diff --git a/.github/workflows/build-plugin.yml b/.github/workflows/build-plugin.yml
index d2c6638..15ee850 100644
--- a/.github/workflows/build-plugin.yml
+++ b/.github/workflows/build-plugin.yml
@@ -17,7 +17,7 @@ jobs:
         config:
         - {
             name: Linux,
-            os: ubuntu-latest,
+            os: ubuntu-16.04,
             prepare-os: sudo apt install -y libglu-dev
           }
         - {
@@ -68,7 +68,7 @@ jobs:
     # only create a release if a tag was created that is called e.g. v1.2.3
     # see also https://vcvrack.com/manual/Manifest#version
     if: startsWith(github.ref, 'refs/tags/v')
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-16.04
     needs: build
     steps:
       - uses: actions/checkout@v2
diff --git a/src/ABC.cpp b/src/ABC.cpp
index 4f8dc04..9e2387d 100644
--- a/src/ABC.cpp
+++ b/src/ABC.cpp
@@ -1,6 +1,7 @@
 #include "plugin.hpp"
 #include "Common.hpp"
-#include "simd_input.hpp"
+
+using simd::float_4;
 
 template <typename T>
 static T clip4(T x) {
@@ -54,126 +55,112 @@ struct ABC : Module {
 		configParam(C2_LEVEL_PARAM, -1.0, 1.0, 0.0, "C2 Level");
 	}
 
-	void process(const ProcessArgs &args) override {
-
-		simd::float_4 a1[4] = {};
-		simd::float_4 b1[4] = {};
-		simd::float_4 c1[4] = {};
-		simd::float_4 out1[4];
-
-		simd::float_4 a2[4] = {};
-		simd::float_4 b2[4] = {};
-		simd::float_4 c2[4] = {};
-		simd::float_4 out2[4];
-
-		int channels_1 = 1;
-		int channels_2 = 1;
-
-		memset(out1, 0, sizeof(out1));
-		memset(out2, 0, sizeof(out2));
+	int processSection(simd::float_4* out, InputIds inputA, InputIds inputB, InputIds inputC,
+	                   ParamIds levelB, ParamIds levelC) {
 
-		// process upper section
-
-		if (outputs[OUT1_OUTPUT].isConnected() || outputs[OUT2_OUTPUT].isConnected()) {
+		float_4 inA[4] = {0.f};
+		float_4 inB[4] = {0.f};
+		float_4 inC[4] = {0.f};
 
-			int channels_A1 = inputs[A1_INPUT].getChannels();
-			int channels_B1 = inputs[B1_INPUT].getChannels();
-			int channels_C1 = inputs[C1_INPUT].getChannels();
+		int channelsA = inputs[inputA].getChannels();
+		int channelsB = inputs[inputB].getChannels();
+		int channelsC = inputs[inputC].getChannels();
 
-			channels_1 = std::max(channels_1, channels_A1);
-			channels_1 = std::max(channels_1, channels_B1);
-			channels_1 = std::max(channels_1, channels_C1);
+		// this sets the number of active engines (according to polyphony standard)
+		// NOTE: A*B + C has the number of active engines set by any one of the three inputs
+		int activeEngines = std::max(1, channelsA);
+		activeEngines = std::max(activeEngines, channelsB);
+		activeEngines = std::max(activeEngines, channelsC);
 
-			float mult_B1 = (2.f / 5.f) * exponentialBipolar80Pade_5_4(params[B1_LEVEL_PARAM].getValue());
-			float mult_C1 = exponentialBipolar80Pade_5_4(params[C1_LEVEL_PARAM].getValue());
+		float mult_B = (2.f / 5.f) * exponentialBipolar80Pade_5_4(params[levelB].getValue());
+		float mult_C = exponentialBipolar80Pade_5_4(params[levelC].getValue());
 
-			if (inputs[A1_INPUT].isConnected())
-				load_input(inputs[A1_INPUT], a1, channels_A1);
-			else
-				memset(a1, 0, sizeof(a1));
-
-			if (inputs[B1_INPUT].isConnected()) {
-				load_input(inputs[B1_INPUT], b1, channels_B1);
-				for (int c = 0; c < channels_1; c += 4)
-					b1[c / 4] *= simd::float_4(mult_B1);
+		if (inputs[inputA].isConnected()) {
+			// if monophonic, broadcast to number of active engines
+			if (channelsA == 1) {
+				for (int c = 0; c < activeEngines; c += 4)
+					inA[c / 4] = float_4(inputs[inputA].getVoltage());
 			}
 			else {
-				for (int c = 0; c < channels_1; c += 4)
-					b1[c / 4] = simd::float_4(5.f * mult_B1);
+				for (int c = 0; c < channelsA; c += 4)
+					inA[c / 4] = inputs[inputA].getVoltageSimd<float_4>(c);
 			}
+		}
 
-			if (inputs[C1_INPUT].isConnected()) {
-				load_input(inputs[C1_INPUT], c1, channels_C1);
-				for (int c = 0; c < channels_1; c += 4)
-					c1[c / 4] *= simd::float_4(mult_C1);
+		if (inputs[inputB].isConnected()) {
+			// if monophonic, broadcast to number of active engines
+			if (channelsB == 1) {
+				for (int c = 0; c < activeEngines; c += 4)
+					inB[c / 4] = float_4(inputs[inputB].getVoltage());
 			}
 			else {
-				for (int c = 0; c < channels_1; c += 4)
-					c1[c / 4] = simd::float_4(10.f * mult_C1);
+				for (int c = 0; c < channelsB; c += 4)
+					inB[c / 4] = inputs[inputB].getVoltageSimd<float_4>(c);
 			}
-
-			for (int c = 0; c < channels_1; c += 4)
-				out1[c / 4] = clip4(a1[c / 4] * b1[c / 4] + c1[c / 4]);
+			for (int c = 0; c < activeEngines; c += 4)
+				inB[c / 4] *= mult_B;
+		}
+		else {
+			for (int c = 0; c < activeEngines; c += 4)
+				inB[c / 4] = 5.f * mult_B;
 		}
 
