eukara/ngunix
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
ngunix/tracker/it/readmod.c
eukos bdddd7aba2 Major restructure for the future
2015-08-07 12:06:39 +02:00

604 lines
16 KiB
C
Raw Blame History

/* _______ ____ __ ___ ___
* \ _ \ \ / \ / \ \ / / ' ' '
* | | \ \ | | || | \/ | . .
* | | | | | | || ||\ /| |
* | | | | | | || || \/ | | ' ' '
* | | | | | | || || | | . .
* | |_/ / \ \__// || | |
* /_______/ynamic \____/niversal /__\ /____\usic /| . . ibliotheque
* / \
* / . \
* readmod.c - Code to read a good old-fashioned / / \ \
* Amiga module from an open file. | < / \_
* | \/ /\ /
* By entheh. \_ / > /
* | \ / /
* | ' /
* \__/
*/
#include <stdlib.h>
#include <string.h>
#include <math.h>
#include "dumb.h"
#include "internal/it.h"
static int it_mod_read_pattern(IT_PATTERN *pattern, DUMBFILE *f, int n_channels, unsigned char *buffer)
{
int pos;
int channel;
int row;
IT_ENTRY *entry;
pattern->n_rows = 64;
if (n_channels == 0) {
/* Read the first four channels, leaving gaps for the rest. */
for (pos = 0; pos < 64*8*4; pos += 8*4)
dumbfile_getnc(buffer + pos, 4*4, f);
/* Read the other channels into the gaps we left. */
for (pos = 4*4; pos < 64*8*4; pos += 8*4)
dumbfile_getnc(buffer + pos, 4*4, f);
n_channels = 8;
} else
dumbfile_getnc(buffer, 64 * n_channels * 4, f);
if (dumbfile_error(f))
return -1;
/* compute number of entries */
pattern->n_entries = 64; /* Account for the row end markers */
pos = 0;
for (row = 0; row < 64; row++) {
for (channel = 0; channel < n_channels; channel++) {
if (buffer[pos+0] | buffer[pos+1] | buffer[pos+2] | buffer[pos+3])
pattern->n_entries++;
pos += 4;
}
}
pattern->entry = malloc(pattern->n_entries * sizeof(*pattern->entry));
if (!pattern->entry)
return -1;
entry = pattern->entry;
pos = 0;
for (row = 0; row < 64; row++) {
for (channel = 0; channel < n_channels; channel++) {
if (buffer[pos+0] | buffer[pos+1] | buffer[pos+2] | buffer[pos+3]) {
unsigned char sample = (buffer[pos+0] & 0xF0) | (buffer[pos+2] >> 4);
int period = ((int)(buffer[pos+0] & 0x0F) << 8) | buffer[pos+1];
entry->channel = channel;
entry->mask = 0;
if (period) {
int note;
entry->mask |= IT_ENTRY_NOTE;
/* frequency = (AMIGA_DIVISOR / 8) / (period * 2)
* C-1: period = 214 -> frequency = 16726
* so, set C5_speed to 16726
* and period = 214 should translate to C5 aka 60
* halve the period, go up an octive
*
* period = 214 / pow(DUMB_SEMITONE_BASE, note - 60)
* pow(DUMB_SEMITONE_BASE, note - 60) = 214 / period
* note - 60 = log(214/period) / log(DUMB_SEMITONE_BASE)
*/
note = (int)floor(log(214.0/period) / log(DUMB_SEMITONE_BASE) + 60.5);
entry->note = MID(0, note, 119);
// or should we preserve the period?
//entry->note = buffer[pos+0] & 0x0F; /* High nibble */
//entry->volpan = buffer[pos+1]; /* Low byte */
// and what about finetune?
}
if (sample) {
entry->mask |= IT_ENTRY_INSTRUMENT;
entry->instrument = sample;
}
_dumb_it_xm_convert_effect(buffer[pos+2] & 0x0F, buffer[pos+3], entry);
entry++;
}
pos += 4;
}
IT_SET_END_ROW(entry);
entry++;
}
return 0;
}
static int it_mod_read_sample_header(IT_SAMPLE *sample, DUMBFILE *f)
{
int finetune;
/**
21 22 Chars Sample 1 name. If the name is not a full
22 chars in length, it will be null
terminated.
If
the sample name begins with a '#' character (ASCII $23 (35)) then this is
assumed not to be an instrument name, and is probably a message.
