sound would have started on June 21, 2020 at 24 minutes and 47 seconds after 5 pm.
AFAIK, most field recorders automatically include the starting date_time as part of the filename. If your sound files do not include this, then you'll need to locate some metadata that includes the recording's start time. Failing that, I don't know how you would be able to determine a start time...