diff --git a/include/audio/aac_decoder.hpp b/include/audio/aac_decoder.hpp
index 6538bce9..8964e1f2 100644
--- a/include/audio/aac_decoder.hpp
+++ b/include/audio/aac_decoder.hpp
@@ -18,7 +18,8 @@ namespace Audio::AAC {
 
 	  public:
 		// Decode function. Takes in a reference to the AAC response & request, and a callback for paddr -> pointer conversions
-		void decode(AAC::Message& response, const AAC::Message& request, PaddrCallback paddrCallback);
+		// We also allow for optionally muting the AAC output (setting all of it to 0) instead of properly decoding it, for debug/research purposes
+		void decode(AAC::Message& response, const AAC::Message& request, PaddrCallback paddrCallback, bool enableAudio = true);
 		~Decoder();
 	};
 }  // namespace Audio::AAC
\ No newline at end of file
diff --git a/include/audio/dsp_core.hpp b/include/audio/dsp_core.hpp
index 5f5efa22..f180e717 100644
--- a/include/audio/dsp_core.hpp
+++ b/include/audio/dsp_core.hpp
@@ -14,6 +14,7 @@
 // The DSP core must have access to the DSP service to be able to trigger interrupts properly
 class DSPService;
 class Memory;
+struct EmulatorConfig;
 
 namespace Audio {
 	// There are 160 stereo samples in 1 audio frame, so 320 samples total
@@ -31,6 +32,7 @@ namespace Audio {
 		Memory& mem;
 		Scheduler& scheduler;
 		DSPService& dspService;
+		EmulatorConfig& settings;
 
 		Samples sampleBuffer;
 		bool audioEnabled = false;
@@ -39,7 +41,8 @@ namespace Audio {
 
 	  public:
 		enum class Type { Null, Teakra, HLE };
-		DSPCore(Memory& mem, Scheduler& scheduler, DSPService& dspService) : mem(mem), scheduler(scheduler), dspService(dspService) {}
+		DSPCore(Memory& mem, Scheduler& scheduler, DSPService& dspService, EmulatorConfig& settings)
+			: mem(mem), scheduler(scheduler), dspService(dspService), settings(settings) {}
 		virtual ~DSPCore() {}
 
 		virtual void reset() = 0;
@@ -62,5 +65,5 @@ namespace Audio {
 		virtual void setAudioEnabled(bool enable) { audioEnabled = enable; }
 	};
 
-	std::unique_ptr<DSPCore> makeDSPCore(DSPCore::Type type, Memory& mem, Scheduler& scheduler, DSPService& dspService);
+	std::unique_ptr<DSPCore> makeDSPCore(EmulatorConfig& config, Memory& mem, Scheduler& scheduler, DSPService& dspService);
 }  // namespace Audio
\ No newline at end of file
diff --git a/include/audio/hle_core.hpp b/include/audio/hle_core.hpp
index 32bbaae8..0e3b8636 100644
--- a/include/audio/hle_core.hpp
+++ b/include/audio/hle_core.hpp
@@ -206,7 +206,7 @@ namespace Audio {
 		SampleBuffer decodeADPCM(const u8* data, usize sampleCount, Source& source);
 
 	  public:
-		HLE_DSP(Memory& mem, Scheduler& scheduler, DSPService& dspService);
+		HLE_DSP(Memory& mem, Scheduler& scheduler, DSPService& dspService, EmulatorConfig& config);
 		~HLE_DSP() override {}
 
 		void reset() override;
diff --git a/include/audio/miniaudio_device.hpp b/include/audio/miniaudio_device.hpp
index f3b212e2..516c1602 100644
--- a/include/audio/miniaudio_device.hpp
+++ b/include/audio/miniaudio_device.hpp
@@ -3,6 +3,7 @@
 #include <string>
 #include <vector>
 
+#include "config.hpp"
 #include "helpers.hpp"
 #include "miniaudio.h"
 #include "ring_buffer.hpp"
@@ -12,12 +13,13 @@ class MiniAudioDevice {
 	static constexpr ma_uint32 sampleRate = 32768;  // 3DS sample rate
 	static constexpr ma_uint32 channelCount = 2;    // Audio output is stereo
 