-		// process lower section
-
-		if (outputs[OUT2_OUTPUT].isConnected()) {
-			int channels_A2 = inputs[A2_INPUT].getChannels();
-			int channels_B2 = inputs[B2_INPUT].getChannels();
-			int channels_C2 = inputs[C2_INPUT].getChannels();
-
-			channels_2 = std::max(channels_2, channels_A2);
-			channels_2 = std::max(channels_2, channels_B2);
-			channels_2 = std::max(channels_2, channels_C2);
-
-			float mult_B2 = (2.f / 5.f) * exponentialBipolar80Pade_5_4(params[B2_LEVEL_PARAM].getValue());
-			float mult_C2 = exponentialBipolar80Pade_5_4(params[C2_LEVEL_PARAM].getValue());
-
-			if (inputs[A2_INPUT].isConnected())
-				load_input(inputs[A2_INPUT], a2, channels_A2);
-			else
-				memset(a2, 0, sizeof(a2));
-
-			if (inputs[B2_INPUT].isConnected()) {
-				load_input(inputs[B2_INPUT], b2, channels_B2);
-				for (int c = 0; c < channels_2; c += 4)
-					b2[c / 4] *= simd::float_4(mult_B2);
+		if (inputs[inputC].isConnected()) {
+			// if monophonic, broadcast to number of active engines
+			if (channelsC == 1) {
+				for (int c = 0; c < activeEngines; c += 4)
+					inC[c / 4] = float_4(inputs[inputC].getVoltage());
 			}
 			else {
-				for (int c = 0; c < channels_2; c += 4)
-					b2[c / 4] = simd::float_4(5.f * mult_B2);
+				for (int c = 0; c < channelsC; c += 4)
+					inC[c / 4] = inputs[inputC].getVoltageSimd<float_4>(c);
 			}
 
-			if (inputs[C2_INPUT].isConnected()) {
-				load_input(inputs[C2_INPUT], c2, channels_C2);
-				for (int c = 0; c < channels_2; c += 4)
-					c2[c / 4] *= simd::float_4(mult_C2);
-			}
-			else {
-				for (int c = 0; c < channels_2; c += 4)
-					c2[c / 4] = simd::float_4(10.f * mult_C2);
-			}
+			for (int c = 0; c < activeEngines; c += 4)
+				inC[c / 4] *= mult_C;
+		}
+		else {
+			for (int c = 0; c < activeEngines; c += 4)
+				inC[c / 4] = float_4(10.f * mult_C);
+		}
+
+		for (int c = 0; c < activeEngines; c += 4)
+			out[c / 4] = clip4(inA[c / 4] * inB[c / 4] + inC[c / 4]);
+
+		return activeEngines;
+	}
+
+	void process(const ProcessArgs& args) override {
 
-			for (int c = 0; c < channels_2; c += 4)
-				out2[c / 4] = clip4(a2[c / 4] * b2[c / 4] + c2[c / 4]);
-		};
+		// process upper section
+		float_4 out1[4] = {0.f};
+		int activeEngines1 = 1;
+		if (outputs[OUT1_OUTPUT].isConnected() || outputs[OUT2_OUTPUT].isConnected()) {
+			activeEngines1 = processSection(out1, A1_INPUT, B1_INPUT, C1_INPUT, B1_LEVEL_PARAM, C1_LEVEL_PARAM);
+		}
+
+		float_4 out2[4] = {0.f};
+		int activeEngines2 = 1;
+		// process lower section
+		if (outputs[OUT2_OUTPUT].isConnected()) {
+			activeEngines2 = processSection(out2, A2_INPUT, B2_INPUT, C2_INPUT, B2_LEVEL_PARAM, C2_LEVEL_PARAM);
+		}
 
 		// Set outputs
 		if (outputs[OUT1_OUTPUT].isConnected()) {
-			outputs[OUT1_OUTPUT].setChannels(channels_1);
-			for (int c = 0; c < channels_1; c += 4)
+			outputs[OUT1_OUTPUT].setChannels(activeEngines1);
+			for (int c = 0; c < activeEngines1; c += 4)
 				out1[c / 4].store(outputs[OUT1_OUTPUT].getVoltages(c));
 		}
-		else {
-			for (int c = 0; c < channels_1; c += 4)
+		else if (outputs[OUT2_OUTPUT].isConnected()) {
+
+			for (int c = 0; c < activeEngines1; c += 4)
 				out2[c / 4] += out1[c / 4];
-			channels_2 = std::max(channels_1, channels_2);
-		}
 
-		if (outputs[OUT2_OUTPUT].isConnected()) {
-			outputs[OUT2_OUTPUT].setChannels(channels_2);
-			for (int c = 0; c < channels_2; c += 4)
+			activeEngines2 = std::max(activeEngines1, activeEngines2);
+
+			outputs[OUT2_OUTPUT].setChannels(activeEngines2);
+			for (int c = 0; c < activeEngines2; c += 4)
 				out2[c / 4].store(outputs[OUT2_OUTPUT].getVoltages(c));
 		}
 
@@ -182,7 +169,7 @@ struct ABC : Module {
 		float light_1;
 		float light_2;
 
-		if (channels_1 == 1) {
+		if (activeEngines1 == 1) {
 			light_1 = out1[0].s[0];
 			lights[OUT1_LIGHT + 0].setSmoothBrightness(light_1 / 5.f, args.sampleTime);
 			lights[OUT1_LIGHT + 1].setSmoothBrightness(-light_1 / 5.f, args.sampleTime);
@@ -195,7 +182,7 @@ struct ABC : Module {
 			lights[OUT1_LIGHT + 2].setBrightness(light_1);
 		}
 
-		if (channels_2 == 1) {
+		if (activeEngines2 == 1) {
 			light_2 = out2[0].s[0];
 			lights[OUT2_LIGHT + 0].setSmoothBrightness(light_2 / 5.f, args.sampleTime);
 			lights[OUT2_LIGHT + 1].setSmoothBrightness(-light_2 / 5.f, args.sampleTime);
@@ -212,7 +199,7 @@ struct ABC : Module {
 
 
 struct ABCWidget : ModuleWidget {
-	ABCWidget(ABC *module) {
+	ABCWidget(ABC* module) {
 		setModule(module);
 		setPanel(APP->window->loadSvg(asset::plugin(pluginInstance, "res/ABC.svg")));
 
@@ -239,4 +226,4 @@ struct ABCWidget : ModuleWidget {
 };
 
