summaryrefslogtreecommitdiffstats
path: root/src/audio_core/sink/sdl2_sink.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'src/audio_core/sink/sdl2_sink.cpp')
-rw-r--r--src/audio_core/sink/sdl2_sink.cpp556
1 files changed, 556 insertions, 0 deletions
diff --git a/src/audio_core/sink/sdl2_sink.cpp b/src/audio_core/sink/sdl2_sink.cpp
new file mode 100644
index 000000000..d6c9ec90d
--- /dev/null
+++ b/src/audio_core/sink/sdl2_sink.cpp
@@ -0,0 +1,556 @@
+// SPDX-FileCopyrightText: Copyright 2018 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include <algorithm>
+#include <atomic>
+
+#include "audio_core/audio_core.h"
+#include "audio_core/audio_event.h"
+#include "audio_core/audio_manager.h"
+#include "audio_core/sink/sdl2_sink.h"
+#include "audio_core/sink/sink_stream.h"
+#include "common/assert.h"
+#include "common/fixed_point.h"
+#include "common/logging/log.h"
+#include "common/reader_writer_queue.h"
+#include "common/ring_buffer.h"
+#include "common/settings.h"
+#include "core/core.h"
+
+// Ignore -Wimplicit-fallthrough due to https://github.com/libsdl-org/SDL/issues/4307
+#ifdef __clang__
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wimplicit-fallthrough"
+#endif
+#include <SDL.h>
+#ifdef __clang__
+#pragma clang diagnostic pop
+#endif
+
+namespace AudioCore::Sink {
+/**
+ * SDL sink stream, responsible for sinking samples to hardware.
+ */
+class SDLSinkStream final : public SinkStream {
+public:
+ /**
+ * Create a new sink stream.
+ *
+ * @param device_channels_ - Number of channels supported by the hardware.
+ * @param system_channels_ - Number of channels the audio systems expect.
+ * @param output_device - Name of the output device to use for this stream.
+ * @param input_device - Name of the input device to use for this stream.
+ * @param type_ - Type of this stream.
+ * @param system_ - Core system.
+ * @param event - Event used only for audio renderer, signalled on buffer consume.
+ */
+ SDLSinkStream(u32 device_channels_, const u32 system_channels_,
+ const std::string& output_device, const std::string& input_device,
+ const StreamType type_, Core::System& system_)
+ : type{type_}, system{system_} {
+ system_channels = system_channels_;
+ device_channels = device_channels_;
+
+ SDL_AudioSpec spec;
+ spec.freq = TargetSampleRate;
+ spec.channels = static_cast<u8>(device_channels);
+ spec.format = AUDIO_S16SYS;
+ if (type == StreamType::Render) {
+ spec.samples = TargetSampleCount;
+ } else {
+ spec.samples = 1024;
+ }
+ spec.callback = &SDLSinkStream::DataCallback;
+ spec.userdata = this;
+
+ playing_buffer.consumed = true;
+
+ std::string device_name{output_device};
+ bool capture{false};
+ if (type == StreamType::In) {
+ device_name = input_device;
+ capture = true;
+ }
+
+ SDL_AudioSpec obtained;
+ if (device_name.empty()) {
+ device = SDL_OpenAudioDevice(nullptr, capture, &spec, &obtained, false);
+ } else {
+ device = SDL_OpenAudioDevice(device_name.c_str(), capture, &spec, &obtained, false);
+ }
+
+ if (device == 0) {
+ LOG_CRITICAL(Audio_Sink, "Error opening SDL audio device: {}", SDL_GetError());
+ return;
+ }
+
+ LOG_DEBUG(Service_Audio,
+ "Opening sdl stream {} with: rate {} channels {} (system channels {}) "
+ " samples {}",
+ device, obtained.freq, obtained.channels, system_channels, obtained.samples);
+ }
+
+ /**
+ * Destroy the sink stream.
+ */
+ ~SDLSinkStream() override {
+ if (device == 0) {
+ return;
+ }
+
+ SDL_CloseAudioDevice(device);
+ }
+
+ /**
+ * Finalize the sink stream.
