Spoken Pronunciation Task Force Teleconference

13 April 2022


Alan_, iali, janina, PaulG, Sam, SteveNoble, SuzanneTaylor
Irfan Ali

Meeting minutes

document from Accessibility for Children CG https://docs.google.com/spreadsheets/d/14EsfShLQ9hacvkc-hpb2MYi-Rn5GZkjO4S0IVg_gmLU/edit#gid=0

Agenda Review & Announcements

janina: sorry I'm late, technical difficulties

<SuzanneTaylor> https://docs.google.com/spreadsheets/d/14EsfShLQ9hacvkc-hpb2MYi-Rn5GZkjO4S0IVg_gmLU/edit?usp=sharing

iali: we have Maud and Susan here to talk about the Accessibility for Children CG's meeting about pronunciation

Children Pronunciation Use Cases (Invited experts)

<SuzanneTaylor> initial organization of use cases from our first meeting on this topic: https://docs.google.com/spreadsheets/d/14EsfShLQ9hacvkc-hpb2MYi-Rn5GZkjO4S0IVg_gmLU/edit?usp=sharing

SuzanneTaylor: we're very preliminary and very surprised to be invited to soon.

SuzanneTaylor: in our initial discussion, people brought all kinds of ideas to contribute to our use cases covering context, tone of voice, and cultural linguistics

SuzanneTaylor: we placed our first discussion notes into this spreadsheet and I expect we will find even more to fill in during our upcoming meetings

SuzanneTaylor: this is of particular interest in educational and crisis situations.

Maud_: we seem to have a lot of information from several contexts and it would be interesting if you had additional contexts for us to consider

Maud_: Is there a privacy issue in pronunciation for children?

Maud_: could it be used to detect their current state or stage of development?

Maud_: it's good to know how far this group is willing to go to accomodate our group's work

janina: I want to address how this group can use what you presented and I want to share information from other groups in W3C that would probably be useful for your work

janina: for Pronunciation TF, we are not all that interested in the recording or microphone side of things. We think we'll make the technologies that use that "smarter" as a byproduct.

janina: we are very welcoming of the use cases that I personally was struggling to verbalize

janina: we only have codes for a TTS engine to express general language/locale. We don't have the technology to specify a child's voice.

janina: it's not enough to say "English as in the UK" but we want "English as in the time of Shakespeare"

janina: we need the ability to express these variations. You gave us use cases for why it's important to include these variations.

janina: as part of APA, we have the ability to recruit others into solving a problem and fill this gap in the specs

janina: With the specs in place, the technology companies can bring their tools into the educational contexts where it's needed

janina: I would strongly suggest we move our use cases to note status. We will eventually have a technical recommendation, our spec, for how to do this.

janina: we have to go down the path with the browser venders. We want to make sure the path we choose will lead to these minimal cases. Specifying children's voices is an important contribution and should be part of our minimum use cases.

SuzanneTaylor: it's exciting that you'd like to formalize this into a problem statement

SuzanneTaylor: possibly more important than a child's voice, we have a need to specify the voice is positive and friendly.

SuzanneTaylor: how an adult might speak to a child

<iali> https://www.w3.org/TR/pronunciation-gap-analysis-and-use-cases/

SuzanneTaylor: if you can tell us the best place to write those use cases, my group would rather contribute to the right format than create something else

PaulG: they could open issues in Github.

<iali> https://github.com/w3c/pronunciation/issues

janina: this will go into our gap analysis to show we don't have the technology to describe these contexts.

janina: and it's beyond accessibility, it's needed by mainstream ed tech

<iali> https://www.thedrum.com/news/2018/11/02/young-people-are-more-likely-favour-female-voices-radio-and-digital-assistants

janina: language learners benefit from tooling that correctly pronounces the words on a page

ACTION: Irfan to update the use case document based upon the github issues.

<trackbot> Created ACTION-41 - Update the use case document based upon the github issues. [on Irfan Ali - due 2022-04-20].

janina: and I think we'll need a new specification to address it

janina: I'll send another link to you for WAI-ADAPT (formerly personalization)

SuzanneTaylor: we still have to bring our document back to the CG. I'd like to work in github directly but it's up to the wider group.

janina: W3C isn't ready to start publishing google docs. We want to capture this notion and we'll need to incorporate it into our HTML docs.

janina: the most educational publishers are in digital publishing group (EPUB)

janina: if you can give us two issues in our gap analysis, it will be up to us to figure out what to do with it

janina: you can always edit, expand, refine, or abandon github issues. I'd recommend starting earlier rather than later and not waiting for a final polished state

SuzanneTaylor: we can always add comments as well

janina: EPUB is github driven as well, your team is free to continue using the google doc if you like

Maud_: we're quite new and it's helpful to understand how the TF works

janina: are we dangerously close to re-opening SSML?

<iali> PaulG: Think that there are some gaps and some of the things in SSML were considered for implementation details. Such as voice option. if you know the target TTS engine, you can give additional hint through the variant element of voice.

<iali> janina: all these use cases are very important. It also covers accessibility

https://www.w3.org/TR/speech-synthesis11/#S3.2.1 talks about variant which is more like a way to target specific TTS engines with hints but we're talking about adding standard characterstics

<iali> PaulG: we are going beyond SSML. with this information we are passing it.

<iali> janina: perhaps a follow up meeting with ARIA will be helpful with these use cases

<iali> janina: we can have explanatory interim meeting with ARIA to discuss this

PaulG: would it be possible to separate pronunciation content (e.g., HTML) from pronunciation presentation (e.g, CSS) so that we divide the use cases and get something out while still iterating on the problem

janina: that would be fine in my opinion

janina: the language meta info wasn't as persuasive but these use cases for children are more compelling

Other Business

Summary of action items

  1. Irfan to update the use case document based upon the github issues.
Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).


Maybe present: Maud_