Re: put TTS off when the user is calling (using phone/whatsapp voice).
A mobile network call could in principle be detected (let's set aside permissions for a moment).
But detecting WhatsApp voice / Skype / VoIP ?
That is a valid concern, and one can extend that line of thinking to other sources (what if you are recording an audio with one of many apps, etc.).
You are right, it is probably impossible to cover
all interference situations.
But the phone call is the most obvious, and still most likely to happen (most frequent). Taking care of that would probably cover at least 80-90% of cases. And that is significant enough to justify doing that.
As my mentor once told me: "You cannot collect all marbles, you just do your best."