+	ma_device device;
 	ma_context context;
 	ma_device_config deviceConfig;
-	ma_device device;
-	ma_resampler resampler;
 	Samples* samples = nullptr;
 
+	const AudioDeviceConfig& audioSettings;
+
 	bool initialized = false;
 	bool running = false;
 
@@ -26,7 +28,8 @@ class MiniAudioDevice {
 	std::vector<std::string> audioDevices;
 
   public:
-	MiniAudioDevice();
+	MiniAudioDevice(const AudioDeviceConfig& audioSettings);
+
 	// If safe is on, we create a null audio device
 	void init(Samples& samples, bool safe = false);
 	void close();
diff --git a/include/audio/null_core.hpp b/include/audio/null_core.hpp
index 7a0f368d..e7ae12dc 100644
--- a/include/audio/null_core.hpp
+++ b/include/audio/null_core.hpp
@@ -23,7 +23,7 @@ namespace Audio {
 		bool loaded = false;  // Have we loaded a component?
 
 	  public:
-		NullDSP(Memory& mem, Scheduler& scheduler, DSPService& dspService) : DSPCore(mem, scheduler, dspService) {}
+		NullDSP(Memory& mem, Scheduler& scheduler, DSPService& dspService, EmulatorConfig& config) : DSPCore(mem, scheduler, dspService, config) {}
 		~NullDSP() override {}
 
 		void reset() override;
diff --git a/include/audio/teakra_core.hpp b/include/audio/teakra_core.hpp
index 17104985..3fd5abc3 100644
--- a/include/audio/teakra_core.hpp
+++ b/include/audio/teakra_core.hpp
@@ -77,7 +77,7 @@ namespace Audio {
 		}
 
 	  public:
-		TeakraDSP(Memory& mem, Scheduler& scheduler, DSPService& dspService);
+		TeakraDSP(Memory& mem, Scheduler& scheduler, DSPService& dspService, EmulatorConfig& config);
 		~TeakraDSP() override {}
 
 		void reset() override;
diff --git a/include/config.hpp b/include/config.hpp
index 81a03385..9364a33d 100644
--- a/include/config.hpp
+++ b/include/config.hpp
@@ -4,6 +4,19 @@
 #include "audio/dsp_core.hpp"
 #include "renderer.hpp"
 
+struct AudioDeviceConfig {
+	float volumeRaw = 1.0f;
+	bool muteAudio = false;
+
+	float getVolume() const {
+		if (muteAudio) {
+			return 0.0f;
+		}
+
+		return volumeRaw;
+	}
+};
+
 // Remember to initialize every field here to its default value otherwise bad things will happen
 struct EmulatorConfig {
 	// Only enable the shader JIT by default on platforms where it's completely tested
@@ -41,6 +54,7 @@ struct EmulatorConfig {
 
 	bool audioEnabled = false;
 	bool vsyncEnabled = true;
+	bool aacEnabled = true;  // Enable AAC audio?
 
 	bool enableRenderdoc = false;
 	bool printAppVersion = true;
@@ -70,6 +84,7 @@ struct EmulatorConfig {
 	};
 
 	WindowSettings windowSettings;
+	AudioDeviceConfig audioDeviceConfig;
 
 	EmulatorConfig(const std::filesystem::path& path);
 	void load();
diff --git a/src/config.cpp b/src/config.cpp
index 4cd18858..833a1b4e 100644
--- a/src/config.cpp
+++ b/src/config.cpp
@@ -96,7 +96,14 @@ void EmulatorConfig::load() {
 
 			auto dspCoreName = toml::find_or<std::string>(audio, "DSPEmulation", "HLE");
 			dspType = Audio::DSPCore::typeFromString(dspCoreName);
+			
 			audioEnabled = toml::find_or<toml::boolean>(audio, "EnableAudio", false);
+			aacEnabled = toml::find_or<toml::boolean>(audio, "EnableAACAudio", true);
+
+			audioDeviceConfig.muteAudio = toml::find_or<toml::boolean>(audio, "MuteAudio", false);
+			// Our volume ranges from 0.0 (muted) to 2.0 (boosted, using a logarithmic scale). 1.0 is the "default" volume, ie we don't adjust the PCM
+			// samples at all.
+			audioDeviceConfig.volumeRaw = float(std::clamp(toml::find_or<toml::floating>(audio, "AudioVolume", 1.0), 0.0, 2.0));
 		}
 	}
 
