diff --git a/lib/lualink.c b/lib/lualink.c
index 5032b943..192593f3 100644
--- a/lib/lualink.c
+++ b/lib/lualink.c
@@ -43,6 +43,7 @@
 #include "lua/calibrate.lua.h"
 #include "lua/sequins.lua.h"
 #include "lua/quote.lua.h"
+#include "lua/scale.lua.h"
 
 #include "build/ii_lualink.h" // generated C header for linking to lua
 
@@ -64,6 +65,7 @@ const struct lua_lib_locator Lua_libs[] =
     , { "lua_calibrate" , lua_calibrate , true}
     , { "lua_sequins"   , lua_sequins   , true}
     , { "lua_quote"     , lua_quote     , true}
+    , { "lua_scale"     , lua_scale     , true}
     , { NULL            , NULL          , true}
     };
 
diff --git a/lua/crowlib.lua b/lua/crowlib.lua
index 388d15c8..60ed8c86 100644
--- a/lua/crowlib.lua
+++ b/lua/crowlib.lua
@@ -14,6 +14,7 @@ public = dofile('lua/public.lua')
 clock  = dofile('lua/clock.lua')
 sequins= dofile('lua/sequins.lua')
 quote  = dofile('lua/quote.lua')
+qscale = dofile('lua/scale.lua')
 
