speaker identification add-on barely changes the economics assembly ai's current async price is 21

hour for universal 3.5 pro plus meter hour for standard diarization equals a 23 surround that is

0.83 to minute per minutes about 0.383 cents per minute adding speaker identification takes it to

0.25 per hour equals 0.417 cents per minute therefore dropping identification saves only

for donna 3v3's minute or 0.33 cents minute for your 64 hour week transcriptions diarization but

without assembly ai speaker identification would be about 14.72 weeks a week or 63.79

month average a 12-hour weekday costs 2.76 a two-hour weekend day costs four six cents

so compared with the 16 week 69.33 a month figure from before you're only saving about 1.28 this

week or 5.55 the expensive bit relatively speaking is the transcription itself not

identification nvidia titanet is much more interesting for your 8 gigabyte gpu the nvidia

model i was referring to is titanet large its job is very specific

what b-ray needs for persistent identity