When NVDA with MathCAT is reading aloud a text phrase that contains multiple in-line Office Math expressions as the result of a keyboard command/navigation, the content between the first Office Math object up-to and including the last Office Math expression are omitted. However, when NVDA speaks the same content as the result of a a mouse trigger (hover), it speaks that content correctly.
Note: This issue does not occur with the JAWS screen reader or when speaking the same content in a web page (HTML + MathML).
Steps to Reproduce:
Launch the NVDA screen reader and ensure it is configured to use the NVDA MathCAT plug-in (see: Recommended NVDA settings)
Restart NVDA if necessary to activate the MathCAT plug-in.
Launch Microsoft Word 365 (desktop app) and start a new blank document
Make sure the mouse pointer is not over the MS Word window
Type the sentence: The expressions a=b and y=3s are examples of equations.
Convert a=b into an Office Math expression by selecting it, pressing Alt+=, and then pressing Ctrl+=
Arrow over to the y=3s expression
Convert y=3s into an Office Math expression by selecting it, pressing Alt+=, and then pressing Ctrl+=
Move to the start of the document and then use the up/down arrow keys as needed to have NVDA read aloud the sentence you just entered
Observe that the sentence is spoken as
The expressions a is equal to b are examples of equations.
instead of as
The expressions a is equal to b and y equals 3 s are examples of equations.
Now use the mouse to hover and dwell over the same sentence in the MS Word document
Observe that NVDA speaks the sentence correctly upon mouse hover
Expected Behavior:
When NVDA is speaking content in a MS Word document as the result of a keyboard command/navigation, none of the content should be skipped regardless of how many in-line Office Math objects are in a text phrase.
Observed Behavior:
When NVDA is speaking content in a MS Word document as the result of a keyboard command/navigation, content after the first Office Math object up-to and including the last Office Math object is skipped over when multiple in-line Office Math objects are in a text phrase.
Version Information:
OS Version:
-- OS Name: Microsoft Windows 11 Enterprise
-- Version: 10.0.22621 Build 22621
-- Locale: United States
Microsoft Word Version:
-- Microsoft 365 Apps for enterprise
-- Version 2402 (Build 17328.20550 Click-to-Run)
NVDA Version and Settings
-- NVDA Version: 2023.3.4 (2023.3.4.31265)
-- MathCAT Version: 0.3.9
-- Configured to use the MathCAT plug-in as outlined in: Recommended NVDA settings
Example Video and Sample Document:
We have provided a link to a short example video demonstrating the issue, and attached the Word document used in the video.
Description:
When NVDA with MathCAT is reading aloud a text phrase that contains multiple in-line Office Math expressions as the result of a keyboard command/navigation, the content between the first Office Math object up-to and including the last Office Math expression are omitted. However, when NVDA speaks the same content as the result of a a mouse trigger (hover), it speaks that content correctly.
Note: This issue does not occur with the JAWS screen reader or when speaking the same content in a web page (HTML + MathML).
Steps to Reproduce:
The expressions a is equal to b are examples of equations.
instead of asThe expressions a is equal to b and y equals 3 s are examples of equations.
Expected Behavior:
When NVDA is speaking content in a MS Word document as the result of a keyboard command/navigation, none of the content should be skipped regardless of how many in-line Office Math objects are in a text phrase.
Observed Behavior:
When NVDA is speaking content in a MS Word document as the result of a keyboard command/navigation, content after the first Office Math object up-to and including the last Office Math object is skipped over when multiple in-line Office Math objects are in a text phrase.
Version Information:
Example Video and Sample Document:
We have provided a link to a short example video demonstrating the issue, and attached the Word document used in the video.
Attachments:
Credits
Thanks to Neil Soiffer, Talking Cat Software for reporting this issue.