midou/citra
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update cycles and explanation

Browse Source
This commit is contained in:
xperia64 2020-04-26 03:14:54 -04:00
parent 3a1601a534
commit a0e8255b65
1 changed files with 6 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
src/audio_core/hle/hle.cpp
View File

@ -48,18 +48,13 @@ void DspHle::serialize(Archive& ar, const unsigned int) {
SERIALIZE_IMPL(DspHle)
// TODO(xperia64): The value below is the "perfect" mathematical ratio
// of ARM11 cycles per audio frame. As per merry, this was calculated to be
// (ARM11 freq)*(samples per frame)/(sample rate)
// = (268,111,855.956 Hz) * (160 samples)/(32,728 Hz)
// = (1310739.946008311...) ~ 1310740
// of ARM11 cycles per audio frame, as per merry's suggestion
// samples per frame * teaklite cycles per sample * 2 ARM11 cycles/teaklite cycle
// (160 * 4096 * 2) = (1310720)
//
// This value was originally set to 1310252, which was determined by measuring it on hardware
// However, as of when this was written, Project Mirai 1/2/DX desync on HLE
// such that the music track runs ahead of the gameplay.
// When the value is set to 1310740, all three games are playable
// The audio track only drifts ~1ms over a 4+ minute song compared to hardware
// and the button presses match as well as I can determine by playing the game/recording
static constexpr u64 audio_frame_ticks = 1310740ull; ///< Units: ARM11 cycles
// As per merry, it may be useful to verify this on hardware with the more recently
// discovered "correct" ARM11 frequency of 268111856 as opposed to 268123480
static constexpr u64 audio_frame_ticks = 160 * 4096 * 2ull; ///< Units: ARM11 cycles
struct DspHle::Impl final {
public: