ks0164: Pitch is floating point [O. Galibert]

This commit is contained in:
Olivier Galibert 2020-10-31 22:53:39 +01:00
parent ac74e7ed24
commit 4726878b3c

View File

@ -286,7 +286,7 @@ void ks0164_device::voice_w(offs_t offset, u16 data, u16 mem_mask)
if(m_cpu->pc() < 0x5f94 || m_cpu->pc() > 0x5fc0)
logerror("voice %02x.%02x = %04x @ %04x (%04x)\n", m_voice_select & 0x1f, offset, m_sregs[m_voice_select & 0x1f][offset], mem_mask, m_cpu->pc());
if(offset == 0 && (data & 1) && !(old & 1))
logerror("keyon %02x mode=%04x (%s %c %c %c) cur=%02x%04x.%04x loop=%02x%04x.%04x end=%02x%04x.%04x pitch=%04x 10=%02x/%02x:%02x/%02x 14=%03x/%03x:%03x/%03x 18=%04x/%04x c=%04x %04x %04x %04x %04x %04x %04x %04x %04x %04x %04x\n",
logerror("keyon %02x mode=%04x (%s %c %c %c) cur=%02x%04x.%04x loop=%02x%04x.%04x end=%02x%04x.%04x pitch=%02x.%03x 10=%02x/%02x:%02x/%02x 14=%03x/%03x:%03x/%03x 18=%04x/%04x c=%04x %04x %04x %04x %04x %04x %04x %04x %04x %04x %04x\n",
m_voice_select,
m_sregs[m_voice_select & 0x1f][0x00],
@ -308,7 +308,8 @@ void ks0164_device::voice_w(offs_t offset, u16 data, u16 mem_mask)
m_sregs[m_voice_select & 0x1f][0x0e],
m_sregs[m_voice_select & 0x1f][0x0f],
m_sregs[m_voice_select & 0x1f][0x08], // pitch
m_sregs[m_voice_select & 0x1f][0x08] & 0x1f, // pitch
m_sregs[m_voice_select & 0x1f][0x08] >> 5,
m_sregs[m_voice_select & 0x1f][0x10] >> 9,
m_sregs[m_voice_select & 0x1f][0x12] >> 9,
@ -431,9 +432,10 @@ void ks0164_device::sound_stream_update(sound_stream &stream, std::vector<read_s
for(int voice = 0; voice < 0x20; voice++) {
u16 *regs = m_sregs[voice];
if(regs[0] & 0x0001) {
u64 current = (u64(regs[1]) << 32) | (regs[2] << 16) | regs[3];
if(current & 0xc00000000000)
u64 current = (u64(regs[1]) << 32) | (u64(regs[2]) << 16) | regs[3];
if(current & 0xc000000000)
continue;
u32 adr = current >> 16;
@ -456,8 +458,14 @@ void ks0164_device::sound_stream_update(sound_stream &stream, std::vector<read_s
}
s16 samp = samp0 + (((samp1 - samp0) * (current & 0xffff)) >> 16);
current += regs[8];
u64 end = (u64(regs[0xd]) << 32) | (regs[0xe] << 16) | regs[0xf];
u32 step = 0x10000 | (regs[8] & ~0x1f);
u32 shift = regs[8] & 0x1f;
if(shift > 0x10)
step >>= 0x20 - shift;
else if(shift)
step <<= shift;
current += step;
u64 end = (u64(regs[0xd]) << 32) | (u64(regs[0xe]) << 16) | regs[0xf];
if(current >= end) {
// Is there a loop enabled flag?
u64 loop = (u64(regs[9]) << 32) | (regs[0xa] << 16) | regs[0xb];