@@ -167,6 +174,9 @@ void EmulatorConfig::save() {
 
 	data["Audio"]["DSPEmulation"] = std::string(Audio::DSPCore::typeToString(dspType));
 	data["Audio"]["EnableAudio"] = audioEnabled;
+	data["Audio"]["EnableAACAudio"] = aacEnabled;
+	data["Audio"]["MuteAudio"] = audioDeviceConfig.muteAudio;
+	data["Audio"]["AudioVolume"] = double(audioDeviceConfig.volumeRaw);
 
 	data["Battery"]["ChargerPlugged"] = chargerPlugged;
 	data["Battery"]["BatteryPercentage"] = batteryPercentage;
diff --git a/src/core/PICA/dynapica/shader_rec_emitter_x64.cpp b/src/core/PICA/dynapica/shader_rec_emitter_x64.cpp
index 142ff8c8..ddec3a36 100644
--- a/src/core/PICA/dynapica/shader_rec_emitter_x64.cpp
+++ b/src/core/PICA/dynapica/shader_rec_emitter_x64.cpp
@@ -370,12 +370,11 @@ void ShaderEmitter::storeRegister(Xmm source, const PICAShader& shader, u32 dest
 	} else if (haveSSE4_1) {
 		// Bit reverse the write mask because that is what blendps expects
 		u32 adjustedMask = ((writeMask >> 3) & 0b1) | ((writeMask >> 1) & 0b10) | ((writeMask << 1) & 0b100) | ((writeMask << 3) & 0b1000);
-		// Don't accidentally overwrite scratch1 if that is what we're writing derp
-		Xmm temp = (source == scratch1) ? scratch2 : scratch1;
 
-		movaps(temp, xword[statePointer + offset]);     // Read current value of dest
-		blendps(temp, source, adjustedMask);            // Blend with source
-		movaps(xword[statePointer + offset], temp);     // Write back
+		// Blend current value of dest with source. We have to invert the bits of the mask, as we do blendps source, dest instead of dest, source
+		// Note: This destroys source
+		blendps(source, xword[statePointer + offset], adjustedMask ^ 0xF);
+		movaps(xword[statePointer + offset], source);     // Write back
 	} else {
 		// Blend algo referenced from Citra
 		const u8 selector = (((writeMask & 0b1000) ? 1 : 0) << 0) |
diff --git a/src/core/audio/aac_decoder.cpp b/src/core/audio/aac_decoder.cpp
index 281539d8..af88485c 100644
--- a/src/core/audio/aac_decoder.cpp
+++ b/src/core/audio/aac_decoder.cpp
@@ -5,7 +5,7 @@
 #include <vector>
 using namespace Audio;
 
-void AAC::Decoder::decode(AAC::Message& response, const AAC::Message& request, AAC::Decoder::PaddrCallback paddrCallback) {
+void AAC::Decoder::decode(AAC::Message& response, const AAC::Message& request, AAC::Decoder::PaddrCallback paddrCallback, bool enableAudio) {
 	// Copy the command and mode fields of the request to the response
 	response.command = request.command;
 	response.mode = request.mode;
@@ -95,9 +95,16 @@ void AAC::Decoder::decode(AAC::Message& response, const AAC::Message& request, A
 				}
 			}
 