 
-Model *modelABC = createModel<ABC, ABCWidget>("ABC");
+Model* modelABC = createModel<ABC, ABCWidget>("ABC");
diff --git a/src/DualAtenuverter.cpp b/src/DualAtenuverter.cpp
index e24e2fd..b32fbb4 100644
--- a/src/DualAtenuverter.cpp
+++ b/src/DualAtenuverter.cpp
@@ -1,6 +1,5 @@
 #include "plugin.hpp"
 #include "Common.hpp"
-#include "simd_input.hpp"
 
 struct DualAtenuverter : Module {
 	enum ParamIds {
@@ -34,7 +33,7 @@ struct DualAtenuverter : Module {
 		configParam(OFFSET2_PARAM, -10.0, 10.0, 0.0, "Ch 2 offset", " V");
 	}
 
-	void process(const ProcessArgs &args) override {
+	void process(const ProcessArgs& args) override {
 		using simd::float_4;
 
 		float_4 out1[4];
@@ -52,10 +51,10 @@ struct DualAtenuverter : Module {
 		float offset2 = params[OFFSET2_PARAM].getValue();
 
 		for (int c = 0; c < channels1; c += 4) {
-			out1[c / 4] = clamp(float_4::load(inputs[IN1_INPUT].getVoltages(c)) * att1 + offset1, -10.f, 10.f);
+			out1[c / 4] = clamp(inputs[IN1_INPUT].getVoltageSimd<float_4>(c) * att1 + offset1, -10.f, 10.f);
 		}
 		for (int c = 0; c < channels2; c += 4) {
-			out2[c / 4] = clamp(float_4::load(inputs[IN2_INPUT].getVoltages(c)) * att2 + offset2, -10.f, 10.f);
+			out2[c / 4] = clamp(inputs[IN2_INPUT].getVoltageSimd<float_4>(c) * att2 + offset2, -10.f, 10.f);
 		}
 
 		outputs[OUT1_OUTPUT].setChannels(channels1);
@@ -97,7 +96,7 @@ struct DualAtenuverter : Module {
 
 
 struct DualAtenuverterWidget : ModuleWidget {
-	DualAtenuverterWidget(DualAtenuverter *module) {
+	DualAtenuverterWidget(DualAtenuverter* module) {
 		setModule(module);
 		setPanel(APP->window->loadSvg(asset::plugin(pluginInstance, "res/DualAtenuverter.svg")));
 
@@ -121,4 +120,4 @@ struct DualAtenuverterWidget : ModuleWidget {
 };
 
 
-Model *modelDualAtenuverter = createModel<DualAtenuverter, DualAtenuverterWidget>("DualAtenuverter");
+Model* modelDualAtenuverter = createModel<DualAtenuverter, DualAtenuverterWidget>("DualAtenuverter");
\ No newline at end of file
diff --git a/src/EvenVCO.cpp b/src/EvenVCO.cpp
index ec69ea7..ec708b3 100644
--- a/src/EvenVCO.cpp
+++ b/src/EvenVCO.cpp
@@ -1,7 +1,8 @@
 #include "plugin.hpp"
-#include "simd_input.hpp"
 #include "Common.hpp"
 
+using simd::float_4;
+
 struct EvenVCO : Module {
 	enum ParamIds {
 		OCTAVE_PARAM,
@@ -26,8 +27,8 @@ struct EvenVCO : Module {
 		NUM_OUTPUTS
 	};
 
-	simd::float_4 phase[4];
-	simd::float_4 tri[4];
+	float_4 phase[4];
+	float_4 tri[4];
 
 	/** The value of the last sync input */
 	float sync = 0.0;
@@ -51,23 +52,14 @@ struct EvenVCO : Module {
 		configParam(PWM_PARAM, -1.0, 1.0, 0.0, "Pulse width");
 
 		for (int i = 0; i < 4; i++) {
-			phase[i] = simd::float_4(0.0f);
-			tri[i] = simd::float_4(0.0f);
+			phase[i] = float_4(0.0f);
+			tri[i] = float_4(0.0f);
 		}
 		for (int c = 0; c < PORT_MAX_CHANNELS; c++)
 			halfPhase[c] = false;
 	}
 
-	void process(const ProcessArgs &args) override {
-		simd::float_4 pitch[4];
-		simd::float_4 pitch_1[4];
-		simd::float_4 pitch_2[4];
-		simd::float_4 pitch_fm[4];
-		simd::float_4 freq[4];
-		simd::float_4 pw[4];
-		simd::float_4 pwm[4];
-		simd::float_4 deltaPhase[4];
-		simd::float_4 oldPhase[4];
+	void process(const ProcessArgs& args) override {
 
 		int channels_pitch1 = inputs[PITCH1_INPUT].getChannels();
 		int channels_pitch2 = inputs[PITCH2_INPUT].getChannels();
@@ -81,56 +73,79 @@ struct EvenVCO : Module {
 		float pitch_0 = 1.f + std::round(params[OCTAVE_PARAM].getValue()) + params[TUNE_PARAM].getValue() / 12.f;
 
 		// Compute frequency, pitch is 1V/oct
-
+		float_4 pitch[4];
 		for (int c = 0; c < channels; c += 4)
-			pitch[c / 4] = simd::float_4(pitch_0);
+			pitch[c / 4] = float_4(pitch_0);
 
 		if (inputs[PITCH1_INPUT].isConnected()) {
-			load_input(inputs[PITCH1_INPUT], pitch_1, channels_pitch1);
-			for (int c = 0; c < channels_pitch1; c += 4)
-				pitch[c / 4] += pitch_1[c / 4];
+			// if pitch_1 monophonic, broadcast
+			if (channels_pitch1 == 1) {
+				for (int c = 0; c < channels; c += 4)
+					pitch[c / 4] += float_4(inputs[PITCH1_INPUT].getVoltage());
+			}
+			else {
+				for (int c = 0; c < std::min(channels, channels_pitch1); c += 4)
+					pitch[c / 4] += inputs[PITCH1_INPUT].getVoltageSimd<float_4>(c);
+			}
 		}
 
