SC-Voice is now able to parse a JSON description of MN1, split it up into 9 sections and expand all ellipses. The expanded MN1 JSON file shows the current output (n.b., it will change). In particular:
Expanded segments are automatically numbered at a more detailed level (e.g., mn1:19.2.1) according to their source segment (i.e., mn1:19.2)
Sections are inferred from the ellipsis grouping. Each expandable section corresponds to a single template, its replaceable segments and an optional concluding expanded template (e.g., the segments for ‘extinguishment’). Non-expandable sections are whatever is left over.
Sections are crucial for voice-navigation. They are also the atomic recitation units. I.e., Raveena would speak MN1/Section 9 The Realized One in its entirety.
The use of ellipses throughout the canon is idiosyncratic. This means that we’ll need to go through each sutta individually to attain voice assistance comparable with MN1. There is no magic wand of voice assistance.
Work continues–this is basically just a status report. I’ve added Vue.js/Vuetify.js as the user interface framework with Nodejs as the server REST backend. I’ve also started working on the HTML voice assistance page with aria tags and assisted HTML elements such as <details>. I am using ChromeVox as the screenreader for testing.
Breaking away from SoundCloud, I’ve created an AWS EC2 instance dedicated to SC-Voice. I’m currently testing the recitation API for sections. A long sutta such as MN1 is broken down into sections. For MN1 we have 9 sections. SC-Voice will return an entire section as a single audio file.
“Take an uneducated ordinary person who has not seen the noble ones, and is neither skilled nor trained in their teaching. They’ve not seen good persons, and are neither skilled nor trained in their teaching.
They perceive earth as earth.
But then they conceive earth, they conceive regarding earth, they conceive as earth, they conceive that ‘earth is mine’, they take pleasure in earth.
Why is that?
Because they haven’t completely understood it, I say. (…)
This section is long at 9:30 minutes. Let us know how it sounds to you…
That was amazingly good! I understood all the pronunciation. Brahma is a little rough.
There is one thing that actually stopped me listening to the end, and that was the speed. It was way too fast for me for effective listening. If nothing can be done about the speed, then I’d suggest at least putting a larger pause between each ‘set’ /paragraph.
In addition to Viveka’s message; marvellous, Karl! Thanks so much for all the work you’ve put into this - it’s kind of incredible that from the first post you expressed interest in going down this line you’ve got to here! Much applause.
Again to pick up from Viveka’s note, to my ear, it’s not the speed per se that is an issue, but the rhythm. I agree that it if were possible to add in ‘natural’ pauses it would make it a lot easier to follow.
That said, it is already very good. A hearty salute.
Aminah, Viveka, thank you for your thoughtful feedback. I will slow down run-on Amy and lend her a pause or two between major numbering changes. Also thanks for the sharp ear on Brahma. That is an odd one!
The SC-Voice web page is coming along and I’ll be releasing it for review next week along with the changes inspired by your recommendations.
