I’ve always been suspicious of getting something for nothing.

A free sample at the market? Fine. A stranger holding the door while my arms are full? Lovely.

But more music than was contained in the original digital file?

That sounds a bit like the fellow selling magic beans.

Take a standard 44.1kHz, 16-bit recording and upsample it to 96kHz, 24 bits. The new file is considerably larger and, in many cases, sounds better.

So where did the additional music come from?

It didn’t.

Upsampling can interpolate new data points between the old ones. If the original pattern is 1, 3, 5, 7, the math can make a good guess that 2, 4, and 6 belong in the gaps.

Useful, but that isn’t where most of the improvement comes from.

At the CD sampling rate, we need a steep filter to keep trouble from folding back into the audible band above 22kHz. Raise the sample rate to 96kHz and we can move that filter farther away and make its slope much gentler.

That gentler filter is, in my experience, responsible for much of what we hear as better sound.

Upsampling doesn’t recover missing music.

It gives the music we already have an easier path home.