Brian is right - one of the first things to try in a "this isn't being read properly" situation is to try it with a different synthesizer.  In this case, OneCore isn't reading any difference between e with an accent é and e.  But you found that SAPI5 DOES read it differently, and I just tested and eSpeak-NG does as well.

Another alternative workaround would be to edit the punctuation / symbol pronunciation list and add an entry for é to read "e acute" (or however you want it read.  

On Wed, Feb 16, 2022 at 1:22 AM Brian Vogel <britechguy@...> wrote:

On Tue, Feb 15, 2022 at 08:04 AM, Suhas D wrote:

Then I changed the synth to SAPI5 and, NVDA started reading out the accented characters.

-
And, again, there you have it (and not just for NVDA, either).

Screen readers don't have anything to do with "the actual reading part," but pass off strings (be they single characters or longer) to a synth and the synth pronounces them.

If a particular letter, word, phrase is not read correctly by a given synth, contacting the maker of that particular synth and registering your displeasure is the way to go.

The only way I can think of to force change this with a voice dictionary (preferable to default dictionary) would be to enter the single character as "whole word" and play around with whatever replacement string you can find that ends up saying what you want that letter to sound like.  If you used the default dictionary the change you made would carry over to every synth and likely screw up the correct pronunciation by those synths that already do so.
--

Brian - Windows 10, 64-Bit, Version 21H2, Build 19044

Under certain circumstances, profanity provides a relief denied even to prayer.

        ~ Mark Twain

--

Quentin Christensen
Training and Support Manager

Web: www.nvaccess.org 

Training: https://www.nvaccess.org/shop/

Certification: https://certification.nvaccess.org/

User group: https://nvda.groups.io/g/nvda

Facebook: http://www.facebook.com/NVAccess 
Twitter: @NVAccess