Ryujinx-uplift/Ryujinx.HLE/HOS/Services/Aud/AudioRenderer/IAudioRenderer.cs

370 lines
12 KiB
C#

using ChocolArm64.Memory;
using Ryujinx.Audio;
using Ryujinx.Audio.Adpcm;
using Ryujinx.Common.Logging;
using Ryujinx.HLE.HOS.Ipc;
using Ryujinx.HLE.HOS.Kernel;
using Ryujinx.HLE.Utilities;
using System;
using System.Collections.Generic;
using System.Runtime.InteropServices;
namespace Ryujinx.HLE.HOS.Services.Aud.AudioRenderer
{
class IAudioRenderer : IpcService, IDisposable
{
//This is the amount of samples that are going to be appended
//each time that RequestUpdateAudioRenderer is called. Ideally,
//this value shouldn't be neither too small (to avoid the player
//starving due to running out of samples) or too large (to avoid
//high latency).
private const int MixBufferSamplesCount = 960;
private Dictionary<int, ServiceProcessRequest> m_Commands;
public override IReadOnlyDictionary<int, ServiceProcessRequest> Commands => m_Commands;
private KEvent UpdateEvent;
private AMemory Memory;
private IAalOutput AudioOut;
private AudioRendererParameter Params;
private MemoryPoolContext[] MemoryPools;
private VoiceContext[] Voices;
private int Track;
private PlayState PlayState;
public IAudioRenderer(
Horizon System,
AMemory Memory,
IAalOutput AudioOut,
AudioRendererParameter Params)
{
m_Commands = new Dictionary<int, ServiceProcessRequest>()
{
{ 0, GetSampleRate },
{ 1, GetSampleCount },
{ 2, GetMixBufferCount },
{ 3, GetState },
{ 4, RequestUpdateAudioRenderer },
{ 5, StartAudioRenderer },
{ 6, StopAudioRenderer },
{ 7, QuerySystemEvent }
};
UpdateEvent = new KEvent(System);
this.Memory = Memory;
this.AudioOut = AudioOut;
this.Params = Params;
Track = AudioOut.OpenTrack(
AudioConsts.HostSampleRate,
AudioConsts.HostChannelsCount,
AudioCallback);
MemoryPools = CreateArray<MemoryPoolContext>(Params.EffectCount + Params.VoiceCount * 4);
Voices = CreateArray<VoiceContext>(Params.VoiceCount);
InitializeAudioOut();
PlayState = PlayState.Stopped;
}
// GetSampleRate() -> u32
public long GetSampleRate(ServiceCtx Context)
{
Context.ResponseData.Write(Params.SampleRate);
return 0;
}
// GetSampleCount() -> u32
public long GetSampleCount(ServiceCtx Context)
{
Context.ResponseData.Write(Params.SampleCount);
return 0;
}
// GetMixBufferCount() -> u32
public long GetMixBufferCount(ServiceCtx Context)
{
Context.ResponseData.Write(Params.MixCount);
return 0;
}
// GetState() -> u32
private long GetState(ServiceCtx Context)
{
Context.ResponseData.Write((int)PlayState);
Logger.PrintStub(LogClass.ServiceAudio, $"Stubbed. Renderer State: {Enum.GetName(typeof(PlayState), PlayState)}");
return 0;
}
private void AudioCallback()
{
UpdateEvent.ReadableEvent.Signal();
}
private static T[] CreateArray<T>(int Size) where T : new()
{
T[] Output = new T[Size];
for (int Index = 0; Index < Size; Index++)
{
Output[Index] = new T();
}
return Output;
}
private void InitializeAudioOut()
{
AppendMixedBuffer(0);
AppendMixedBuffer(1);
AppendMixedBuffer(2);
AudioOut.Start(Track);
}
public long RequestUpdateAudioRenderer(ServiceCtx Context)
{
long OutputPosition = Context.Request.ReceiveBuff[0].Position;
long OutputSize = Context.Request.ReceiveBuff[0].Size;
AMemoryHelper.FillWithZeros(Context.Memory, OutputPosition, (int)OutputSize);
long InputPosition = Context.Request.SendBuff[0].Position;
StructReader Reader = new StructReader(Context.Memory, InputPosition);
StructWriter Writer = new StructWriter(Context.Memory, OutputPosition);
UpdateDataHeader InputHeader = Reader.Read<UpdateDataHeader>();
Reader.Read<BehaviorIn>(InputHeader.BehaviorSize);
MemoryPoolIn[] MemoryPoolsIn = Reader.Read<MemoryPoolIn>(InputHeader.MemoryPoolSize);
for (int Index = 0; Index < MemoryPoolsIn.Length; Index++)
{
MemoryPoolIn MemoryPool = MemoryPoolsIn[Index];
if (MemoryPool.State == MemoryPoolState.RequestAttach)
{
MemoryPools[Index].OutStatus.State = MemoryPoolState.Attached;
}
else if (MemoryPool.State == MemoryPoolState.RequestDetach)
{
MemoryPools[Index].OutStatus.State = MemoryPoolState.Detached;
}