*/
dumbfile_getnc(sample->name, 22, f);
sample->name[22] = 0;
sample->filename[0] = 0;
sample->length = dumbfile_mgetw(f) << 1;
finetune = (signed char)(dumbfile_getc(f) << 4) >> 4; /* signed nibble */
/** Each finetune step changes the note 1/8th of a semitone. */
sample->global_volume = 64;
sample->default_volume = dumbfile_getc(f); // Should we be setting global_volume to this instead?
sample->loop_start = dumbfile_mgetw(f) << 1;
sample->loop_end = sample->loop_start + (dumbfile_mgetw(f) << 1);
/**
Once this sample has been played completely from beginning
to end, if the repeat length (next field) is greater than two bytes it
will loop back to this position in the sample and continue playing. Once
it has played for the repeat length, it continues to loop back to the
repeat start offset. This means the sample continues playing until it is
told to stop.
*/
if (sample->length <= 0) {
sample->flags = 0;
return 0;
}
sample->flags = IT_SAMPLE_EXISTS;
sample->default_pan = 0;
sample->C5_speed = (long)(16726.0*pow(DUMB_PITCH_BASE, finetune*32));
// the above line might be wrong
if (sample->loop_end > sample->length)
sample->loop_end = sample->length;
if (sample->loop_end - sample->loop_start > 2)
sample->flags |= IT_SAMPLE_LOOP;
sample->vibrato_speed = 0;
sample->vibrato_depth = 0;
sample->vibrato_rate = 0;
sample->vibrato_waveform = 0; // do we have to set _all_ these?
return dumbfile_error(f);
}
static int it_mod_read_sample_data(IT_SAMPLE *sample, DUMBFILE *f)
{
long i;
long truncated_size;
/* let's get rid of the sample data coming after the end of the loop */
if ((sample->flags & IT_SAMPLE_LOOP) && sample->loop_end < sample->length) {
truncated_size = sample->length - sample->loop_end;
sample->length = sample->loop_end;
} else {
truncated_size = 0;
}
if (sample->length) {
sample->data = malloc(sample->length);
if (!sample->data)
return -1;
/* Sample data are stored in "8-bit two's compliment format" (sic). */
for (i = 0; i < sample->length; i++)
((signed char *)sample->data)[i] = dumbfile_getc(f);
} else
sample->flags &= ~IT_SAMPLE_EXISTS;
/* skip truncated data */
dumbfile_skip(f, truncated_size);
// Should we be truncating it?
if (dumbfile_error(f))
return -1;
return 0;
}
typedef struct BUFFERED_MOD BUFFERED_MOD;
struct BUFFERED_MOD
{
unsigned char *buffered;
long ptr, len;
DUMBFILE *remaining;
};
static int buffer_mod_skip(void *f, long n)
{
BUFFERED_MOD *bm = f;
if (bm->buffered) {
bm->ptr += n;
if (bm->ptr >= bm->len) {
free(bm->buffered);
bm->buffered = NULL;
return dumbfile_skip(bm->remaining, bm->ptr - bm->len);
}
return 0;
}
return dumbfile_skip(bm->remaining, n);
}
static int buffer_mod_getc(void *f)
{
BUFFERED_MOD *bm = f;
if (bm->buffered) {
int rv = bm->buffered[bm->ptr++];
if (bm->ptr >= bm->len) {
free(bm->buffered);
bm->buffered = NULL;
}
return rv;
}
return dumbfile_getc(bm->remaining);
}
static long buffer_mod_getnc(char *ptr, long n, void *f)
{
BUFFERED_MOD *bm = f;
if (bm->buffered) {
int left = bm->len - bm->ptr;
if (n >= left) {
int rv;
memcpy(ptr, bm->buffered + bm->ptr, left);
free(bm->buffered);
bm->buffered = NULL;
rv = dumbfile_getnc(ptr + left, n - left, bm->remaining);
return left + MAX(rv, 0);
}
memcpy(ptr, bm->buffered + bm->ptr, n);
bm->ptr += n;
return n;
}
return dumbfile_getnc(ptr, n, bm->remaining);
}
static void buffer_mod_close(void *f)
{
BUFFERED_MOD *bm = f;
