I'm particularly keen on promoting this [Nikolai] Medtner piece, Night Wind, because I feel it is a fantastic work that hasn't been given its due yet. Admittedly, it's difficult for both the performer and the listener. The work is quite dense, comparatively long, as these things go, and demands rather active listening. But it's gripping, and I believe it should be heard much more than it has been.

The process of putting intangible thoughts into an imperfect system of notation - which is difficult enough, depending on your ideas - acquaints you with how best to express your ideas so that it is as clear as possible to the performer.

I think one of the reasons [Nikolai] Medtner hasn't had a chance is that his music needs very, very committed performances. If you play his works passively, the juice of his music is really not going to be extracted - it's simply not going to come out.

My father ran the gamut: a favorite of his was Josef Hofmann, but there was also Friedman, Rachmaninoff, Godowsky, Lhevinne, Moiseiwitsch, Paderewski, etc.

I don't listen to recordings very much now, to be perfectly honest. I listened to them a lot when I was younger.

That is my way of doing things, and I wouldn't necessarily recommend this to anybody else; if you need to do technical exercises, you do them. The whole point of practicing is to get to know yourself, to know your weaknesses and to zero in on them and target them. It's not really about employing anybody else's formulas, because you really have to find what is best for you and what you need.

Very rarely do I work on mechanics now. I tend to solve problems within the music itself.