-			for (int sample = 0; sample < info->frameSize; sample++) {
+			if (enableAudio) {
+				for (int sample = 0; sample < info->frameSize; sample++) {
+					for (int stream = 0; stream < channels; stream++) {
+						audioStreams[stream].push_back(frame[(sample * channels) + stream]);
+					}
+				}
+			} else {
+				// If audio is not enabled, push 0s
 				for (int stream = 0; stream < channels; stream++) {
-					audioStreams[stream].push_back(frame[(sample * channels) + stream]);
+					audioStreams[stream].resize(audioStreams[stream].size() + info->frameSize, 0);
 				}
 			}
 		} else {
diff --git a/src/core/audio/dsp_core.cpp b/src/core/audio/dsp_core.cpp
index 01cee11e..c793fcf8 100644
--- a/src/core/audio/dsp_core.cpp
+++ b/src/core/audio/dsp_core.cpp
@@ -8,17 +8,17 @@
 #include "audio/null_core.hpp"
 #include "audio/teakra_core.hpp"
 
-std::unique_ptr<Audio::DSPCore> Audio::makeDSPCore(DSPCore::Type type, Memory& mem, Scheduler& scheduler, DSPService& dspService) {
+std::unique_ptr<Audio::DSPCore> Audio::makeDSPCore(EmulatorConfig& config, Memory& mem, Scheduler& scheduler, DSPService& dspService) {
 	std::unique_ptr<DSPCore> core;
 
-	switch (type) {
-		case DSPCore::Type::Null: core = std::make_unique<NullDSP>(mem, scheduler, dspService); break;
-		case DSPCore::Type::Teakra: core = std::make_unique<TeakraDSP>(mem, scheduler, dspService); break;
-		case DSPCore::Type::HLE: core = std::make_unique<HLE_DSP>(mem, scheduler, dspService); break;
+	switch (config.dspType) {
+		case DSPCore::Type::Null: core = std::make_unique<NullDSP>(mem, scheduler, dspService, config); break;
+		case DSPCore::Type::Teakra: core = std::make_unique<TeakraDSP>(mem, scheduler, dspService, config); break;
+		case DSPCore::Type::HLE: core = std::make_unique<HLE_DSP>(mem, scheduler, dspService, config); break;
 
 		default:
 			Helpers::warn("Invalid DSP core selected!");
-			core = std::make_unique<NullDSP>(mem, scheduler, dspService);
+			core = std::make_unique<NullDSP>(mem, scheduler, dspService, config);
 			break;
 	}
 
diff --git a/src/core/audio/hle_core.cpp b/src/core/audio/hle_core.cpp
index 7e82a139..73d39adb 100644
--- a/src/core/audio/hle_core.cpp
+++ b/src/core/audio/hle_core.cpp
@@ -8,6 +8,7 @@
 
 #include "audio/aac_decoder.hpp"
 #include "audio/dsp_simd.hpp"
+#include "config.hpp"
 #include "services/dsp.hpp"
 
 namespace Audio {
@@ -20,7 +21,8 @@ namespace Audio {
 		};
 	}
 
-	HLE_DSP::HLE_DSP(Memory& mem, Scheduler& scheduler, DSPService& dspService) : DSPCore(mem, scheduler, dspService) {
+	HLE_DSP::HLE_DSP(Memory& mem, Scheduler& scheduler, DSPService& dspService, EmulatorConfig& config)
+		: DSPCore(mem, scheduler, dspService, config) {
 		// Set up source indices
 		for (int i = 0; i < sources.size(); i++) {
 			sources[i].index = i;
@@ -702,24 +704,9 @@ namespace Audio {
 		AAC::Message response;
 
 		switch (request.command) {
-			case AAC::Command::EncodeDecode: {
-				// Dummy response to stop games from hanging
-				response.resultCode = AAC::ResultCode::Success;
-				response.decodeResponse.channelCount = 2;
-				response.decodeResponse.sampleCount = 1024;
-				response.decodeResponse.size = 0;
-				response.decodeResponse.sampleRate = AAC::SampleRate::Rate48000;
-
-				response.command = request.command;
-				response.mode = request.mode;
-
-				// TODO: Make this a toggle in config.toml. Currently we have it on by default.
-				constexpr bool enableAAC = true;
-				if (enableAAC) {
-					aacDecoder->decode(response, request, [this](u32 paddr) { return getPointerPhys<u8>(paddr); });
-				}
+			case AAC::Command::EncodeDecode:
+				aacDecoder->decode(response, request, [this](u32 paddr) { return getPointerPhys<u8>(paddr); }, settings.aacEnabled);
 				break;
-			}
 
