It isn't. MIDI variable-length quantities don't include an indication of length in a header, and the last byte in any multi-byte sequence is a valid single-byte value. It is more bit efficient than UTF-8, but I believe it lacks other properties around synchronization, possibly others.

I'm less sure here, but I believe the MIDI version is more similar to the thing the X/Open folks were considering before UTF-8 supplanted it.

On Feb 7, 2016, at 20:09, Bruce Ellis <bruce.ellis@gmail.com> wrote:

Yes. But the midi version is utf-8.

brucee

On 7 February 2016 at 16:42, Skip Tavakkolian <skip.tavakkolian@gmail.com> wrote:
but following that line of reasoning, aren't they all specialized versions of Huffman encoding?

On Sat, Feb 6, 2016 at 6:04 PM Bruce Ellis <bruce.ellis@gmail.com> wrote:
Not to burst a balloon but check out variable length ints in the Midi File Format for utf-8 in the early 80s.

brucee