NVDA incorrectly speaks 'normalizes' for characters with combining circumflex (ê, ô)

[ObjectInSpace]

certain accented characters will cause the word "normalized" to be appended to the spoken output. This occurs irrespective of the "unicode normalization" and "report normalized characters when moving by character" settings, even while not moving by character.

Steps to reproduce:

1. Confirm that unicode normalization is disabled.
2. Type the character "ê or "ô"
3. Observe NVDA's spoken output.

Actual behavior:

NVDA says "e normalizes" or "o normalizes." This also occurs if the character is are at the end of a word, Eg "robô" or "você." is read as "robo normalizes" and "voechay normalizes."

Expected behavior:

NVDA should only say "normalizes" for these unicode characters if:

• the user has enabled the "enable unicode normalization" option, as well as the "read normalized characters when reviewing by character" setting from NVDA's speech preferences.
• the user is reading character-by-character.

NVDA logs, crash dumps and other attachments:

System configuration

NVDA installed/portable/running from source:

installed

NVDA version:

2024.4.2

Windows version:

11

Name and version of other software in use when reproducing the issue:

Other information about your system:

Other questions

Does the issue still occur after restarting your computer?

yes

Have you tried any other versions of NVDA? If so, please report their behaviors.

no

If NVDA add-ons are disabled, is your problem still occurring?

yes

Does the issue still occur after you run the COM Registration Fixing Tool in NVDA's tools menu?

yes

[CyrilleB79]

@ObjectInSpace here NVDA is just reporting the character "combining circumflex" (char 770 = 0x302) as "normalizes". Looking for "normalizes" in the symbol dialog, you can modify its name as well as the minimal level to which it is reported.

This character has been introduced (or modified) in #11467 adding various math symbols.

Though I acknowledge that this symbol is not only used for maths so forcing it to a math definition and at the level "Never" may be confusing, especially now that we have a character normalization option.

Maybe at least renaming to "circumflex" would be useful? People doing maths / statistics will know what "circumflex" mean and will be able to change the name of the symbol if they prefer to have it reported as "normalizes".

What do you think?

[gerald-hartig]

Plan is to change the term from "normalizes" to "hat" as this is a colloquial term that is understood in spoken language and the various math domains that use the diacritic.

Are you reading predominantly Portuguese language text with a PT speech synthesiser? What would you expect to be read out?

[CyrilleB79]

"hat" seems quite nice.

Though, @ObjectInSpace could you indicate in which context you have encountered these words? In French, we usually have the letters with the diacritic combined in a single Unicode character, and I guess that it's the most common for languages using this diacritic. Thus, I wonder if the use of a separate combining diacritic character is not more commonly used in math where any letter can be diacriticized...

The answer to this question would help to decide which level should be associated to this character.

[ruifontes]

This problem was arised by an update in WhatsApp, from Windows store or the Beta version...
The only solution is to delete the line of symbols.dic where that character is described.
In other apps letters like "ê" or "ô" are correctly read.

[CyrilleB79]

@ruifontes wrote:

The only solution is to delete the line of symbols.dic where that character is described.

What about the following alternate solution: raise the level to which the symbol should be reported?