Hi,

Based on NVDA source code, a common event NVDA will (or will not) handle is show/hide/UI tree changes. More specifically, NVDA does not handle many show window events as they can happen anytime anywhere. NVDA drops them conditionally (with some exceptions) because of performance considerations - it takes some time (at the level of microseconds) to handle events, involving looking up what the event is, which accessibility API is the event coming from, which object fired it, object construction to locate properties such as name and role, looping through global plugins and app modules (if any) to handle the event for this object, and destroying the object. Because show/hide/UI tree changes happen hundreds to thousands of times a second, it will overwhelm NVDA. So it is not really "gain focus" event that NVDA will handle internally - it is show window event.

Practically speaking, contrary to what people may think, gain focus event is not the event NVDA will respond the most with audio (sounds and speech included) - rather, it is "value change." You can observe this in action if you edit a text document, change sliders, choose combo box items, and so on. While it can be said that some value changes somewhat amount to focus changes, it is really values that are being changed. For example, if you do enable speak typed characters and type into Notepad, you will hear NVDA announce the typed character because it will handle value change event (there is a specific internal NVDA event called "typed character" that is responsible for announcing typed characters, but for simplicity, we can say that NVDA is responding to value changes).

Another reason why gain focus event is not the most common event that can result in sounds is internals of object navigation. When you move to the next/previous/parent/first child object, NVDA will notice that you are not moving focus but will announce where you are going. The event responsible for this is called "become navigator object" and is handled regardless of how you move to a different control - focus movement, object navigation, mouse movement, touch interaction, using Windows API to forcefully switch foreground windows (yes, there is at least one add-on that does this in response to user requests).

Just because an event can be run whenever the navigator object (and/or focus) changes does not guarantee that sounds will be played (speech and sounds). Why? I will answer in January 2023.

Cheers,

Joseph

Hi,

As an addendum: value is not the only property NVDA cares about - it will respond to other property changes such as name, state, description, location (sometimes), and API specific propertiess. While focus change may result in audio output, property changes do cause NVDA to visit event handling facility quite often.

As for my statement on the last post about answering something in January 2023: that's the subject of the next Inside Story post.

Cheers,

Joseph

This seems more like a developer discussion to me.

NVDA Developers Group

This group is focused on the development of the NVDA Screen Reader.  It is not an end-user support group.

• Post: nvda-devel@groups.io
• Subscribe: nvda-devel+subscribe@groups.io
• Unsubscribe: nvda-devel+unsubscribe@groups.io
• Group Owner: nvda-devel+owner@groups.io
• Help: nvda-devel+help@groups.io

Brian - Virginia, USA - Windows 10 Pro, 64-Bit, Version 22H2, Build 19045; Office 2016, Version 16.0.15726.20188, 32-bit

"Be Yourself" is the worst advice you can give to some people.

       ~ Tom Masson