 		if (inputs[PITCH2_INPUT].isConnected()) {
-			load_input(inputs[PITCH2_INPUT], pitch_2, channels_pitch2);
-			for (int c = 0; c < channels_pitch2; c += 4)
-				pitch[c / 4] += pitch_2[c / 4];
+			// if pitch_2 monophonic, broadcast
+			if (channels_pitch2 == 1) {
+				for (int c = 0; c < channels; c += 4)
+					pitch[c / 4] += float_4(inputs[PITCH2_INPUT].getVoltage());
+			}
+			else {
+				for (int c = 0; c < std::min(channels, channels_pitch2); c += 4)
+					pitch[c / 4] += inputs[PITCH2_INPUT].getVoltageSimd<float_4>(c);
+			}
 		}
 
 		if (inputs[FM_INPUT].isConnected()) {
-			load_input(inputs[FM_INPUT], pitch_fm, channels_fm);
-			for (int c = 0; c < channels_fm; c += 4)
-				pitch[c / 4] += pitch_fm[c / 4] / 4.f;
+			// if FM is monophonic, broadcast
+			if (channels_fm == 1) {
+				for (int c = 0; c < channels; c += 4)
+					pitch[c / 4] += float_4(inputs[FM_INPUT].getVoltage() / 4.f);
+			}
+			else {
+				for (int c = 0; c < std::min(channels, channels_fm); c += 4)
+					pitch[c / 4] += inputs[FM_INPUT].getVoltageSimd<float_4>(c) / 4.f;
+			}
 		}
 
+		float_4 freq[4];
 		for (int c = 0; c < channels; c += 4) {
 			freq[c / 4] = dsp::FREQ_C4 * simd::pow(2.f, pitch[c / 4]);
 			freq[c / 4] = clamp(freq[c / 4], 0.f, 20000.f);
 		}
 
-
 		// Pulse width
-
 		float pw_0 = params[PWM_PARAM].getValue();
-
+		float_4 pw[4];
 		for (int c = 0; c < channels; c += 4)
-			pw[c / 4] = simd::float_4(pw_0);
+			pw[c / 4] = float_4(pw_0);
 
 		if (inputs[PWM_INPUT].isConnected()) {
-			load_input(inputs[PWM_INPUT], pwm, channels_pwm);
-			for (int c = 0; c < channels_pwm; c += 4)
-				pw[c / 4] += pwm[c / 4] / 5.f;
+			if (channels_pwm == 1) {
+				for (int c = 0; c < channels; c += 4)
+					pw[c / 4] += float_4(inputs[PWM_INPUT].getVoltage() / 5.f);
+			}
+			else {
+				for (int c = 0; c < std::min(channels, channels_pwm); c += 4)
+					pw[c / 4] += inputs[PWM_INPUT].getVoltageSimd<float_4>(c) / 5.f;
+			}
 		}
 
-		const simd::float_4 minPw_4 = simd::float_4(0.05f);
-		const simd::float_4 m_one_4 = simd::float_4(-1.0f);
-		const simd::float_4 one_4 = simd::float_4(1.0f);
-
+		const float_4 minPw_4 = float_4(0.05f);
+		const float_4 m_one_4 = float_4(-1.0f);
+		const float_4 one_4 = float_4(1.0f);
+		float_4 deltaPhase[4];
+		float_4 oldPhase[4];
 		for (int c = 0; c < channels; c += 4) {
 			pw[c / 4] = rescale(clamp(pw[c / 4], m_one_4, one_4), m_one_4, one_4, minPw_4, one_4 - minPw_4);
 
 			// Advance phase
-			deltaPhase[c / 4] = clamp(freq[c / 4] * args.sampleTime, simd::float_4(1e-6f), simd::float_4(0.5f));
+			deltaPhase[c / 4] = clamp(freq[c / 4] * args.sampleTime, float_4(1e-6f), float_4(0.5f));
 			oldPhase[c / 4] = phase[c / 4];
 			phase[c / 4] += deltaPhase[c / 4];
 		}
@@ -163,19 +178,19 @@ struct EvenVCO : Module {
 			}
 		}
 
-		simd::float_4 triSquareMinBlepOut[4];
-		simd::float_4 doubleSawMinBlepOut[4];
-		simd::float_4 sawMinBlepOut[4];
-		simd::float_4 squareMinBlepOut[4];
+		float_4 triSquareMinBlepOut[4];
+		float_4 doubleSawMinBlepOut[4];
+		float_4 sawMinBlepOut[4];
+		float_4 squareMinBlepOut[4];
 
-		simd::float_4 triSquare[4];
-		simd::float_4 sine[4];
-		simd::float_4 doubleSaw[4];
+		float_4 triSquare[4];
+		float_4 sine[4];
+		float_4 doubleSaw[4];
 
-		simd::float_4 even[4];
-		simd::float_4 saw[4];
-		simd::float_4 square[4];
-		simd::float_4 triOut[4];
+		float_4 even[4];
+		float_4 saw[4];
+		float_4 square[4];
+		float_4 triOut[4];
 
 		for (int c = 0; c < channels; c++) {
 			triSquareMinBlepOut[c / 4].s[c % 4] = triSquareMinBlep[c].process();
@@ -231,7 +246,7 @@ struct EvenVCO : Module {
 
 
 struct EvenVCOWidget : ModuleWidget {
-	EvenVCOWidget(EvenVCO *module) {
+	EvenVCOWidget(EvenVCO* module) {
 		setModule(module);
 		setPanel(APP->window->loadSvg(asset::plugin(pluginInstance, "res/EvenVCO.svg")));
 
@@ -260,4 +275,4 @@ struct EvenVCOWidget : ModuleWidget {
 };
 
 
-Model *modelEvenVCO = createModel<EvenVCO, EvenVCOWidget>("EvenVCO");
+Model* modelEvenVCO = createModel<EvenVCO, EvenVCOWidget>("EvenVCO");
diff --git a/src/Mixer.cpp b/src/Mixer.cpp
index 8ea8c2c..f8e96ff 100644
--- a/src/Mixer.cpp
+++ b/src/Mixer.cpp
@@ -1,6 +1,7 @@
 #include "plugin.hpp"
 #include "Common.hpp"
-#include "simd_input.hpp"
+
+using simd::float_4;
 
 struct Mixer : Module {
 	enum ParamIds {
@@ -37,7 +38,7 @@ struct Mixer : Module {
 		configParam(CH4_PARAM, 0.0, 1.0, 0.0, "Ch 4 level", "%", 0, 100);
 	}
 
