suyu/src/audio_core/command_generator.h
Chloe Marcec 4a7fd91857 audren: Implement I3dl2Reverb
Most notable fix is the voices in Fire Emblem Three Houses
2021-02-12 18:48:10 -08:00

107 lines
5.1 KiB
C++

// Copyright 2020 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <array>
#include "audio_core/common.h"
#include "audio_core/voice_context.h"
#include "common/common_types.h"
namespace Core::Memory {
class Memory;
}
namespace AudioCore {
class MixContext;
class SplitterContext;
class ServerSplitterDestinationData;
class ServerMixInfo;
class EffectContext;
class EffectBase;
struct AuxInfoDSP;
struct I3dl2ReverbParams;
struct I3dl2ReverbState;
using MixVolumeBuffer = std::array<float, AudioCommon::MAX_MIX_BUFFERS>;
class CommandGenerator {
public:
explicit CommandGenerator(AudioCommon::AudioRendererParameter& worker_params_,
VoiceContext& voice_context_, MixContext& mix_context_,
SplitterContext& splitter_context_, EffectContext& effect_context_,
Core::Memory::Memory& memory_);
~CommandGenerator();
void ClearMixBuffers();
void GenerateVoiceCommands();
void GenerateVoiceCommand(ServerVoiceInfo& voice_info);
void GenerateSubMixCommands();
void GenerateFinalMixCommands();
void PreCommand();
void PostCommand();
[[nodiscard]] s32* GetChannelMixBuffer(s32 channel);
[[nodiscard]] const s32* GetChannelMixBuffer(s32 channel) const;
[[nodiscard]] s32* GetMixBuffer(std::size_t index);
[[nodiscard]] const s32* GetMixBuffer(std::size_t index) const;
[[nodiscard]] std::size_t GetMixChannelBufferOffset(s32 channel) const;
[[nodiscard]] std::size_t GetTotalMixBufferCount() const;
private:
void GenerateDataSourceCommand(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 channel);
void GenerateBiquadFilterCommandForVoice(ServerVoiceInfo& voice_info, VoiceState& dsp_state,
s32 mix_buffer_count, s32 channel);
void GenerateVolumeRampCommand(float last_volume, float current_volume, s32 channel,
s32 node_id);
void GenerateVoiceMixCommand(const MixVolumeBuffer& mix_volumes,
const MixVolumeBuffer& last_mix_volumes, VoiceState& dsp_state,
s32 mix_buffer_offset, s32 mix_buffer_count, s32 voice_index,
s32 node_id);
void GenerateSubMixCommand(ServerMixInfo& mix_info);
void GenerateMixCommands(ServerMixInfo& mix_info);
void GenerateMixCommand(std::size_t output_offset, std::size_t input_offset, float volume,
s32 node_id);
void GenerateFinalMixCommand();
void GenerateBiquadFilterCommand(s32 mix_buffer, const BiquadFilterParameter& params,
std::array<s64, 2>& state, std::size_t input_offset,
std::size_t output_offset, s32 sample_count, s32 node_id);
void GenerateDepopPrepareCommand(VoiceState& dsp_state, std::size_t mix_buffer_count,
std::size_t mix_buffer_offset);
void GenerateDepopForMixBuffersCommand(std::size_t mix_buffer_count,
std::size_t mix_buffer_offset, s32 sample_rate);
void GenerateEffectCommand(ServerMixInfo& mix_info);
void GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled);
void GenerateBiquadFilterEffectCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled);
void GenerateAuxCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled);
[[nodiscard]] ServerSplitterDestinationData* GetDestinationData(s32 splitter_id, s32 index);
s32 WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u32 max_samples, const s32* data,
u32 sample_count, u32 write_offset, u32 write_count);
s32 ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples, s32* out_data,
u32 sample_count, u32 read_offset, u32 read_count);
void InitializeI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
std::vector<u8>& work_buffer);
void UpdateI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state, bool should_clear);
// DSP Code
s32 DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_count,
s32 channel, std::size_t mix_offset);
s32 DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_count,
s32 channel, std::size_t mix_offset);
void DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* output, VoiceState& dsp_state,
s32 channel, s32 target_sample_rate, s32 sample_count, s32 node_id);
AudioCommon::AudioRendererParameter& worker_params;
VoiceContext& voice_context;
MixContext& mix_context;
SplitterContext& splitter_context;
EffectContext& effect_context;
Core::Memory::Memory& memory;
std::vector<s32> mix_buffer{};
std::vector<s32> sample_buffer{};
std::vector<s32> depop_buffer{};
bool dumping_frame{false};
};
} // namespace AudioCore