W3C

– DRAFT –
Spoken Pronunciation Task Force Teleconference

30 March 2022

Attendees

Present
Alan, Dee, iali, janina, NeilS, PaulG, Sam, Sarah_Wood, SteveNoble
Regrets
-
Chair
Paul
Scribe
iali, Irfan Ali

Meeting minutes

Agenda Review, Membership & Announcements

Action Items

introduction to Sarah

<Sarah_Wood> +Present

<PaulG> https://www.w3.org/WAI/APA/task-forces/pronunciation/track/actions/open

Github Issues and examples

https://github.com/w3c/pronunciation/issues

Other Business (Pronunciation Task Force and Children)

<PaulG> in the accessibility for children meeting, Susan Taylor asked about support specific to use cases for children using AT

<PaulG> iali: we discussed it last week and need to invite them to a future meeting

<PaulG> janina: we may need to research the various use cases

janina: is it useful in education to lead someone to provide an example or mimic the voice for children to make it easier to understand for the children.

janina: would mimicking voice useful in some cases?

<PaulG> AAVE is a dialect/ethnolect/sociolect

Sarah_Wood: from literacy department, there are some studies which indicates that it is helpful providing the mimicking voice is helpful in learning

SteveNoble: In video description or audio description, one of the best practice is to use children or younger sounding voices. Using adult voice for children character is not very effective.

<PaulG> Hawaiian pidgin is another example of how language is learned. Not sure if we have ability to spec that as well.

SteveNoble: there is a known factor but not sure how it might come play in literacy

Dee: there is lot of interest in general including children in pronunciation.

PaulG: we agree that we want to talk tot his group to get more information to write up some more use-cases.

janina: including children to support better support for pronunciation is helpful. Giving heads up about the use cases that we have discussed is good idea.

Designating lang for TTS

<PaulG> I'm also curious how cognates should be marked up. Should they remain in the context language or the borrowed language?

janina: I explained it a problem in APA meeting. but this is a subset of the main problem.

It is good idea to pronounce foreign language correctly.

we need to bring expertise in accessibility, we have other working groups in w3 who do the linguistic stuff. we do not have specific way about the language support for tts.

it has no provision for child voice vs male voice or female voice.

looked for bc-47 which is unicode. it has no concept of ancient language.

<PaulG> https://www.rfc-editor.org/bcp/bcp47.txt

most common language in the roman empire was ancient greek. Sacred scripture was written in greek. pronunciation that we have is not good to support these languages. All the specs we have today is bc 47 unicode.

<PaulG> https://datatracker.ietf.org/doc/html/rfc5646

I am proposing that what we have for language is doing what shoes up on the screen but doesn't help much in the world of TTS. Joint meeting at TPAC with some groups to help to specify the problems and work on the solutions

<PaulG> https://www.w3.org/TR/speech-synthesis11/#edef_voice covers "age" but not these other issues

Alan: it seems like we are moving beyond the schema. we also need the way to identify the voice pack

PaulG: we also need meta tagging which will allow author to write the required properties

PaulG: not sure if the browser parters or AT would like to implement this.

janina: internationalization group and other group that are working on linguistic could be the part of the joint meeting at TPAC. we have to show them that there is issue 144 in github which is blocking to go it to the CR. there is lot of correspondence which make internationalization work well.

janina: we need to request a join meeting by describing all these problems. It is a bigger deal than we can solve in github.

Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).

Diagnostics

No scribenick or scribe found. Guessed: iali