-	void process(const ProcessArgs &args) override {
+	void process(const ProcessArgs& args) override {
 		int channels1 = inputs[IN1_INPUT].getChannels();
 		int channels2 = inputs[IN2_INPUT].getChannels();
 		int channels3 = inputs[IN3_INPUT].getChannels();
@@ -49,37 +50,35 @@ struct Mixer : Module {
 		out_channels = std::max(out_channels, channels3);
 		out_channels = std::max(out_channels, channels4);
 
-		simd::float_4 mult1 = simd::float_4(params[CH1_PARAM].getValue());
-		simd::float_4 mult2 = simd::float_4(params[CH2_PARAM].getValue());
-		simd::float_4 mult3 = simd::float_4(params[CH3_PARAM].getValue());
-		simd::float_4 mult4 = simd::float_4(params[CH4_PARAM].getValue());
+		float_4 mult1 = float_4(params[CH1_PARAM].getValue());
+		float_4 mult2 = float_4(params[CH2_PARAM].getValue());
+		float_4 mult3 = float_4(params[CH3_PARAM].getValue());
+		float_4 mult4 = float_4(params[CH4_PARAM].getValue());
 
-		simd::float_4 out[4];
+		float_4 out[4];
 
 		std::memset(out, 0, sizeof(out));
 
-
 		if (inputs[IN1_INPUT].isConnected()) {
 			for (int c = 0; c < channels1; c += 4)
-				out[c / 4] += simd::float_4::load(inputs[IN1_INPUT].getVoltages(c)) * mult1;
+				out[c / 4] += inputs[IN1_INPUT].getVoltageSimd<float_4>(c) * mult1;
 		}
 
 		if (inputs[IN2_INPUT].isConnected()) {
 			for (int c = 0; c < channels2; c += 4)
-				out[c / 4] += simd::float_4::load(inputs[IN2_INPUT].getVoltages(c)) * mult2;
+				out[c / 4] += inputs[IN2_INPUT].getVoltageSimd<float_4>(c) * mult2;
 		}
 
 		if (inputs[IN3_INPUT].isConnected()) {
 			for (int c = 0; c < channels3; c += 4)
-				out[c / 4] += simd::float_4::load(inputs[IN3_INPUT].getVoltages(c)) * mult3;
+				out[c / 4] += inputs[IN3_INPUT].getVoltageSimd<float_4>(c) * mult3;
 		}
 
 		if (inputs[IN4_INPUT].isConnected()) {
 			for (int c = 0; c < channels4; c += 4)
-				out[c / 4] += simd::float_4::load(inputs[IN4_INPUT].getVoltages(c)) * mult4;
+				out[c / 4] += inputs[IN4_INPUT].getVoltageSimd<float_4>(c) * mult4;
 		}
 
-
 		outputs[OUT1_OUTPUT].setChannels(out_channels);
 		outputs[OUT2_OUTPUT].setChannels(out_channels);
 
@@ -105,13 +104,12 @@ struct Mixer : Module {
 			lights[OUT_NEG_LIGHT].setBrightness(0.0f);
 			lights[OUT_BLUE_LIGHT].setSmoothBrightness(light / 5.f, args.sampleTime);
 		}
-
 	}
 };
 
 
 struct MixerWidget : ModuleWidget {
-	MixerWidget(Mixer *module) {
+	MixerWidget(Mixer* module) {
 		setModule(module);
 		setPanel(APP->window->loadSvg(asset::plugin(pluginInstance, "res/Mixer.svg")));
 
@@ -137,4 +135,4 @@ struct MixerWidget : ModuleWidget {
 };
 
 
-Model *modelMixer = createModel<Mixer, MixerWidget>("Mixer");
+Model* modelMixer = createModel<Mixer, MixerWidget>("Mixer");
diff --git a/src/Rampage.cpp b/src/Rampage.cpp
index 0a4d7df..c4061c2 100644
--- a/src/Rampage.cpp
+++ b/src/Rampage.cpp
@@ -1,16 +1,17 @@
 #include "plugin.hpp"
 #include "Common.hpp"
-#include "simd_input.hpp"
 #include "PulseGenerator_4.hpp"
 
-static simd::float_4 shapeDelta(simd::float_4 delta, simd::float_4 tau, float shape) {
-	simd::float_4 lin = simd::sgn(delta) * 10.f / tau;
+using simd::float_4;
+
+static float_4 shapeDelta(float_4 delta, float_4 tau, float shape) {
+	float_4 lin = simd::sgn(delta) * 10.f / tau;
 	if (shape < 0.f) {
-		simd::float_4 log = simd::sgn(delta) * simd::float_4(40.f) / tau / (simd::fabs(delta) + simd::float_4(1.f));
+		float_4 log = simd::sgn(delta) * float_4(40.f) / tau / (simd::fabs(delta) + float_4(1.f));
 		return simd::crossfade(lin, log, -shape * 0.95f);
 	}
 	else {
-		simd::float_4 exp = M_E * delta / tau;
+		float_4 exp = M_E * delta / tau;
 		return simd::crossfade(lin, exp, shape * 0.90f);
 	}
 }
@@ -75,10 +76,10 @@ struct Rampage : Module {
 	};
 
 
-	simd::float_4 out[2][4];
-	simd::float_4 gate[2][4]; // use simd __m128 logic instead of bool
+	float_4 out[2][4];
+	float_4 gate[2][4]; // use simd __m128 logic instead of bool
 
-	dsp::TSchmittTrigger<simd::float_4> trigger_4[2][4];
+	dsp::TSchmittTrigger<float_4> trigger_4[2][4];
 	PulseGenerator_4 endOfCyclePulse[2][4];
 
 	// ChannelMask channelMask;
@@ -103,10 +104,10 @@ struct Rampage : Module {
 		std::memset(gate, 0, sizeof(gate));
 	}
 
-	void process(const ProcessArgs &args) override {
+	void process(const ProcessArgs& args) override {
 		int channels_in[2];
 		int channels_trig[2];
-		int channels[2];
+		int channels[2]; 	// the larger of in or trig (per-part)
 
 		// determine number of channels:
 