if (bm->buffered) free(bm->buffered);
/* Do NOT close bm->remaining */
free(f);
}
DUMBFILE_SYSTEM buffer_mod_dfs = {
NULL,
&buffer_mod_skip,
&buffer_mod_getc,
&buffer_mod_getnc,
&buffer_mod_close
};
#define MOD_FFT_OFFSET (20 + 31*(22+2+1+1+2+2) + 1 + 1 + 128)
static DUMBFILE *dumbfile_buffer_mod(DUMBFILE *f, unsigned long *fft)
{
BUFFERED_MOD *bm = malloc(sizeof(*bm));
if (!bm) return NULL;
bm->buffered = malloc(MOD_FFT_OFFSET + 4);
if (!bm->buffered) {
free(bm);
return NULL;
}
bm->len = dumbfile_getnc(bm->buffered, MOD_FFT_OFFSET + 4, f);
if (bm->len > 0) {
if (bm->len >= MOD_FFT_OFFSET + 4)
*fft = (unsigned long)bm->buffered[MOD_FFT_OFFSET ] << 24
| (unsigned long)bm->buffered[MOD_FFT_OFFSET+1] << 16
| (unsigned long)bm->buffered[MOD_FFT_OFFSET+2] << 8
| (unsigned long)bm->buffered[MOD_FFT_OFFSET+3];
else
*fft = 0;
bm->ptr = 0;
} else {
free(bm->buffered);
bm->buffered = NULL;
}
bm->remaining = f;
return dumbfile_open_ex(bm, &buffer_mod_dfs);
}
static DUMB_IT_SIGDATA *it_mod_load_sigdata(DUMBFILE *f)
{
DUMB_IT_SIGDATA *sigdata;
int n_channels;
int i;
unsigned long fft;
f = dumbfile_buffer_mod(f, &fft);
if (!f)
return NULL;
sigdata = malloc(sizeof(*sigdata));
if (!sigdata) {
dumbfile_close(f);
return NULL;
}
/**
1 20 Chars Title of the song. If the title is not a
full 20 chars in length, it will be null-
terminated.
*/
if (dumbfile_getnc(sigdata->name, 20, f) < 20) {
free(sigdata);
dumbfile_close(f);
return NULL;
}
sigdata->name[20] = 0;
sigdata->n_samples = 31;
switch (fft) {
case DUMB_ID('M','.','K','.'):
case DUMB_ID('M','!','K','!'):
case DUMB_ID('M','&','K','!'):
case DUMB_ID('N','.','T','.'):
case DUMB_ID('F','L','T','4'):
n_channels = 4;
break;
case DUMB_ID('F','L','T','8'):
n_channels = 0;
/* 0 indicates a special case; two four-channel patterns must be
* combined into one eight-channel pattern. Pattern indexes must
* be halved. Why oh why do they obfuscate so?
*/
for (i = 0; i < 128; i++)
sigdata->order[i] >>= 1;
break;
case DUMB_ID('C','D','8','1'):
case DUMB_ID('O','C','T','A'):
case DUMB_ID('O','K','T','A'):
n_channels = 8;
break;
case DUMB_ID('1','6','C','N'):
n_channels = 16;
break;
case DUMB_ID('3','2','C','N'):
n_channels = 32;
break;
default:
/* If we get an illegal tag, assume 4 channels 15 samples. */
if ((fft & 0x0000FFFFL) == DUMB_ID(0,0,'C','H')) {
if (fft >= '1' << 24 && fft < '4' << 24) {
n_channels = ((fft & 0x00FF0000L) >> 16) - '0';
if ((unsigned int)n_channels >= 10) {
/* Rightmost character wasn't a digit. */
n_channels = 4;
sigdata->n_samples = 15;
} else {
n_channels += (((fft & 0xFF000000L) >> 24) - '0') * 10;
/* MODs should really only go up to 32 channels, but we're lenient. */
if ((unsigned int)(n_channels - 1) >= DUMB_IT_N_CHANNELS - 1) {
/* No channels or too many? Can't be right... */
n_channels = 4;
sigdata->n_samples = 15;
}
}
} else {
n_channels = 4;
sigdata->n_samples = 15;
}
} else if ((fft & 0x00FFFFFFL) == DUMB_ID(0,'C','H','N')) {
n_channels = (fft >> 24) - '0';
if ((unsigned int)(n_channels - 1) >= 9) {
/* Character was '0' or it wasn't a digit */
n_channels = 4;
sigdata->n_samples = 15;
}
} else if ((fft & 0xFFFFFF00L) == DUMB_ID('T','D','Z',0)) {
n_channels = (fft & 0x000000FFL) - '0';
if ((unsigned int)(n_channels - 1) >= 9) {
/* We've been very lenient, given that it should have
* been 1, 2 or 3, but this MOD has been very naughty and
* must be punished.
