Comment 1 by briang1 on 2013-01-11 20:04
Well I've obviously led a shelteredlife, as I'd never heard of this one. Is this more specialised than most people might use though?

Comment 2 by jteh on 2013-01-11 22:32
This needs to be done in individual speech synthesisers. NVDA can't do this itself.

If you're talking about eSpeak, there might be a chance to consider this. Otherwise, contact the developer of your speech synthesiser.

Comment 3 by pigsonthewing (in reply to comment 1) on 2013-01-12 00:45
Replying to briang1:

Well I've obviously led a shelteredlife, as I'd never heard of this one. Is this more specialised than most people might use though?

Depending on sources, Wikipedia is the 5th or 6th most visited website on the Internet. Its IPA help page alone is visited between 20-40,000 times /per day/: http://stats.grok.se/en/latest30/Help:IPA_for_English And Wikipedia is not the only site using IPA.

I have also received a feature requirement for IPA from a linguist, and many dictionaries also seem to use IPA to indicate pronunciation.

Synthesiser may not be able to provide IPA sounds as IPA symbols use greek, latin and other symbols that are also used for other purposes so context switching is needed in order to read IPA. For example, symbol β read as beta by ellequence and s set by eSpeak. This symbol should be read as va sound for IPA.

How about an addon which could turn on or off IPA mode as and when required so that users could navigate character by character with IPA mode on and symbols would be read as such.

Does anyone have any thoughts/inputs for this ticket, particularly in response to #2922 (comment)?

For anyone looking at this issue, a user who also requested this recently provided the following links:
https://www.internationalphoneticassociation.org/content/ipa-chart (Official page)

https://linguistics.ucla.edu/people/keating/IPA/IPA_charts_2018.html (The "List of symbols and diacritics with descriptions & identifiers: PDF (plain)" file linked to on this page has more detailed pronunciation).

My understanding from the user is that these symbols should be pronounced the same way all the time so a solution which enables turning pronunciation on and off may not be required over one which simply ensures the correct pronunciation is always used.

My thought then was that one solution could be if reading text with the β character for instance, that it should be pronounced as it is designed to be (the va sound according to the previous comment), and if spelt phonetically, then should be described as "Beta".

Are there plans to add support for this? I am the user who found the links, and with the full pronunciation, unicode values and character names, adding this into the symbols dic should be very easy.

I can imagine that this is quite har to implement in all languages. What would be the benefit of this for an user? Cjould someone describe some use cases?

For example, this symbol which is used as an example in this issue is used very often in german in many words instead of ss and is pronounced exactly like a double s in a word. But when navigating letter by letter, the symbol is pronounced sz which is in line with the german gramar.

Replying here to @DavidBitSource request in #11585 :
Pleas specify clearly your need:

• voice feedback: please specify what you expect:
  • do you need to have IPA character names pronounced correctly?
  • do you need to have IPA written word spoken according to IPA notation by the synthesis?
  • or both?
  • what do you expect for each IPA character name? Is it something language specific or not?
• braille feedback: You suggest X-SAMPA notations. Please specify (or give examples) what is missing in braille IPA representation that is present in X-SAMPA notations.

I had thought to create an add-on to support such capability, at least the voice part, since I do not use regularly braille and do not know so much about braille IPA (even if I know that it exists). But this seems to be a big work and I do not think I will have time to dedicate to it so soon.

the IPA chart describes place of articulation and manner of articulation, diacritics specify details.
I need the symbols named, according to the table and its markers. The chart doesn't say "b", it gives you a voiced bilabial plosive, diacritics can modify this further, as it appears on the chart.
I do not need a morph pronounced, since I do not want to study a certain speaker, which would be the synthesizer engine in this case.
This is definitely not language specific, it is a matter of phonetics.
There are several braille notation approaches, but they also lack features, e.g. diacritics or parts of the IPA itself. This makes it impossible to use this for a thesis or academic research.

@nvaccess - I'm interested in updates on IPA (International Phonetic Alphabet) symbol support. Is there any progress on enabling full pronunciation, Unicode values, and spoken names for these characters? Adding complete IPA support with speech synthesis so NVDA users can accurately read IPA notation would be extremely valuable for linguists, language learners, language teachers and so on. Please provide any updates on the status or priority for this in future NVDA releases.

To progress this issue, it would be nice that the people requesting this feature provide a description of the goal to achieve and an example of steps that the user would perform to use this new feature.

Here is an example:
Use case: Have an IPA mode allowing to report unambiguously each IPA character when navigating character by character, so that the reader can figure the pronunciation of the word.
Steps:

• Go to Wikipedia Glasgow page
• Search for "Glasgow (UK: " string which is just before the IPA transcription of the name
• Press a gesture to enable IPA mode
• Press right arrow various times
  Result: I want NVDA to report the Unicode name of the characters that make up the word /ˈɡlɑːzɡoʊ/, i.e.:
  "SOLIDUS", "MODIFIER LETTER VERTICAL LINE", "LATIN SMALL LETTER SCRIPT G", "LATIN SMALL LETTER L", "LATIN SMALL LETTER ALPHA", "MODIFIER LETTER TRIANGULAR COLON", "LATIN SMALL LETTER Z", "LATIN SMALL LETTER SCRIPT G", "LATIN SMALL LETTER O", "LATIN SMALL LETTER UPSILON", "SOLIDUS".

Of course, this description has to be modified according to your needs. You will probably not need the Unicode character name but another one: indicate the source document (e.g. link of webpage or PDF, paragraph, etc.) as well as the precise location (e.g. column n in table "xyz" or column named "foo" in table named "bar").
It is also not clear to me if you want just the characters reported a specific way or if you want the whole word pronounced as indicated by IPA.
Thanks.