+ */
+ void Finalize() override {
+ if (device == 0) {
+ return;
+ }
+
+ SDL_CloseAudioDevice(device);
+ }
+
+ /**
+ * Start the sink stream.
+ *
+ * @param resume - Set to true if this is resuming the stream a previously-active stream.
+ * Default false.
+ */
+ void Start(const bool resume = false) override {
+ if (device == 0) {
+ return;
+ }
+
+ if (resume && was_playing) {
+ SDL_PauseAudioDevice(device, 0);
+ paused = false;
+ } else if (!resume) {
+ SDL_PauseAudioDevice(device, 0);
+ paused = false;
+ }
+ }
+
+ /**
+ * Stop the sink stream.
+ */
+ void Stop() {
+ if (device == 0) {
+ return;
+ }
+ SDL_PauseAudioDevice(device, 1);
+ paused = true;
+ }
+
+ /**
+ * Append a new buffer and its samples to a waiting queue to play.
+ *
+ * @param buffer - Audio buffer information to be queued.
+ * @param samples - The s16 samples to be queue for playback.
+ */
+ void AppendBuffer(::AudioCore::Sink::SinkBuffer& buffer, std::vector<s16>& samples) override {
+ if (type == StreamType::In) {
+ queue.enqueue(buffer);
+ queued_buffers++;
+ } else {
+ constexpr s32 min = std::numeric_limits<s16>::min();
+ constexpr s32 max = std::numeric_limits<s16>::max();
+
+ auto yuzu_volume{Settings::Volume()};
+ auto volume{system_volume * device_volume * yuzu_volume};
+
+ if (system_channels == 6 && device_channels == 2) {
+ // We're given 6 channels, but our device only outputs 2, so downmix.
+ constexpr std::array<f32, 4> down_mix_coeff{1.0f, 0.707f, 0.251f, 0.707f};
+
+ for (u32 read_index = 0, write_index = 0; read_index < samples.size();
+ read_index += system_channels, write_index += device_channels) {
+ const auto left_sample{
+ ((Common::FixedPoint<49, 15>(
+ samples[read_index + static_cast<u32>(Channels::FrontLeft)]) *
+ down_mix_coeff[0] +
+ samples[read_index + static_cast<u32>(Channels::Center)] *
+ down_mix_coeff[1] +
+ samples[read_index + static_cast<u32>(Channels::LFE)] *
+ down_mix_coeff[2] +
+ samples[read_index + static_cast<u32>(Channels::BackLeft)] *
+ down_mix_coeff[3]) *
+ volume)
+ .to_int()};
+
+ const auto right_sample{
+ ((Common::FixedPoint<49, 15>(
+ samples[read_index + static_cast<u32>(Channels::FrontRight)]) *
+ down_mix_coeff[0] +
+ samples[read_index + static_cast<u32>(Channels::Center)] *
+ down_mix_coeff[1] +
+ samples[read_index + static_cast<u32>(Channels::LFE)] *
+ down_mix_coeff[2] +
+ samples[read_index + static_cast<u32>(Channels::BackRight)] *
+ down_mix_coeff[3]) *
+ volume)
+ .to_int()};
+
+ samples[write_index + static_cast<u32>(Channels::FrontLeft)] =
+ static_cast<s16>(std::clamp(left_sample, min, max));
+ samples[write_index + static_cast<u32>(Channels::FrontRight)] =
+ static_cast<s16>(std::clamp(right_sample, min, max));
+ }
+
+ samples.resize(samples.size() / system_channels * device_channels);
+
+ } else if (system_channels == 2 && device_channels == 6) {
+ // We need moar samples! Not all games will provide 6 channel audio.
+ // TODO: Implement some upmixing here. Currently just passthrough, with other
+ // channels left as silence.
