FWIW, I'm pretty certain that video sound is down to two things.

One, as already mentioned, it's a camera phone, and

Two, there's probably no tweaking been done on that keyboard - no matter how good any keyboard sounds OOTB, they all benefit from some amount of tweaking.

Plus, while I can see the intent of demonstrating it using its own speakers is to attract the 'home buyer' - I hate that term - anyone who would be using it on stage most probably wouldn't be relying on its built-in speakers, anyway.

Roland has been making quality gear for long enough that I can't believe they'd put out anything that initially sounds as substandard as that.