spacetime.lua

-- spacetime
-- norns study 3
--
-- ENC 1 - sweep filter
-- ENC 2 - select edit position
-- ENC 3 - choose command
-- KEY 3 - randomize command set
--
-- spacetime is a weird function sequencer.
-- it plays a note on each step.
-- each step is a symbol for the action.
-- + = increase note
-- - = decrease note
-- < = go to bottom note
-- > = go to top note
-- * = random note
-- M = fast metro
-- m = slow metro
-- # = jump random position
--
-- augment/change this script with new functions!

engine.name = "PolyPerc"

note = 40
position = 1
steps = {1,1,1,1 ,1,1,1,1 ,1,1,1,1 ,1,1,1,1}
steps_len = #steps
edit = 1

-- helpers to index into commands table
label = 1
act   = 2
commands =
{ { "+", function inc() note = util.clamp(note + 5, 40, 120) end }
, { "-", function dec() note = util.clamp(note - 5, 40, 120) end }
, { "<", function bottom() note = 40 end }
, { ">", function top() note = 120 end }
, { "*", function rand() note = math.random(80) + 40 end }
, { "M", function metrofast() counter.time = 0.125 end }
, { "m", function metroslow() counter.time = 0.25 end }
, { "#", function positionrand() position = math.random(steps_len) end }
}
commands_len = #commands

function init()
  params:add_control("cutoff",controlspec.new(50,5000,'exp',0,555,'hz'))
  params:set_action("cutoff", function(x) engine.cutoff(x) end)
  counter = metro.alloc(count, 0.125, -1)
  counter:start()
end

function count()
  position = (position % steps_len) + 1
  commands[steps[position]].act()
  engine.hz(midi_to_hz(note))
  redraw()
end

function redraw()
  screen.clear()
  for i=1,steps_len do
    screen.level((i == edit) and 15 or 2)
    screen.move(i*8-8,40)
    screen.text(commands[steps[i]].label)
    if i == position then
      screen.move(i*8-8, 45)
      screen.line_rel(6,0)
      screen.stroke()
    end
  end
  screen.update()
end

function enc(n,d)
  if n == 1 then
    params:delta("cutoff",d)
  elseif n == 2 then
    edit = util.clamp(edit + d, 1, steps_len)
  elseif n == 3 then
    steps[edit] = util.clamp(steps[edit]+d, 1, commands_len)
  end
  redraw()
end

function key(n,z)
  if n==3 and z==1 then
    randomize_steps()
  end
end

function midi_to_hz(note)
  return (440 / 32) * (2 ^ ((note - 9) / 12))
end

function randomize_steps()
  for i=1,steps_len do
    steps[i] = math.random(commands_len)
  end
end

real minor stuff:

• the # operator in lua iterates over the table, it's not efficient. if the table is constant size, i'd just store the size once.

• i'm little confused by the all-caps variable names. convention would imply these are constants. but COMMANDS is a constant and STEPS isn't.

Just a note for people coming here who hadn't seen this before (and I'm thinking this is how this works in lua, it's how it works in JS)

screen.level((i == edit) and 15 or 2) makes use of programming techniques known as "take right" and "take left". When many items are and'ed together, the right-most term that is truthy is given. So if i == 2 is evaluated to be true, the next term's truthiness is evaluated. Since non-zero numbers are truth-y in lua, 15 is returned.

"or" works similarly but the left most truth-y item is returned. you can think of this expression like this: ((i == 2) and 15) or 2. If ((i == 2) and 15) returns false, the truthiness of the next value is checked. And since non-zero numbers are truthy as mentioned above, 2 is returned.

@catfact
thanks for the pointer on #. had no idea about the performance there. excuse the globals everywhere too (was trying not to refactor the whole study).

point taken on the all-caps. it was a holdover from the study which used allcaps to identify the table-lengths. i just reused them and didn't really think about it. updated now to all lowercase with the idea that they don't need to be constant, and there's nothing stopping someone making the command list change at runtime.

speaking of convention, i just read the LuaStyleGuide again which covers both the allcaps constants, but also the idiom explained by @jlmitch5. worth the 10minute read!