 			case AAC::Command::Init:
 			case AAC::Command::Shutdown:
diff --git a/src/core/audio/miniaudio_device.cpp b/src/core/audio/miniaudio_device.cpp
index cd537aef..b513450e 100644
--- a/src/core/audio/miniaudio_device.cpp
+++ b/src/core/audio/miniaudio_device.cpp
@@ -1,10 +1,14 @@
 #include "audio/miniaudio_device.hpp"
 
+#include <algorithm>
+#include <cmath>
 #include <cstring>
+#include <limits>
 
 #include "helpers.hpp"
 
-MiniAudioDevice::MiniAudioDevice() : initialized(false), running(false), samples(nullptr) {}
+MiniAudioDevice::MiniAudioDevice(const AudioDeviceConfig& audioSettings)
+	: initialized(false), running(false), samples(nullptr), audioSettings(audioSettings) {}
 
 void MiniAudioDevice::init(Samples& samples, bool safe) {
 	this->samples = &samples;
@@ -106,6 +110,40 @@ void MiniAudioDevice::init(Samples& samples, bool safe) {
 			std::memcpy(&self->lastStereoSample[0], &output[(samplesWritten - 1) * 2], sizeof(lastStereoSample));
 		}
 
+		// Adjust the volume of our samples based on the emulator's volume slider
+		float audioVolume = self->audioSettings.getVolume();
+		// If volume is 1.0 we don't need to do anything
+		if (audioVolume != 1.0f) {
+			s16* sample = output;
+
+			// If our volume is > 1.0 then we boost samples using a logarithmic scale,
+			// In this case we also have to clamp samples to make sure they don't wrap around
+			if (audioVolume > 1.0f) {
+				audioVolume = 0.6 + 20 * std::log10(audioVolume);
+
+				constexpr s32 min = s32(std::numeric_limits<s16>::min());
+				constexpr s32 max = s32(std::numeric_limits<s16>::max());
+
+				for (usize i = 0; i < samplesWritten; i += 2) {
+					s16 l = s16(std::clamp<s32>(s32(float(sample[0]) * audioVolume), min, max));
+					s16 r = s16(std::clamp<s32>(s32(float(sample[1]) * audioVolume), min, max));
+
+					*sample++ = l;
+					*sample++ = r;
+				}
+			} else {
+				// If our volume is in [0.0, 1.0) then just multiply by the volume. No need to clamp, since there is no danger of our samples wrapping
+				// around due to overflow
+				for (usize i = 0; i < samplesWritten; i += 2) {
+					s16 l = s16(float(sample[0]) * audioVolume);
+					s16 r = s16(float(sample[1]) * audioVolume);
+
+					*sample++ = l;
+					*sample++ = r;
+				}
+			}
+		}
+
 		// If underruning, copy the last output sample
 		{
 			s16* pointer = &output[samplesWritten * 2];
diff --git a/src/core/audio/teakra_core.cpp b/src/core/audio/teakra_core.cpp
index da2e5a5a..cf1484f8 100644
--- a/src/core/audio/teakra_core.cpp
+++ b/src/core/audio/teakra_core.cpp
@@ -36,8 +36,8 @@ struct Dsp1 {
 	Segment segments[10];
 };
 
-TeakraDSP::TeakraDSP(Memory& mem, Scheduler& scheduler, DSPService& dspService)
-	: DSPCore(mem, scheduler, dspService), pipeBaseAddr(0), running(false) {
+TeakraDSP::TeakraDSP(Memory& mem, Scheduler& scheduler, DSPService& dspService, EmulatorConfig& config)
+	: DSPCore(mem, scheduler, dspService, config), pipeBaseAddr(0), running(false) {
 	// Set up callbacks for Teakra
 	Teakra::AHBMCallback ahbm;
 
diff --git a/src/emulator.cpp b/src/emulator.cpp
index 9851adb1..fc25eacb 100644
--- a/src/emulator.cpp
+++ b/src/emulator.cpp
@@ -20,7 +20,7 @@ __declspec(dllexport) DWORD AmdPowerXpressRequestHighPerformance = 1;
 
