Singing the Portal Theme

Quite a few years ago I posted something about creating music on the microbit and used the Portal ending theme as written by Jonathan Coulton.

This was some code that used the microbit's music DSL to play the theme using music.play and ran through an array of notes representing that theme. I just came across an alternative version where I have used speech.sing instead to work through another simple DSL I created to sing the words at the correct pitch and with the correct timing. I paired invidiual words with note names and octave numbers, like in the previous example, but for sing to work it requires frequency values rather than note names. I added a "rest" command to make sure that pauses in the music were handled.

from microbit import *
import speech

note_name_to_value = {
	"c3"  : "58",
	"cs3" : "55",
	"d3"  : "52",
	"ds3" : "49",
	"e3"  : "46",
	"f3"  : "44",
	"fs3" : "42",
	"g3"  : "39",
	"gs3" : "37",
	"a3"  : "35",
	"as3" : "33",
	"b3"  : "31",
	"c4"  : "29",
	"cs4" : "28",
	"d4"  : "26",
	"ds4" : "25",
	"e4"  : "23",
	"f4"  : "22",
	"fs4" : "21",
	"g4"  : "20"
}

song = [
	"This:g4",
	"was:fs4",
	"a:e4",
	"try:e4",
	"umph:fs4"
	"rest:1000"
	"I'm:a3"
	"making:g4",
	"a:fs4",
	"note:e4",
	"here:fs4",
	"rest:500",
	"huge:d4",
	"rest:200",
	"suck:e4",
	"sess:a3"
	"rest:1000",
	"it's:a3",
	"hard:e4",
	"to:fs4",
	"oh:g4",
	"ver:e4"
	"state:cs4"
	"my:d4",
	"sat:e4",
	"tis:a3",
	"fack:a3",
	"shun:fs4"
]

while True:
	for chunk in song:
		word, value = chunk.split(":") 
		
		if word == "rest":
			sleep(int(value))
		else:
			phoneme = speech.translate(word)
			speech.sing("#" + note_name_to_value[value] + phoneme)

Words need to be translated into phonemes before singing so that they are in the correct format for the sing API.

This is incomplete - mainly because it took a lot of time to get even this far - but I like the approach, it got quite good results and it might be something I try again in the future.