1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
|
/*
* FILE: fresample.h
*
* The configuration constants below govern
* the number of bits in the input sample and filter coefficients, the
* number of bits to the right of the binary-point for fixed-point math, etc.
*
*/
/* #include "stdefs.h" */
typedef double SAMPLE_TYPE;
#define SCALE_FACTOR_TO_SHORT 32767
/* Conversion constants */
#define Nhc 8
/* #define Na 7 */
/* #define Np (Nhc+Na)*/
#define Npc (1<<Nhc)
/* #define Amask ((1<<Na)-1) */
/* #define Pmask ((1<<Np)-1) */
/* #define Nh 16 */
/* #define Nb 16 */
/* #define Nhxn 14 */
/* #define Nhg (Nh-Nhxn) */
/* #define NLpScl 13 */
/* Description of constants:
*
* Npc - is the number of look-up values available for the lowpass filter
* between the beginning of its impulse response and the "cutoff time"
* of the filter. The cutoff time is defined as the reciprocal of the
* lowpass-filter cut off frequence in Hz. For example, if the
* lowpass filter were a sinc function, Npc would be the index of the
* impulse-response lookup-table corresponding to the first zero-
* crossing of the sinc function. (The inverse first zero-crossing
* time of a sinc function equals its nominal cutoff frequency in Hz.)
* Npc must be a power of 2 due to the details of the current
* implementation. The default value of 512 is sufficiently high that
* using linear interpolation to fill in between the table entries
* gives approximately 16-bit accuracy in filter coefficients.
*
* Nhc - is log base 2 of Npc.
*
* Na - is the number of bits devoted to linear interpolation of the
* filter coefficients.
*
* Np - is Na + Nhc, the number of bits to the right of the binary point
* in the integer "time" variable. To the left of the point, it indexes
* the input array (X), and to the right, it is interpreted as a number
* between 0 and 1 sample of the input X. Np must be less than 16 in
* this implementation.
*
* Nh - is the number of bits in the filter coefficients. The sum of Nh and
* the number of bits in the input data (typically 16) cannot exceed 32.
* Thus Nh should be 16. The largest filter coefficient should nearly
* fill 16 bits (32767).
*
* Nb - is the number of bits in the input data. The sum of Nb and Nh cannot
* exceed 32.
*
* Nhxn - is the number of bits to right shift after multiplying each input
* sample times a filter coefficient. It can be as great as Nh and as
* small as 0. Nhxn = Nh-2 gives 2 guard bits in the multiply-add
* accumulation. If Nhxn=0, the accumulation will soon overflow 32 bits.
*
* Nhg - is the number of guard bits in mpy-add accumulation (equal to Nh-Nhxn)
*
* NLpScl - is the number of bits allocated to the unity-gain normalization
* factor. The output of the lowpass filter is multiplied by LpScl and
* then right-shifted NLpScl bits. To avoid overflow, we must have
* Nb+Nhg+NLpScl < 32.
*/
|
