SutaCentral, read me a sutta (Emma)

Text to speech vendors typically provide a selection of “voices”. There are a LOT of different voices, including Indian English (!). American voices tend to be direct and action-oriente. Indian English voices do well at packing lots of information in a short time. Etc. And UK voices seem to invite a careful consideration of novel and nuanced perspectives (e.g., “Mind the gap!”).

IPA (or International Phonetic alphabet) is a standard alphabet for representing the sounds of the world’s languages. IPA is usually combined with SSML to specify things like volume, pitch, rate and pauses. Although IPA and SSML are the building blocks, artificial intelligence (AI) is the framework that joins it all together to give us text-to-speech as well as speech-to-text. This technology enables conversational interaction with software agents such as Siri that act on our behalf. For those with visual disabilities such as myself, the technology offers hope. Also, the more I study the suttas, the more I realize the benefit of recitation–the oral tradition is amazing!

3 Likes

I know the IPA; I never heard of SSML until this moment, but I did once know some HTML and it looks like there’d be a limited amount of SSML syntax to absorb (I was a teacher of applied linguistics but have never been an IT nerd). If I can help I’d love to.

It looks time-consuming but would be so worthwhile doing (a couple of my friends are losing their sight and I have no idea which of my six senses will fade out first). Obviously before starting there’d be good consultation with the folks who are organising the real voice recordings; it would be hideous to find we’ve doubled up!

2 Likes

Absolutely. I’m so glad the PaliAudio folks spoke up. Although SlowAmy could read entire suttas, we already have most excellent human voice recordings thanks to PaliAudio. SlowAmy should fill in the gaps. For example, one might use SlowAmy to ask, “Hey SuttaCentral, what is the root of suffering?” and would get back 13 one-liners spoken by SlowAmy. This would permit progressive discovery by means of voice interaction and provide direct assistance in searching the suttas. This has been my own interaction style with typed in queries. I follow little sparks of inquiry here and there and am introduced to new suttas all the time in a way that connects them all in my head. I’d like to bring voice assistance to this mode of search in the hopes that others may also find such interaction helpful.

Thank you for your very kind offer of help. It is indeed a painstaking task. I spent hours getting the right IPA for bhikkhu. and Tathagata. However, with more of us helping a little bit, the task actually feels doable!

:pray:

5 Likes

Dave, it just occurred to me that providing search links to PaliAudio suttas would make for quite an interesting search experience. SuttaCentral could help “find a voice recording of a sutta discussing root of suffering”, and then provide a Google link to your site.. This would clearly take a lot of work, but I think it may be of value.

2 Likes

We would, rather, integrate the audio by hosting the files ourselves. With attribution to paliaudio, of course.

2 Likes