We’re currently addressing Voice Pali pronunciation fixes for version 2.3. Our goal is not to have perfect Pali pronunciation, which is a bit out of reach for current AI. Instead, our goal is to prevent confusion and misunderstanding.
For example, here is an example where Aditi hiccups unintelligibly on vīthiṃ:
Idha, bhikkhave, gopālako na rūpaññū hoti, na lakkhaṇakusalo hoti, na āsāṭikaṃ hāretā hoti, na vaṇaṃ paṭicchādetā hoti, na dhūmaṃ kattā hoti, na titthaṃ jānāti, na pītaṃ jānāti, na vīthiṃ jānāti, na gocarakusalo hoti, anavasesadohī ca hoti, ye te usabhā gopitaro gopariṇāyakā te na atirekapūjāya pūjetā hoti—
Getting pronunciation right is very very tricky. We are using Amazon Web Services Polly API to correct Aditi’s speech. In particular, we are using custom IPA for Aditi.
Let’s discuss Voice Pali pronunciation problems and/or fixes in this thread. If you are passionate or curious about pronunciation fixes, you can use the AWS Polly Console to experiment and help us find the “right sound”.
Oh my, I think Aditi has the hiccups! That really must be some super sophisticated simulation technology to even be able to capture the human hiccup. Sadhu!
The proposed change affects all Pali long ī. Here is the passage after curing Aditi’s hiccups. Note that long ī occurs in pītaṃ as well as vīthiṃ and anavasesadohī:
Idha, bhikkhave, gopālako na rūpaññū hoti, na lakkhaṇakusalo hoti, na āsāṭikaṃ hāretā hoti, na vaṇaṃ paṭicchādetā hoti, na dhūmaṃ kattā hoti, na titthaṃ jānāti, na pītaṃ jānāti, na vīthiṃ jānāti, na gocarakusalo hoti, anavasesadohī ca hoti, ye te usabhā gopitaro gopariṇāyakā te na atirekapūjāya pūjetā hoti.
Sounds good to me, thank you so much! I’m actually glad that there is another long “i” in this passage, so we can see there is no devastating effect on it. I hope this is so throughout—otherwise we’d have to live with the hiccups…