@@ -122,25 +123,25 @@ struct Rampage : Module {
 			outputs[FALLING_A_OUTPUT + part].setChannels(channels[part]);
 			outputs[EOC_A_OUTPUT + part].setChannels(channels[part]);
 		}
-		int channels_max = std::max(channels[0], channels[1]);
+		
+		// total number of active polyphony engines, accounting for both halves
+		// (channels[0] / channels[1] are the number of active engines per section)
+		const int channels_max = std::max(channels[0], channels[1]);
 
 		outputs[COMPARATOR_OUTPUT].setChannels(channels_max);
 		outputs[MIN_OUTPUT].setChannels(channels_max);
 		outputs[MAX_OUTPUT].setChannels(channels_max);
 
 		// loop over two parts of Rampage:
-
 		for (int part = 0; part < 2; part++) {
 
-			simd::float_4 in[4];
-			simd::float_4 in_trig[4];
-			simd::float_4 expCV[4];
-			simd::float_4 riseCV[4];
-			simd::float_4 fallCV[4];
-			simd::float_4 cycle[4];
+			float_4 in[4];
+			float_4 in_trig[4];
+			float_4 riseCV[4];
+			float_4 fallCV[4];
+			float_4 cycle[4];
 
 			// get parameters:
-
 			float shape = params[SHAPE_A_PARAM + part].getValue();
 			float minTime;
 			switch ((int) params[RANGE_A_PARAM + part].getValue()) {
@@ -155,91 +156,142 @@ struct Rampage : Module {
 					break;
 			}
 
-			simd::float_4 param_rise  = simd::float_4(params[RISE_A_PARAM  + part].getValue() * 10.0f);
-			simd::float_4 param_fall  = simd::float_4(params[FALL_A_PARAM  + part].getValue() * 10.0f);
-			simd::float_4 param_trig  = simd::float_4(params[TRIGG_A_PARAM + part].getValue() * 20.0f);
-			simd::float_4 param_cycle = simd::float_4(params[CYCLE_A_PARAM + part].getValue() * 10.0f);
+			float_4 param_rise  = float_4(params[RISE_A_PARAM  + part].getValue() * 10.0f);
+			float_4 param_fall  = float_4(params[FALL_A_PARAM  + part].getValue() * 10.0f);
+			float_4 param_trig  = float_4(params[TRIGG_A_PARAM + part].getValue() * 20.0f);
+			float_4 param_cycle = float_4(params[CYCLE_A_PARAM + part].getValue() * 10.0f);
 
 			for (int c = 0; c < channels[part]; c += 4) {
 				riseCV[c / 4] = param_rise;
 				fallCV[c / 4] = param_fall;
 				cycle[c / 4] = param_cycle;
 				in_trig[c / 4] = param_trig;
-
 			}
 
-
 			// read inputs:
-
 			if (inputs[IN_A_INPUT + part].isConnected()) {
-				load_input(inputs[IN_A_INPUT + part], in, channels_in[part]);
-				// channelMask.apply_all(in, channels_in[part]);
+				// if IN_<A,B>_INPUT is monophonic, broadcast to the active number of engines (channels[part])
+				if (inputs[IN_A_INPUT + part].getChannels() == 1) {
+					for (int c = 0; c < channels[part]; c += 4)
+						in[c / 4] = float_4(inputs[IN_A_INPUT + part].getVoltage());
+				}
+				else {
+					for (int c = 0; c < channels[part]; c += 4)
+						in[c / 4] = inputs[IN_A_INPUT + part].getVoltageSimd<float_4>(c);
+				}
 			}
 			else {
 				std::memset(in, 0, sizeof(in));
 			}
 
 			if (inputs[TRIGG_A_INPUT + part].isConnected()) {
-				add_input(inputs[TRIGG_A_INPUT + part], in_trig, channels_trig[part]);
-				// channelMask.apply_all(in_trig, channels_trig[part]);
+				// if TRIGG_<A,B>_INPUT is monophonic, broadcast to the active number of engines (channels[part])
+				if (inputs[TRIGG_A_INPUT + part].getChannels() == 1) {
+					for (int c = 0; c < channels[part]; c += 4)
+						in_trig[c / 4] += float_4(inputs[TRIGG_A_INPUT + part].getVoltage());
+				}
+				else {
+					for (int c = 0; c < channels[part]; c += 4)
+						in_trig[c / 4] += inputs[TRIGG_A_INPUT + part].getVoltageSimd<float_4>(c);
+				}
 			}
 
 			if (inputs[EXP_CV_A_INPUT + part].isConnected()) {
-				load_input(inputs[EXP_CV_A_INPUT + part], expCV, channels[part]);
+				float_4 expCV[4];
+				int expCVChannels = inputs[EXP_CV_A_INPUT + part].getChannels();
+				// if EXP_CV_<A,B>_INPUT is monophonic, broadcast to the active number of engines (channels[part])
+				if (expCVChannels == 1) {
+					for (int c = 0; c < channels[part]; c += 4)
+						expCV[c / 4] = float_4(inputs[EXP_CV_A_INPUT + part].getVoltage());
+				}
+				else {
+					// otherwise read in the polyphonic expCV data, either to the number of active engines (channels[part])
+					// or the number of channels of expCV, whichever is smaller
+					for (int c = 0; c < std::min(channels[part], expCVChannels); c += 4)
+						expCV[c / 4] = inputs[EXP_CV_A_INPUT + part].getVoltageSimd<float_4>(c);
+				}
+
 				for (int c = 0; c < channels[part]; c += 4) {
 					riseCV[c / 4] -= expCV[c / 4];
 					fallCV[c / 4] -= expCV[c / 4];
 				}
 			}
 