*/
n_channels = 4;
sigdata->n_samples = 15;
}
} else {
n_channels = 4;
sigdata->n_samples = 15;
}
}
sigdata->sample = malloc(sigdata->n_samples * sizeof(*sigdata->sample));
if (!sigdata->sample) {
free(sigdata);
dumbfile_close(f);
return NULL;
}
sigdata->song_message = NULL;
sigdata->order = NULL;
sigdata->instrument = NULL;
sigdata->pattern = NULL;
sigdata->midi = NULL;
sigdata->checkpoint = NULL;
sigdata->n_instruments = 0;
for (i = 0; i < sigdata->n_samples; i++)
sigdata->sample[i].data = NULL;
for (i = 0; i < sigdata->n_samples; i++) {
if (it_mod_read_sample_header(&sigdata->sample[i], f)) {
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
}
sigdata->n_orders = dumbfile_getc(f);
sigdata->restart_position = dumbfile_getc(f);
// what if this is >= 127? what about with Fast Tracker II?
if (sigdata->n_orders <= 0 || sigdata->n_orders > 128) { // is this right?
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
//if (sigdata->restart_position >= sigdata->n_orders)
//sigdata->restart_position = 0;
sigdata->order = malloc(128); /* We may need to scan the extra ones! */
if (!sigdata->order) {
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
if (dumbfile_getnc(sigdata->order, 128, f) < 128) {
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
/* "The old NST format contains only 15 samples (instead of 31). Further
* it doesn't contain a file format tag (id). So Pattern data offset is
* at 20+15*30+1+1+128."
* - Then I shall assume the File Format Tag never exists if there are
* only 15 samples. I hope this isn't a faulty assumption...
*/
if (sigdata->n_samples == 31)
dumbfile_skip(f, 4);
/* Work out how many patterns there are. */
sigdata->n_patterns = -1;
for (i = 0; i < 128; i++)
if (sigdata->n_patterns < sigdata->order[i])
sigdata->n_patterns = sigdata->order[i];
sigdata->n_patterns++;
/* May as well try to save a tiny bit of memory. */
if (sigdata->n_orders < 128) {
unsigned char *order = realloc(sigdata->order, sigdata->n_orders);
if (order) sigdata->order = order;
}
sigdata->pattern = malloc(sigdata->n_patterns * sizeof(*sigdata->pattern));
if (!sigdata->pattern) {
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
for (i = 0; i < sigdata->n_patterns; i++)
sigdata->pattern[i].entry = NULL;
/* Read in the patterns */
{
unsigned char *buffer = malloc(256 * n_channels); /* 64 rows * 4 bytes */
if (!buffer) {
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
for (i = 0; i < sigdata->n_patterns; i++) {
if (it_mod_read_pattern(&sigdata->pattern[i], f, n_channels, buffer) != 0) {
free(buffer);
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
}
free(buffer);
}
/* And finally, the sample data */
for (i = 0; i < sigdata->n_samples; i++) {
if (it_mod_read_sample_data(&sigdata->sample[i], f)) {
_dumb_it_unload_sigdata(sigdata);
dumbfile_close(f);
return NULL;
}
}
dumbfile_close(f); /* Destroy the BUFFERED_MOD DUMBFILE we were using. */
/* The DUMBFILE originally passed to our function is intact. */
/* Now let's initialise the remaining variables, and we're done! */
sigdata->flags = IT_WAS_AN_XM | IT_WAS_A_MOD | IT_OLD_EFFECTS | IT_COMPATIBLE_GXX | IT_STEREO;
sigdata->global_volume = 128;
sigdata->mixing_volume = 48;
/* We want 50 ticks per second; 50/6 row advances per second;
* 50*10=500 row advances per minute; 500/4=125 beats per minute.
*/
sigdata->speed = 6;
sigdata->tempo = 125;
sigdata->pan_separation = 128;
memset(sigdata->channel_volume, 64, DUMB_IT_N_CHANNELS);
for (i = 0; i < DUMB_IT_N_CHANNELS; i += 4) {
sigdata->channel_pan[i+0] = 16;
sigdata->channel_pan[i+1] = 48;
sigdata->channel_pan[i+2] = 48;
sigdata->channel_pan[i+3] = 16;
}
_dumb_it_fix_invalid_orders(sigdata);
return sigdata;
}
DUH *dumb_read_mod_quick(DUMBFILE *f)
{
sigdata_t *sigdata;
DUH_SIGTYPE_DESC *descptr = &_dumb_sigtype_it;
sigdata = it_mod_load_sigdata(f);
if (!sigdata)
return NULL;
{
const char *tag[1][2];
tag[0][0] = "TITLE";
tag[0][1] = ((DUMB_IT_SIGDATA *)sigdata)->name;
return make_duh(-1, 1, (const char *const (*)[2])tag, 1, &descptr, &sigdata);
}
}
Reference in a new issue View git blame Copy permalink