 Emulator::Emulator()
 	: config(getConfigPath()), kernel(cpu, memory, gpu, config), cpu(memory, kernel, *this), gpu(memory, config), memory(cpu.getTicksRef(), config),
-	  cheats(memory, kernel.getServiceManager().getHID()), lua(*this), running(false)
+	  cheats(memory, kernel.getServiceManager().getHID()), audioDevice(config.audioDeviceConfig), lua(*this), running(false)
 #ifdef PANDA3DS_ENABLE_HTTP_SERVER
 	  ,
 	  httpServer(this)
@@ -28,7 +28,7 @@ Emulator::Emulator()
 {
 	DSPService& dspService = kernel.getServiceManager().getDSP();
 
-	dsp = Audio::makeDSPCore(config.dspType, memory, scheduler, dspService);
+	dsp = Audio::makeDSPCore(config, memory, scheduler, dspService);
 	dspService.setDSPCore(dsp.get());
 
 	audioDevice.init(dsp->getSamples());
diff --git a/src/libretro_core.cpp b/src/libretro_core.cpp
index 319b0a72..1e5b532d 100644
--- a/src/libretro_core.cpp
+++ b/src/libretro_core.cpp
@@ -172,12 +172,16 @@ static void configInit() {
 		{"panda3ds_use_vsync", "Enable VSync; enabled|disabled"},
 		{"panda3ds_dsp_emulation", "DSP emulation; HLE|LLE|Null"},
 		{"panda3ds_use_audio", "Enable audio; disabled|enabled"},
+		{"panda3ds_audio_volume", "Audio volume; 100|0|10|20|40|60|80|90|100|120|140|150|180|200"},
+		{"panda3ds_mute_audio", "Mute audio; disabled|enabled"},
+		{"panda3ds_enable_aac", "Enable AAC audio; enabled|disabled"},
+
+		{"panda3ds_ubershader_lighting_override", "Force shadergen when rendering lights; enabled|disabled"},
+		{"panda3ds_ubershader_lighting_override_threshold", "Light threshold for forcing shadergen; 1|2|3|4|5|6|7|8"},
 		{"panda3ds_use_virtual_sd", "Enable virtual SD card; enabled|disabled"},
 		{"panda3ds_write_protect_virtual_sd", "Write protect virtual SD card; disabled|enabled"},
 		{"panda3ds_battery_level", "Battery percentage; 5|10|20|30|50|70|90|100"},
 		{"panda3ds_use_charger", "Charger plugged; enabled|disabled"},
-		{"panda3ds_ubershader_lighting_override", "Force shadergen when rendering lights; enabled|disabled"},
-		{"panda3ds_ubershader_lighting_override_threshold", "Light threshold for forcing shadergen; 1|2|3|4|5|6|7|8"},
 		{nullptr, nullptr},
 	};
 
@@ -194,6 +198,10 @@ static void configUpdate() {
 	config.batteryPercentage = fetchVariableRange("panda3ds_battery_level", 5, 100);
 	config.dspType = Audio::DSPCore::typeFromString(fetchVariable("panda3ds_dsp_emulation", "null"));
 	config.audioEnabled = fetchVariableBool("panda3ds_use_audio", false);
+	config.aacEnabled = fetchVariableBool("panda3ds_enable_aac", true);
+	config.audioDeviceConfig.muteAudio = fetchVariableBool("panda3ds_mute_audio", false);
+	config.audioDeviceConfig.volumeRaw = float(fetchVariableRange("panda3ds_audio_volume", 0, 200)) / 100.0f;
+
 	config.sdCardInserted = fetchVariableBool("panda3ds_use_virtual_sd", true);
 	config.sdWriteProtected = fetchVariableBool("panda3ds_write_protect_virtual_sd", false);
 	config.accurateShaderMul = fetchVariableBool("panda3ds_accurate_shader_mul", false);