-			add_input(inputs[RISE_CV_A_INPUT + part], riseCV, channels[part]);
-			add_input(inputs[FALL_CV_A_INPUT + part], fallCV, channels[part]);
-			add_input(inputs[CYCLE_A_INPUT + part], cycle, channels[part]);
-			// channelMask.apply(cycle, channels[part]); // check whether this is necessary
+			const int riseCVChannels = inputs[RISE_CV_A_INPUT + part].getChannels();
+			// if EXP_CV_<A,B>_INPUT is monophonic, broadcast to the active number of engines (channels[part])
+			if (riseCVChannels == 1) {
+				for (int c = 0; c < channels[part]; c += 4)
+					riseCV[c / 4] += float_4(inputs[RISE_CV_A_INPUT + part].getVoltage());
+			}
+			else {
+				// otherwise read in the polyphonic rise CV data, either to the number of active engines (channels[part])
+				// or the number of channels of expCV, whichever is smaller
+				for (int c = 0; c < std::min(channels[part], riseCVChannels); c += 4)
+					riseCV[c / 4] += inputs[RISE_CV_A_INPUT + part].getVoltageSimd<float_4>(c);
+			}
+
+			const int fallCVChannels = inputs[FALL_CV_A_INPUT + part].getChannels();
+			if (fallCVChannels == 1) {
+				for (int c = 0; c < channels[part]; c += 4)
+					fallCV[c / 4] += float_4(inputs[FALL_CV_A_INPUT + part].getVoltage());
+			}
+			else {
+				for (int c = 0; c < std::min(channels[part], fallCVChannels); c += 4)
+					fallCV[c / 4] += inputs[FALL_CV_A_INPUT + part].getVoltageSimd<float_4>(c);
+			}
 
-			// start processing:
+			const int cycleChannels = inputs[CYCLE_A_INPUT + part].getChannels();
+			if (cycleChannels == 1) {
+				for (int c = 0; c < channels[part]; c += 4)
+					cycle[c / 4] += float_4(inputs[CYCLE_A_INPUT + part].getVoltage());
+			}
+			else {
+				for (int c = 0; c < std::min(channels[part], cycleChannels); c += 4)
+					cycle[c / 4] += inputs[CYCLE_A_INPUT + part].getVoltageSimd<float_4>(c);
+			}
 
+			// start processing:
 			for (int c = 0; c < channels[part]; c += 4) {
 
 				// process SchmittTriggers
 
-				simd::float_4 trig_mask = trigger_4[part][c / 4].process(in_trig[c / 4] / 2.0);
-				gate[part][c / 4] = ifelse(trig_mask, simd::float_4::mask(), gate[part][c / 4]);
-				in[c / 4] = ifelse(gate[part][c / 4], simd::float_4(10.0f), in[c / 4]);
+				float_4 trig_mask = trigger_4[part][c / 4].process(in_trig[c / 4] / 2.0);
+				gate[part][c / 4] = ifelse(trig_mask, float_4::mask(), gate[part][c / 4]);
+				in[c / 4] = ifelse(gate[part][c / 4], float_4(10.0f), in[c / 4]);
 
-				simd::float_4 delta = in[c / 4] - out[part][c / 4];
+				float_4 delta = in[c / 4] - out[part][c / 4];
 
 				// rise / fall branching
 
-				simd::float_4 delta_gt_0 = delta > simd::float_4::zero();
-				simd::float_4 delta_lt_0 = delta < simd::float_4::zero();
-				simd::float_4 delta_eq_0 = ~(delta_lt_0 | delta_gt_0);
+				float_4 delta_gt_0 = delta > float_4::zero();
+				float_4 delta_lt_0 = delta < float_4::zero();
+				float_4 delta_eq_0 = ~(delta_lt_0 | delta_gt_0);
 
-				simd::float_4 rateCV = ifelse(delta_gt_0, riseCV[c / 4], simd::float_4::zero());
+				float_4 rateCV = ifelse(delta_gt_0, riseCV[c / 4], float_4::zero());
 				rateCV = ifelse(delta_lt_0, fallCV[c / 4], rateCV);
-				rateCV = clamp(rateCV, simd::float_4::zero(), simd::float_4(10.0f));
+				rateCV = clamp(rateCV, float_4::zero(), float_4(10.0f));
 
-				simd::float_4 rate = minTime * simd::pow(2.0f, rateCV);
+				float_4 rate = minTime * simd::pow(2.0f, rateCV);
 
 				out[part][c / 4] += shapeDelta(delta, rate, shape) * args.sampleTime;
 
-				simd::float_4 rising  = (in[c / 4] - out[part][c / 4]) > simd::float_4(1e-3);
-				simd::float_4 falling = (in[c / 4] - out[part][c / 4]) < simd::float_4(-1e-3);
-				simd::float_4 end_of_cycle = simd::andnot(falling, delta_lt_0);
+				float_4 rising  = (in[c / 4] - out[part][c / 4]) > float_4(1e-3);
+				float_4 falling = (in[c / 4] - out[part][c / 4]) < float_4(-1e-3);
+				float_4 end_of_cycle = simd::andnot(falling, delta_lt_0);
 
 				endOfCyclePulse[part][c / 4].trigger(end_of_cycle, 1e-3);
 
-				gate[part][c / 4] = ifelse(simd::andnot(rising, delta_gt_0),                 simd::float_4::zero(), gate[part][c / 4]);
-				gate[part][c / 4] = ifelse(end_of_cycle & (cycle[c / 4] >= simd::float_4(4.0f)), simd::float_4::mask(), gate[part][c / 4]);
-				gate[part][c / 4] = ifelse(delta_eq_0,                                       simd::float_4::zero(), gate[part][c / 4]);
+				gate[part][c / 4] = ifelse(simd::andnot(rising, delta_gt_0),                 float_4::zero(), gate[part][c / 4]);
+				gate[part][c / 4] = ifelse(end_of_cycle & (cycle[c / 4] >= float_4(4.0f)), float_4::mask(), gate[part][c / 4]);
+				gate[part][c / 4] = ifelse(delta_eq_0,                                       float_4::zero(), gate[part][c / 4]);
 
 				out[part][c / 4]  = ifelse(rising | falling, out[part][c / 4], in[c / 4]);
 
-				simd::float_4 out_rising  = ifelse(rising,  simd::float_4(10.0f), simd::float_4::zero());
-				simd::float_4 out_falling = ifelse(falling, simd::float_4(10.0f), simd::float_4::zero());
+				float_4 out_rising  = ifelse(rising,  float_4(10.0f), float_4::zero());
+				float_4 out_falling = ifelse(falling, float_4(10.0f), float_4::zero());
 
-				simd::float_4 pulse = endOfCyclePulse[part][c / 4].process(args.sampleTime);
-				simd::float_4 out_EOC = ifelse(pulse, simd::float_4(10.f), simd::float_4::zero());
+				float_4 pulse = endOfCyclePulse[part][c / 4].process(args.sampleTime);
+				float_4 out_EOC = ifelse(pulse, float_4(10.f), float_4::zero());
 
 				out[part][c / 4].store(outputs[OUT_A_OUTPUT + part].getVoltages(c));
 