}
Reader.Read<VoiceChannelResourceIn>(InputHeader.VoiceResourceSize);
VoiceIn[] VoicesIn = Reader.Read<VoiceIn>(InputHeader.VoiceSize);
for (int Index = 0; Index < VoicesIn.Length; Index++)
{
VoiceIn Voice = VoicesIn[Index];
VoiceContext VoiceCtx = Voices[Index];
VoiceCtx.SetAcquireState(Voice.Acquired != 0);
if (Voice.Acquired == 0)
{
continue;
}
if (Voice.FirstUpdate != 0)
{
VoiceCtx.AdpcmCtx = GetAdpcmDecoderContext(
Voice.AdpcmCoeffsPosition,
Voice.AdpcmCoeffsSize);
VoiceCtx.SampleFormat = Voice.SampleFormat;
VoiceCtx.SampleRate = Voice.SampleRate;
VoiceCtx.ChannelsCount = Voice.ChannelsCount;
VoiceCtx.SetBufferIndex(Voice.BaseWaveBufferIndex);
}
VoiceCtx.WaveBuffers[0] = Voice.WaveBuffer0;
VoiceCtx.WaveBuffers[1] = Voice.WaveBuffer1;
VoiceCtx.WaveBuffers[2] = Voice.WaveBuffer2;
VoiceCtx.WaveBuffers[3] = Voice.WaveBuffer3;
VoiceCtx.Volume = Voice.Volume;
VoiceCtx.PlayState = Voice.PlayState;
}
UpdateAudio();
UpdateDataHeader OutputHeader = new UpdateDataHeader();
int UpdateHeaderSize = Marshal.SizeOf<UpdateDataHeader>();
OutputHeader.Revision = IAudioRendererManager.RevMagic;
OutputHeader.BehaviorSize = 0xb0;
OutputHeader.MemoryPoolSize = (Params.EffectCount + Params.VoiceCount * 4) * 0x10;
OutputHeader.VoiceSize = Params.VoiceCount * 0x10;
OutputHeader.EffectSize = Params.EffectCount * 0x10;
OutputHeader.SinkSize = Params.SinkCount * 0x20;
OutputHeader.PerformanceManagerSize = 0x10;
OutputHeader.TotalSize = UpdateHeaderSize +
OutputHeader.BehaviorSize +
OutputHeader.MemoryPoolSize +
OutputHeader.VoiceSize +
OutputHeader.EffectSize +
OutputHeader.SinkSize +
OutputHeader.PerformanceManagerSize;
Writer.Write(OutputHeader);
foreach (MemoryPoolContext MemoryPool in MemoryPools)
{
Writer.Write(MemoryPool.OutStatus);
}
foreach (VoiceContext Voice in Voices)
{
Writer.Write(Voice.OutStatus);
}
return 0;
}
public long StartAudioRenderer(ServiceCtx Context)
{
Logger.PrintStub(LogClass.ServiceAudio, "Stubbed.");
PlayState = PlayState.Playing;
return 0;
}
public long StopAudioRenderer(ServiceCtx Context)
{
Logger.PrintStub(LogClass.ServiceAudio, "Stubbed.");
PlayState = PlayState.Stopped;
return 0;
}
public long QuerySystemEvent(ServiceCtx Context)
{
if (Context.Process.HandleTable.GenerateHandle(UpdateEvent.ReadableEvent, out int Handle) != KernelResult.Success)
{
throw new InvalidOperationException("Out of handles!");
}
Context.Response.HandleDesc = IpcHandleDesc.MakeCopy(Handle);
return 0;
}
private AdpcmDecoderContext GetAdpcmDecoderContext(long Position, long Size)
{
if (Size == 0)
{
return null;
}
AdpcmDecoderContext Context = new AdpcmDecoderContext();
Context.Coefficients = new short[Size >> 1];
for (int Offset = 0; Offset < Size; Offset += 2)
{
Context.Coefficients[Offset >> 1] = Memory.ReadInt16(Position + Offset);
}
return Context;
}
private void UpdateAudio()
{
long[] Released = AudioOut.GetReleasedBuffers(Track, 2);
for (int Index = 0; Index < Released.Length; Index++)
{
AppendMixedBuffer(Released[Index]);
}
}
private void AppendMixedBuffer(long Tag)
{
int[] MixBuffer = new int[MixBufferSamplesCount * AudioConsts.HostChannelsCount];
foreach (VoiceContext Voice in Voices)
{
if (!Voice.Playing)
{
continue;
}
int OutOffset = 0;
int PendingSamples = MixBufferSamplesCount;
while (PendingSamples > 0)
{
int[] Samples = Voice.GetBufferData(Memory, PendingSamples, out int ReturnedSamples);
if (ReturnedSamples == 0)
{
break;
}
PendingSamples -= ReturnedSamples;
for (int Offset = 0; Offset < Samples.Length; Offset++)
{
int Sample = (int)(Samples[Offset] * Voice.Volume);
MixBuffer[OutOffset++] += Sample;
}
}
}
AudioOut.AppendBuffer(Track, Tag, GetFinalBuffer(MixBuffer));
}
private static short[] GetFinalBuffer(int[] Buffer)
{
short[] Output = new short[Buffer.Length];
for (int Offset = 0; Offset < Buffer.Length; Offset++)
{
Output[Offset] = DspUtils.Saturate(Buffer[Offset]);
}
return Output;
}
public void Dispose()
{
Dispose(true);
}
protected virtual void Dispose(bool Disposing)
{
if (Disposing)
{
AudioOut.CloseTrack(Track);
}
}
}
}