Addendum to make it more clear:
Audio1 is a high level audio API, which has a software implementation in Kha which is used if possible (that's the mixer).
Audio2 is a low level audio API which always talks to a system API as directly as possible. This uses Kore's audio API directly when you use a C++ target in Kha.

When you use a C++ target, Kore is used as Kha's backend for everything, including audio. Kha still does audio mixing itself, maybe that's what got you confused.

When you build a project in Kode Studio it tells you which version of Kha it uses in the first line of output

Mostly the same targets but they differ in some small things - but how a window/canvas is handled is one of the things which work differently.

If you only need it for the debug-html5 target, we can add that. Create a GitHub issue please so I don't forget about it.

No, it isn't and it's not planned. Canvas element resizing depends too much on how it is embedded in a website to put that functionality in Kha directly. Add a bit of JavaScript if you need it. Also the flash backend is purely in maintenance mode - I fix it when it breaks but I don't add new functionality.

You're running some very old Kha version then.

See http://api.kha.tech/kha/Window.html

Try building again please, just realized I hadn't yet merged rblsb's fixes.

The matrix is not quite right because the Kha VR support hasn't been updated for recent changes in Kore's VR support. That said, Lubos has demo code hidden somewhere (see http://forums.armory3d.org/t/vr-demo-dino/316) and you can find current Kore based VR things at https://github.com/CatCuddler/BodyTracking