+ std::vector<s16> new_samples(samples.size() / system_channels * device_channels, 0);
+
+ for (u32 read_index = 0, write_index = 0; read_index < samples.size();
+ read_index += system_channels, write_index += device_channels) {
+ const auto left_sample{static_cast<s16>(std::clamp(
+ static_cast<s32>(
+ static_cast<f32>(
+ samples[read_index + static_cast<u32>(Channels::FrontLeft)]) *
+ volume),
+ min, max))};
+
+ new_samples[write_index + static_cast<u32>(Channels::FrontLeft)] = left_sample;
+
+ const auto right_sample{static_cast<s16>(std::clamp(
+ static_cast<s32>(
+ static_cast<f32>(
+ samples[read_index + static_cast<u32>(Channels::FrontRight)]) *
+ volume),
+ min, max))};
+
+ new_samples[write_index + static_cast<u32>(Channels::FrontRight)] =
+ right_sample;
+ }
+ samples = std::move(new_samples);
+
+ } else if (volume != 1.0f) {
+ for (u32 i = 0; i < samples.size(); i++) {
+ samples[i] = static_cast<s16>(std::clamp(
+ static_cast<s32>(static_cast<f32>(samples[i]) * volume), min, max));
+ }
+ }
+
+ samples_buffer.Push(samples);
+ queue.enqueue(buffer);
+ queued_buffers++;
+ }
+ }
+
+ /**
+ * Release a buffer. Audio In only, will fill a buffer with recorded samples.
+ *
+ * @param num_samples - Maximum number of samples to receive.
+ * @return Vector of recorded samples. May have fewer than num_samples.
+ */
+ std::vector<s16> ReleaseBuffer(const u64 num_samples) override {
+ static constexpr s32 min = std::numeric_limits<s16>::min();
+ static constexpr s32 max = std::numeric_limits<s16>::max();
+
+ auto samples{samples_buffer.Pop(num_samples)};
+
+ // TODO: Up-mix to 6 channels if the game expects it.
+ // For audio input this is unlikely to ever be the case though.
+
+ // Incoming mic volume seems to always be very quiet, so multiply by an additional 8 here.
+ // TODO: Play with this and find something that works better.
+ auto volume{system_volume * device_volume * 8};
+ for (u32 i = 0; i < samples.size(); i++) {
+ samples[i] = static_cast<s16>(
+ std::clamp(static_cast<s32>(static_cast<f32>(samples[i]) * volume), min, max));
+ }
+
+ if (samples.size() < num_samples) {
+ samples.resize(num_samples, 0);
+ }
+ return samples;
+ }
+
+ /**
+ * Check if a certain buffer has been consumed (fully played).
+ *
+ * @param tag - Unique tag of a buffer to check for.
+ * @return True if the buffer has been played, otherwise false.
+ */
+ bool IsBufferConsumed(const u64 tag) override {
+ if (released_buffer.tag == 0) {
+ if (!released_buffers.try_dequeue(released_buffer)) {
+ return false;
+ }
+ }
+
+ if (released_buffer.tag == tag) {
+ released_buffer.tag = 0;
+ return true;
+ }
+ return false;
+ }
+
+ /**
+ * Empty out the buffer queue.
+ */
+ void ClearQueue() override {
+ samples_buffer.Pop();
+ while (queue.pop()) {
+ }
+ while (released_buffers.pop()) {
+ }
+ released_buffer = {};
+ playing_buffer = {};
+ playing_buffer.consumed = true;
+ queued_buffers = 0;
+ }
+
+private:
+ /**
+ * Signal events back to the audio system that a buffer was played/can be filled.
+ *
+ * @param buffer - Consumed audio buffer to be released.
+ */
+ void SignalEvent(const ::AudioCore::Sink::SinkBuffer& buffer) {
+ auto& manager{system.AudioCore().GetAudioManager()};
+ switch (type) {
+ case StreamType::Out:
+ released_buffers.enqueue(buffer);
+ manager.SetEvent(Event::Type::AudioOutManager, true);
+ break;
+ case StreamType::In:
+ released_buffers.enqueue(buffer);
+ manager.SetEvent(Event::Type::AudioInManager, true);
+ break;
+ case StreamType::Render:
+ break;
+ }
+ }
+
+ /**
+ * Main callback from SDL. Either expects samples from us (audio render/audio out), or will
+ * provide samples to be copied (audio in).
+ *
+ * @param userdata - Custom data pointer passed along, points to a SDLSinkStream.
+ * @param stream - Buffer of samples to be filled or read.
+ * @param len - Length of the stream in bytes.
