hashtags/lmms
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases 1 Wiki Activity

Fix incorrect calculation of IntBits

Browse Source 
Fix over-shifting of int by explicitly making type ULL.
Taken from upstream:
aaca6447de
This commit is contained in:
Tres Finocchiaro
2015-08-15 13:23:32 -04:00
parent 6df512dc76
commit 5375bda6ac
1 changed files with 7 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
plugins/LadspaEffect/calf/src/calf/fixed_point.h
View File

@@ -37,7 +37,7 @@ inline T shr(T v, int bits = 1) {
template<class T, int FracBits> class fixed_point {
T value;
enum { IntBits = (sizeof(T)/8) - FracBits };
enum { IntBits = (sizeof(T)*8) - FracBits };
public:
/// default constructor, does not initialize the value, just like - say - float doesn't
@@ -210,13 +210,13 @@ public:
template<class U, int UseBits, class MulType>
inline U lerp_by_fract_int(U v1, U v2) const {
int fp = fpart<UseBits>();
assert ( fp >=0 && fp <= (1<<UseBits));
assert ( fp >=0 && fp <= (1ULL<<UseBits));
// printf("diff =
return v1 + shr(((MulType)(v2-v1) * fp), UseBits);
}
template<class U, int UseBits>
inline U lerp_table_lookup_int(U data[(unsigned int)(1<<IntBits)+1]) const {
inline U lerp_table_lookup_int(U data[(unsigned int)(1ULL<<IntBits)+1]) const {
unsigned int pos = uipart();
return lerp_by_fract_int<U, UseBits>(data[pos], data[pos+1]);
}
@@ -224,19 +224,19 @@ public:
/// Untested... I've started it to get a sin/cos readout for rotaryorgan, but decided to use table-less solution instead
/// Do not assume it works, because it most probably doesn't
template<class U, int UseBits>
inline U lerp_table_lookup_int_shift(U data[(unsigned int)(1<<IntBits)+1], unsigned int shift) {
unsigned int pos = (uipart() + shift) & ((1 << IntBits) - 1);
inline U lerp_table_lookup_int_shift(U data[(unsigned int)(1ULL<<IntBits)+1], unsigned int shift) {
unsigned int pos = (uipart() + shift) & ((1ULL << IntBits) - 1);
return lerp_by_fract_int<U, UseBits>(data[pos], data[pos+1]);
}
template<class U>
inline U lerp_table_lookup_float(U data[(unsigned int)(1<<IntBits)+1]) const {
inline U lerp_table_lookup_float(U data[(unsigned int)(1ULL<<IntBits)+1]) const {
unsigned int pos = uipart();
return data[pos] + (data[pos+1]-data[pos]) * fpart_as_double();
}
template<class U>
inline U lerp_table_lookup_float_mask(U data[(unsigned int)(1<<IntBits)+1], unsigned int mask) const {
inline U lerp_table_lookup_float_mask(U data[(unsigned int)(1ULL<<IntBits)+1], unsigned int mask) const {
unsigned int pos = ui64part() & mask;
// printf("full = %lld pos = %d + %f\n", value, pos, fpart_as_double());
return data[pos] + (data[pos+1]-data[pos]) * fpart_as_double();