XML DOCTYPE SINGING or equivalent, with best quality voice (on Linux)?

Not sure where to ask. So, I’ve just asked in the chat but thought I’d ask here as well:

Festival understands XML files that implement the “singing” DTD

<!DOCTYPE SINGING PUBLIC "-//SINGING//DTD SINGING mark up//EN" "Singing.v0_1.dtd" []>

However, kal_diphone is the only voice I’ve found that actually sings, and sadly, the intelligibility leaves a bit to be desired. I’ve been pledging a bit of money which makes my Mark 1’s speech much nicer. Can I get that voice onto a regular Linux box and will it sing using the SINGING doctype?

Or is there an alternative to make Linux sing with a better voice?

(And, by extension, eventually, the Mark 1 as well – though I’m more interested in making canned, synthetic singing WAV, OGG, or MP3 files?)