+ */
+ static void DataCallback(void* userdata, Uint8* stream, int len) {
+ auto* impl = static_cast<SDLSinkStream*>(userdata);
+
+ if (!impl) {
+ return;
+ }
+
+ const std::size_t num_channels = impl->GetDeviceChannels();
+ const std::size_t frame_size = num_channels;
+ const std::size_t frame_size_bytes = frame_size * sizeof(s16);
+ const std::size_t num_frames{len / num_channels / sizeof(s16)};
+ size_t frames_written{0};
+ [[maybe_unused]] bool underrun{false};
+
+ if (impl->type == StreamType::In) {
+ std::span<s16> input_buffer{reinterpret_cast<s16*>(stream), num_frames * frame_size};
+
+ while (frames_written < num_frames) {
+ auto& playing_buffer{impl->playing_buffer};
+
+ // If the playing buffer has been consumed or has no frames, we need a new one
+ if (playing_buffer.consumed || playing_buffer.frames == 0) {
+ if (!impl->queue.try_dequeue(impl->playing_buffer)) {
+ // If no buffer was available we've underrun, just push the samples and
+ // continue.
+ underrun = true;
+ impl->samples_buffer.Push(&input_buffer[frames_written * frame_size],
+ (num_frames - frames_written) * frame_size);
+ frames_written = num_frames;
+ continue;
+ } else {
+ impl->queued_buffers--;
+ impl->SignalEvent(impl->playing_buffer);
+ }
+ }
+
+ // Get the minimum frames available between the currently playing buffer, and the
+ // amount we have left to fill
+ size_t frames_available{
+ std::min(playing_buffer.frames - playing_buffer.frames_played,
+ num_frames - frames_written)};
+
+ impl->samples_buffer.Push(&input_buffer[frames_written * frame_size],
+ frames_available * frame_size);
+
+ frames_written += frames_available;
+ playing_buffer.frames_played += frames_available;
+
+ // If that's all the frames in the current buffer, add its samples and mark it as
+ // consumed
+ if (playing_buffer.frames_played >= playing_buffer.frames) {
+ impl->AddPlayedSampleCount(playing_buffer.frames_played * num_channels);
+ impl->playing_buffer.consumed = true;
+ }
+ }
+
+ std::memcpy(&impl->last_frame[0], &input_buffer[(frames_written - 1) * frame_size],
+ frame_size_bytes);
+ } else {
+ std::span<s16> output_buffer{reinterpret_cast<s16*>(stream), num_frames * frame_size};
+
+ while (frames_written < num_frames) {
+ auto& playing_buffer{impl->playing_buffer};
+
+ // If the playing buffer has been consumed or has no frames, we need a new one
+ if (playing_buffer.consumed || playing_buffer.frames == 0) {
+ if (!impl->queue.try_dequeue(impl->playing_buffer)) {
+ // If no buffer was available we've underrun, fill the remaining buffer with
+ // the last written frame and continue.
