Except a novelization of a movie is entirely different than the movie itself. Why is this any different than doing text-to-speech of an ebook? As long as this i

Except a novelization of a movie is entirely different than the movie itself. Why is this any different than doing text-to-speech of an ebook? As long as this is done on-the-fly and the text isn't stored afterwards, what difference does it make in how the original piece of media is consumed? Amazon isn't selling either form of transcription. What if someone made a device for the deaf that would transcribe spoken word to text? Would it be fair for the manufacturer to be sued for possibly doing speech-to-text on some copyrighted recording?