 
 function C.reset()
diff --git a/lua/scale.lua b/lua/scale.lua
new file mode 100644
index 00000000..ba3f44d2
--- /dev/null
+++ b/lua/scale.lua
@@ -0,0 +1,34 @@
+--- scale quantizer generator library
+-- call the library similarly to input.scale or output.scale
+-- returns a function that will perform the quantization on an input
+
+local Q = {}
+
+Q.ji = function(rs, ins, outs)
+    -- create a 12TET-ified version of rs
+    return Q.__call(S, just12(rs), ins, outs)
+end
+
+local chrom = {0,1,2,3,4,5,6,7,8,9,10,11}
+Q.__call = function(self, ns, ins, outs)
+    -- defaults
+    ins, outs = ins or 12, outs or 1
+    -- chromatic shortcut with empty table or empty call
+    ns = (not ns or #ns == 0) and chrom or ns
+
+    -- optimize by precalculating constants
+    local _INS = 1/ins -- inverse of in-scaling (mul cheaper than div!)
+    local TET = 12 -- TODO configurable
+    local _TET = 1/TET -- TODO configurable
+    local OFF = 0.5 * ins / TET -- half an input-window of offset
+    local LEN = #ns -- memoize length
+
+    return function(n)
+        local norm = (n+OFF) * _INS -- normalize to input scaling
+        local octs = math.floor(norm) -- extract octaves
+        local ix   = math.floor((norm - octs) * LEN)
+        return (ns[ix+1]*_TET + octs) * outs -- scale lookup & reconstruct
+    end
+end
+
+return setmetatable(Q,Q)
diff --git a/tests/scale.lua b/tests/scale.lua
new file mode 100644
index 00000000..7ee6e4ad
--- /dev/null
+++ b/tests/scale.lua
@@ -0,0 +1,126 @@
+--- scale library test
+
+--- Just Intonation helpers
+-- convert a single fraction, or table of fractions to just intonation
+-- optional 'offset' is itself a just ratio
+-- justvolts converts to volts-per-octave
+-- just12 converts to 12TET representation (for *.scale libs)
+-- just12 will convert a fraction or table of fractions into 12tet 'semitones'
+function _justint(fn, f, off)
+    off = off and fn(off) or 0 -- optional offset is a just ratio
+    if type(f) == 'table' then
+        local t = {}
+        for k,v in ipairs(f) do
+            t[k] = fn(v) + off
+        end
+        return t
+    else -- assume number
+        return fn(f) + off
+    end
+end
+JIVOLT = 1 / math.log(2)
+JI12TET = 12 * JIVOLT
+function _jiv(f) return math.log(f) * JIVOLT end
+function _ji12(f) return math.log(f) * JI12TET end
+-- public functions
+function justvolts(f, off) return _justint(_jiv, f, off) end
+function just12(f, off) return _justint(_ji12, f, off) end
+function hztovolts(hz, ref)
+    ref = ref or 261.63 -- optional. defaults to middle-C
+    return justvolts(hz/ref)
+end
+
+qscale = dofile("../lua/scale.lua")
+
+
+--- major scale quantizer
+-- input[1].mode('scale',{0,2,4,7,9},12,1.0)
+-- output[1].scale({0,2,4,7,9},12,1.0)
+
+-- myquantizer = scale({0,2,4,7,9}, 12, 1.0) -- convert to volts
+-- myquantizerN = scale({0,2,4,7,9}, 12, 12) -- input/output mapping the same
+
+-- think of the 'divs'/'scaling' as input/output ranges
+-- so if divs == 12, then our table should be in 12TET
+-- if divs == 'ji', then our table should be in just fractions
+-- if scaling == divs, then output matches input
+-- if divs==12 & scaling==1 then convert note table to voltage
+--
+-- quantize to octaves, staying in 12TET
+local s1 = qscale({0},12,12)
+assert(type(s1) == 'function')
+assert(s1(0) == 0)
+assert(s1(7) == 0)
+assert(s1(11) == 0) -- always round down
+assert(s1(11.99) == 12) -- capture marginally beneath bounds
+assert(s1(12) == 12)
+assert(s1(-12) == -12) -- test negative numbers
+assert(s1(-11) == -12)
+assert(s1(-13) == -24)
+
+-- quantize to octaves, convert to volts
+local s1 = qscale({0},12,1)
+assert(s1(0) == 0)
+assert(s1(7) == 0)
+assert(s1(11) == 0) -- always round down
+assert(s1(11.99) == 1) -- capture marginally beneath bounds
+assert(s1(12) == 1)
+assert(s1(-12) == -1) -- test negative numbers
+assert(s1(-11) == -1)
+assert(s1(-13) == -2)
+
+-- quantize to 2 values, splitting the octave evenly
+local s1 = qscale({0,1},12,12)
+assert(s1(0) == 0)
+assert(s1(5) == 0) -- round down
+assert(s1(6) == 1) -- round up
+assert(s1(12) == 12) -- round up
+
+-- as above but confirm default in/out is 12,1
+local s1 = qscale({0,36})
+assert(s1(0) == 0)
+assert(s1(5) == 0) -- round down
+assert(s1(6) == 3) -- round up (note out-of-octave vals allowed)
+assert(s1(12) == 1) -- round up
+
+-- chromatic n->v scaler
+local s1 = qscale()
+assert(s1(0) == 0)
+assert(s1(1) == 1/12)
+assert(s1(11) == 11/12)
+assert(s1(12) == 1)
+
+-- chromatic n->v scaler: table-call syntax
+local s1 = qscale{}
+assert(s1(0) == 0)
+assert(s1(1) == 1/12)
+assert(s1(11) == 11/12)
+assert(s1(12) == 1)
+
+-- chromatic n->n scaler
+local s1 = qscale({}, 12, 12)
+assert(s1(0) == 0)
+assert(s1(1) == 1)
+assert(s1(11) == 11)
+assert(s1(12) == 12)
+
+-- just intonation support
+-- use a separate function
+local sj1 = qscale.ji({1/1, 2/1}, 12, 1)
+assert(sj1(0) == 0)
+assert(sj1(5) == 0) -- round down
+-- note we need to account for floating point inaccuracies here
+assert(sj1(6) >= 0.9999 and sj1(6) <= 1.0001) -- round up
+assert(sj1(12) == 1.0) -- round up
+
+-- input as voltage, convert to 12TET
+local sv1 = qscale({0,7},1.0,12)
+assert(sv1(0) == 0)
+assert(sv1(0.5) >= 6.9999 and sv1(0.5) <= 7.0001) -- round up
+assert(sv1(1) == 12)
+
+-- input as voltage, output to voltage
+local sv1 = qscale({0,7},1.0,1.0)
+assert(sv1(0) == 0)
+assert(sv1(0.5) >= 0.583 and sv1(0.5) <= 0.584) -- round up
+assert(sv1(1) == 1)