+ underrun = true;
+ for (size_t i = frames_written; i < num_frames; i++) {
+ std::memcpy(&output_buffer[i * frame_size], &impl->last_frame[0],
+ frame_size_bytes);
+ }
+ frames_written = num_frames;
+ continue;
+ } else {
+ impl->queued_buffers--;
+ impl->SignalEvent(impl->playing_buffer);
+ }
+ }
+
+ // Get the minimum frames available between the currently playing buffer, and the
+ // amount we have left to fill
+ size_t frames_available{
+ std::min(playing_buffer.frames - playing_buffer.frames_played,
+ num_frames - frames_written)};
+
+ impl->samples_buffer.Pop(&output_buffer[frames_written * frame_size],
+ frames_available * frame_size);
+
+ frames_written += frames_available;
+ playing_buffer.frames_played += frames_available;
+
+ // If that's all the frames in the current buffer, add its samples and mark it as
+ // consumed
+ if (playing_buffer.frames_played >= playing_buffer.frames) {
+ impl->AddPlayedSampleCount(playing_buffer.frames_played * num_channels);
+ impl->playing_buffer.consumed = true;
+ }
+ }
+
+ std::memcpy(&impl->last_frame[0], &output_buffer[(frames_written - 1) * frame_size],
+ frame_size_bytes);
+ }
+ }
+
+ /// SDL device id of the opened input/output device
+ SDL_AudioDeviceID device{};
+ /// Type of this stream
+ StreamType type;
+ /// Core system
+ Core::System& system;
+ /// Ring buffer of the samples waiting to be played or consumed
+ Common::RingBuffer<s16, 0x10000> samples_buffer;
+ /// Audio buffers queued and waiting to play
+ Common::ReaderWriterQueue<::AudioCore::Sink::SinkBuffer> queue;
+ /// The currently-playing audio buffer
+ ::AudioCore::Sink::SinkBuffer playing_buffer{};
+ /// Audio buffers which have been played and are in queue to be released by the audio system
+ Common::ReaderWriterQueue<::AudioCore::Sink::SinkBuffer> released_buffers{};
+ /// Currently released buffer waiting to be taken by the audio system
+ ::AudioCore::Sink::SinkBuffer released_buffer{};
+ /// The last played (or received) frame of audio, used when the callback underruns
+ std::array<s16, MaxChannels> last_frame{};
+};
+
+SDLSink::SDLSink(std::string_view target_device_name) {
+ if (!SDL_WasInit(SDL_INIT_AUDIO)) {
+ if (SDL_InitSubSystem(SDL_INIT_AUDIO) < 0) {
+ LOG_CRITICAL(Audio_Sink, "SDL_InitSubSystem audio failed: {}", SDL_GetError());
+ return;
+ }
+ }
+
+ if (target_device_name != auto_device_name && !target_device_name.empty()) {
+ output_device = target_device_name;
+ } else {
+ output_device.clear();
+ }
+
+ device_channels = 2;
+}
+
+SDLSink::~SDLSink() = default;
+
+SinkStream* SDLSink::AcquireSinkStream(Core::System& system, const u32 system_channels,
+ const std::string&, const StreamType type) {
+ SinkStreamPtr& stream = sink_streams.emplace_back(std::make_unique<SDLSinkStream>(
+ device_channels, system_channels, output_device, input_device, type, system));
+ return stream.get();
+}
+
+void SDLSink::CloseStream(const SinkStream* stream) {
+ for (size_t i = 0; i < sink_streams.size(); i++) {
+ if (sink_streams[i].get() == stream) {
+ sink_streams[i].reset();
+ sink_streams.erase(sink_streams.begin() + i);
+ break;
+ }
+ }
+}
+
+void SDLSink::CloseStreams() {
+ sink_streams.clear();
+}
+
+void SDLSink::PauseStreams() {
+ for (auto& stream : sink_streams) {
+ stream->Stop();
+ }
+}
+
+void SDLSink::UnpauseStreams() {
+ for (auto& stream : sink_streams) {
+ stream->Start();
+ }
+}
+
+f32 SDLSink::GetDeviceVolume() const {
+ if (sink_streams.empty()) {
+ return 1.0f;
+ }
+
+ return sink_streams[0]->GetDeviceVolume();
+}
+
+void SDLSink::SetDeviceVolume(const f32 volume) {
+ for (auto& stream : sink_streams) {
+ stream->SetDeviceVolume(volume);
+ }
+}
+
+void SDLSink::SetSystemVolume(const f32 volume) {
+ for (auto& stream : sink_streams) {
+ stream->SetSystemVolume(volume);
+ }
+}
+
+std::vector<std::string> ListSDLSinkDevices(const bool capture) {
+ std::vector<std::string> device_list;
+
+ if (!SDL_WasInit(SDL_INIT_AUDIO)) {
+ if (SDL_InitSubSystem(SDL_INIT_AUDIO) < 0) {
+ LOG_CRITICAL(Audio_Sink, "SDL_InitSubSystem audio failed: {}", SDL_GetError());
+ return {};
+ }
+ }
+
+ const int device_count = SDL_GetNumAudioDevices(capture);
+ for (int i = 0; i < device_count; ++i) {
+ device_list.emplace_back(SDL_GetAudioDeviceName(i, 0));
+ }
+
+ return device_list;
+}
+
+} // namespace AudioCore::Sink