mirror of
https://github.com/irmen/prog8.git
synced 2024-11-27 03:50:27 +00:00
129 lines
4.4 KiB
Lua
129 lines
4.4 KiB
Lua
%import textio
|
|
|
|
main {
|
|
|
|
sub start() {
|
|
|
|
; cx16.set_rasterirq(irqhandler, 10)
|
|
uword adpcm_size = &adpcm_data_end - &adpcm_data
|
|
uword num_adpcm_blocks = adpcm_size / 256
|
|
uword nibblesptr = &adpcm_data
|
|
|
|
num_adpcm_blocks = 1
|
|
nibblesptr = &nibbles
|
|
|
|
repeat num_adpcm_blocks {
|
|
uword @zp sample = peekw(nibblesptr)
|
|
nibblesptr += 2
|
|
adpcm.init(sample, @(nibblesptr))
|
|
nibblesptr += 2
|
|
|
|
repeat 252 {
|
|
ubyte @zp nibble = @(nibblesptr)
|
|
adpcm.decode_nibble(nibble & 15)
|
|
txt.print_w(adpcm.predict as word)
|
|
txt.spc()
|
|
adpcm.decode_nibble(nibble>>4)
|
|
txt.print_w(adpcm.predict as word)
|
|
txt.spc()
|
|
nibblesptr++
|
|
}
|
|
}
|
|
|
|
repeat {
|
|
}
|
|
}
|
|
|
|
; sub irqhandler() {
|
|
; cx16.vpoke(1, $fa00, 255)
|
|
;
|
|
; uword @zp sample = peekw(nibbles)
|
|
; adpcm.init(sample, nibbles[2])
|
|
;
|
|
; ubyte @zp idx = 4
|
|
; while idx<100 {
|
|
; ubyte @zp nibble = nibbles[idx]
|
|
; adpcm.decode_nibble(nibble & 15)
|
|
; sample = adpcm.predict
|
|
; nibble >>= 4
|
|
; adpcm.decode_nibble(nibble)
|
|
; sample = adpcm.predict
|
|
; idx++
|
|
; }
|
|
;
|
|
; cx16.vpoke(1, $fa00,0)
|
|
; }
|
|
|
|
adpcm_data:
|
|
%asmbinary "adpcm-mono.bin"
|
|
adpcm_data_end:
|
|
|
|
ubyte[256] nibbles = [216, 227, 59, 0, 32, 2, 186, 139, 24, 34, 84, 115, 19, 185, 173, 8, 0, 56, 162,
|
|
169, 144, 113, 55, 130, 186, 12, 33, 144, 10, 136, 10, 129, 41, 66, 210, 207,
|
|
9, 0, 153, 129, 185, 173, 155, 0, 37, 162, 191, 27, 36, 129, 144, 137, 82, 18,
|
|
87, 18, 152, 153, 136, 0, 32, 23, 0, 40, 24, 117, 19, 202, 138, 8, 176, 171,
|
|
24, 131, 234, 9, 32, 39, 160, 10, 153, 33, 52, 209, 137, 200, 73, 33, 233, 155,
|
|
136, 66, 146, 8, 152, 56, 131, 113, 23, 168, 169, 185, 9, 160, 65, 37, 49, 22,
|
|
64, 68, 168, 138, 153, 1, 171, 88, 35, 25, 65, 1, 176, 173, 185, 175, 168, 154,
|
|
153, 205, 138, 144, 152, 154, 204, 43, 53, 51, 35, 154, 184, 96, 71, 34, 152,
|
|
172, 138, 160, 9, 24, 16, 204, 88, 48, 54, 160, 172, 152, 65, 132, 168, 48, 4,
|
|
0, 190, 9, 161, 155, 218, 139, 154, 185, 207, 9, 50, 130, 184, 252, 155, 24,
|
|
53, 144, 172, 170, 88, 51, 1, 176, 173, 17, 40, 115, 129, 136, 200, 11, 70,
|
|
129, 8, 0, 0, 48, 37, 160, 42, 35, 32, 17, 18, 220, 174, 170, 143, 9, 137, 160,
|
|
171, 104, 4, 169, 154, 219, 142, 65, 129, 128, 168, 9, 69, 50, 130, 219, 128,
|
|
138, 49, 52, 146, 237, 141, 50, 21, 8, 136, 8, 153, 35, 193, 170, 2, 169, 224,
|
|
186, 170, 40, 253, 12, 17, 25]
|
|
|
|
}
|
|
|
|
|
|
adpcm {
|
|
ubyte[] t_index = [ -1, -1, -1, -1, 2, 4, 6, 8, -1, -1, -1, -1, 2, 4, 6, 8]
|
|
uword[] t_step = [
|
|
7, 8, 9, 10, 11, 12, 13, 14,
|
|
16, 17, 19, 21, 23, 25, 28, 31,
|
|
34, 37, 41, 45, 50, 55, 60, 66,
|
|
73, 80, 88, 97, 107, 118, 130, 143,
|
|
157, 173, 190, 209, 230, 253, 279, 307,
|
|
337, 371, 408, 449, 494, 544, 598, 658,
|
|
724, 796, 876, 963, 1060, 1166, 1282, 1411,
|
|
1552, 1707, 1878, 2066, 2272, 2499, 2749, 3024,
|
|
3327, 3660, 4026, 4428, 4871, 5358, 5894, 6484,
|
|
7132, 7845, 8630, 9493, 10442, 11487, 12635, 13899,
|
|
15289, 16818, 18500, 20350, 22385, 24623, 27086, 29794,
|
|
32767]
|
|
|
|
uword @zp predict
|
|
ubyte @zp index
|
|
uword @zp pstep
|
|
|
|
sub init(uword startPredict, ubyte startIndex) {
|
|
predict = startPredict
|
|
index = startIndex
|
|
pstep = t_step[index]
|
|
}
|
|
|
|
sub decode_nibble(ubyte nibble) {
|
|
word @zp difference = 0
|
|
if nibble & 4
|
|
difference += pstep
|
|
pstep >>= 1
|
|
if nibble & 2
|
|
difference += pstep
|
|
pstep >>= 1
|
|
if nibble & 1
|
|
difference += pstep
|
|
pstep >>= 1
|
|
difference += pstep
|
|
if nibble & 8
|
|
difference = -difference
|
|
predict += difference as uword
|
|
index += t_index[nibble]
|
|
if index & 128
|
|
index = 0
|
|
else if index > 88
|
|
index = 88
|
|
pstep = t_step[index]
|
|
}
|
|
}
|