tumbledemerald-legacy/tools/aif2pcm/extended.c
2022-05-19 17:14:13 +00:00

173 lines
4.2 KiB
C

/* $Id: extended.c,v 1.8 2006/12/23 11:17:49 toad32767 Exp $ */
/*-
* Copyright (c) 2005, 2006 by Marco Trillo <marcotrillo@gmail.com>
*
* Permission is hereby granted, free of charge, to any
* person obtaining a copy of this software and associated
* documentation files (the "Software"), to deal in the
* Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of the
* Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice
* shall be included in all copies or substantial portions of
* the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY
* KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
* WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
* PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS
* OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
* OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#include <math.h>
#include <string.h>
#include <stdint.h>
/*
* Infinite & NAN values
* for non-IEEE systems
*/
#ifndef HUGE_VAL
#ifdef HUGE
#define INFINITE_VALUE HUGE
#define NAN_VALUE HUGE
#endif
#else
#define INFINITE_VALUE HUGE_VAL
#define NAN_VALUE HUGE_VAL
#endif
/*
* IEEE 754 Extended Precision
*
* Implementation here is the 80-bit extended precision
* format of Motorola 68881, Motorola 68882 and Motorola
* 68040 FPUs, as well as Intel 80x87 FPUs.
*
* See:
* http://www.freescale.com/files/32bit/doc/fact_sheet/BR509.pdf
*/
/*
* Exponent range: [-16383,16383]
* Precision for mantissa: 64 bits with no hidden bit
* Bias: 16383
*/
/*
* Write IEEE Extended Precision Numbers
*/
void
ieee754_write_extended(double in, uint8_t* out)
{
int sgn, exp, shift;
double fraction, t;
unsigned int lexp, hexp;
unsigned long low, high;
if (in == 0.0) {
memset(out, 0, 10);
return;
}
if (in < 0.0) {
in = fabs(in);
sgn = 1;
} else
sgn = 0;
fraction = frexp(in, &exp);
if (exp == 0 || exp > 16384) {
if (exp > 16384) /* infinite value */
low = high = 0;
else {
low = 0x80000000;
high = 0;
}
exp = 32767;
goto done;
}
fraction = ldexp(fraction, 32);
t = floor(fraction);
low = (unsigned long) t;
fraction -= t;
t = floor(ldexp(fraction, 32));
high = (unsigned long) t;
/* Convert exponents < -16382 to -16382 (then they will be
* stored as -16383) */
if (exp < -16382) {
shift = 0 - exp - 16382;
high >>= shift;
high |= (low << (32 - shift));
low >>= shift;
exp = -16382;
}
exp += 16383 - 1; /* bias */
done:
lexp = ((unsigned int) exp) >> 8;
hexp = ((unsigned int) exp) & 0xFF;
/* big endian */
out[0] = ((uint8_t) sgn) << 7;
out[0] |= (uint8_t) lexp;
out[1] = (uint8_t) hexp;
out[2] = (uint8_t) (low >> 24);
out[3] = (uint8_t) ((low >> 16) & 0xFF);
out[4] = (uint8_t) ((low >> 8) & 0xFF);
out[5] = (uint8_t) (low & 0xFF);
out[6] = (uint8_t) (high >> 24);
out[7] = (uint8_t) ((high >> 16) & 0xFF);
out[8] = (uint8_t) ((high >> 8) & 0xFF);
out[9] = (uint8_t) (high & 0xFF);
return;
}
/*
* Read IEEE Extended Precision Numbers
*/
double
ieee754_read_extended(uint8_t* in)
{
int sgn, exp;
unsigned long low, high;
double out;
/* Extract the components from the big endian buffer */
sgn = (int) (in[0] >> 7);
exp = ((int) (in[0] & 0x7F) << 8) | ((int) in[1]);
low = (((unsigned long) in[2]) << 24)
| (((unsigned long) in[3]) << 16)
| (((unsigned long) in[4]) << 8) | (unsigned long) in[5];
high = (((unsigned long) in[6]) << 24)
| (((unsigned long) in[7]) << 16)
| (((unsigned long) in[8]) << 8) | (unsigned long) in[9];
if (exp == 0 && low == 0 && high == 0)
return (sgn ? -0.0 : 0.0);
switch (exp) {
case 32767:
if (low == 0 && high == 0)
return (sgn ? -INFINITE_VALUE : INFINITE_VALUE);
else
return (sgn ? -NAN_VALUE : NAN_VALUE);
default:
exp -= 16383; /* unbias exponent */
}
out = ldexp((double) low, -31 + exp);
out += ldexp((double) high, -63 + exp);
return (sgn ? -out : out);
}