@@ -247,7 +299,6 @@ struct Rampage : Module {
 				out_falling.store(outputs[FALLING_A_OUTPUT + part].getVoltages(c));
 				out_EOC.store(outputs[EOC_A_OUTPUT + part].getVoltages(c));
 
-
 			} // for(int c, ...)
 
 			if (channels[part] == 1) {
@@ -281,17 +332,17 @@ struct Rampage : Module {
 
 		for (int c = 0; c < channels_max; c += 4) {
 
-			simd::float_4 a = out[0][c / 4];
-			simd::float_4 b = out[1][c / 4];
+			float_4 a = out[0][c / 4];
+			float_4 b = out[1][c / 4];
 
 			if (balance < 0.5)
 				b *= 2.0f * balance;
 			else if (balance > 0.5)
 				a *= 2.0f * (1.0 - balance);
 
-			simd::float_4 comp    = ifelse(b > a, simd::float_4(10.0f), simd::float_4::zero());
-			simd::float_4 out_min = simd::fmin(a, b);
-			simd::float_4 out_max = simd::fmax(a, b);
+			float_4 comp    = ifelse(b > a, float_4(10.0f), float_4::zero());
+			float_4 out_min = simd::fmin(a, b);
+			float_4 out_max = simd::fmax(a, b);
 
 			comp.store(outputs[COMPARATOR_OUTPUT].getVoltages(c));
 			out_min.store(outputs[MIN_OUTPUT].getVoltages(c));
@@ -325,10 +376,8 @@ struct Rampage : Module {
 };
 
 
-
-
 struct RampageWidget : ModuleWidget {
-	RampageWidget(Rampage *module) {
+	RampageWidget(Rampage* module) {
 		setModule(module);
 		setPanel(APP->window->loadSvg(asset::plugin(pluginInstance, "res/Rampage.svg")));
 
@@ -389,4 +438,4 @@ struct RampageWidget : ModuleWidget {
 };
 
 
-Model *modelRampage = createModel<Rampage, RampageWidget>("Rampage");
+Model* modelRampage = createModel<Rampage, RampageWidget>("Rampage");
diff --git a/src/SlewLimiter.cpp b/src/SlewLimiter.cpp
index f87c735..c47677b 100644
--- a/src/SlewLimiter.cpp
+++ b/src/SlewLimiter.cpp
@@ -1,6 +1,7 @@
 #include "plugin.hpp"
 #include "Common.hpp"
-#include "simd_input.hpp"
+
+using simd::float_4;
 
 struct SlewLimiter : Module {
 	enum ParamIds {
@@ -20,7 +21,7 @@ struct SlewLimiter : Module {
 		NUM_OUTPUTS
 	};
 
-	simd::float_4 out[4];
+	float_4 out[4];
 
 	SlewLimiter() {
 		config(NUM_PARAMS, NUM_INPUTS, NUM_OUTPUTS);
@@ -31,14 +32,14 @@ struct SlewLimiter : Module {
 		memset(out, 0, sizeof(out));
 	}
 
-	void process(const ProcessArgs &args) override {
-
-		simd::float_4 in[4];
-		simd::float_4 riseCV[4];
-		simd::float_4 fallCV[4];
+	void process(const ProcessArgs& args) override {
 
-		int channels = inputs[IN_INPUT].getChannels();
+		float_4 in[4] = {0.f};
+		float_4 riseCV[4] = {0.f};
+		float_4 fallCV[4] = {0.f};
 
+		// this is the number of active polyphony engines, defined by the input
+		int numPolyphonyEngines = inputs[IN_INPUT].getChannels();
 
 		// minimum and std::maximum slopes in volts per second
 		const float slewMin = 0.1;
@@ -46,32 +47,45 @@ struct SlewLimiter : Module {
 		// Amount of extra slew per voltage difference
 		const float shapeScale = 1 / 10.f;
 
-		const simd::float_4 shape = simd::float_4(params[SHAPE_PARAM].getValue());
-		const simd::float_4 param_rise = simd::float_4(params[RISE_PARAM].getValue() * 10.f);
-		const simd::float_4 param_fall = simd::float_4(params[FALL_PARAM].getValue() * 10.f);
-
-		outputs[OUT_OUTPUT].setChannels(channels);
+		const float_4 shape = float_4(params[SHAPE_PARAM].getValue());
+		const float_4 param_rise = float_4(params[RISE_PARAM].getValue() * 10.f);
+		const float_4 param_fall = float_4(params[FALL_PARAM].getValue() * 10.f);
+
+		outputs[OUT_OUTPUT].setChannels(numPolyphonyEngines);
+
+		for (int c = 0; c < numPolyphonyEngines; c += 4) {
+			in[c / 4] = inputs[IN_INPUT].getVoltageSimd<float_4>(c);
+
+			if (inputs[RISE_INPUT].isConnected()) {
+				if (inputs[RISE_INPUT].getChannels() == 1) {
+					riseCV[c / 4] = float_4(inputs[RISE_INPUT].getVoltage());
+				}
+				else {
+					riseCV[c / 4] = inputs[RISE_INPUT].getVoltageSimd<float_4>(c);
+				}
+			}
+			if (inputs[FALL_INPUT].isConnected()) {
+				if (inputs[FALL_INPUT].getChannels() == 1) {
+					fallCV[c / 4] = float_4(inputs[FALL_INPUT].getVoltage());
+				}
+				else {
+					fallCV[c / 4] = inputs[FALL_INPUT].getVoltageSimd<float_4>(c);
+				}
+			}
 
-		load_input(inputs[IN_INPUT], in, channels);
-		load_input(inputs[RISE_INPUT], riseCV, channels);
-		load_input(inputs[FALL_INPUT], fallCV, channels);
-
-		for (int c = 0; c < channels; c += 4) {
 			riseCV[c / 4] += param_rise;
 			fallCV[c / 4] += param_fall;
 
-			simd::float_4 delta = in[c / 4] - out[c / 4];
-
-			simd::float_4 delta_gt_0 = delta > simd::float_4::zero();
-			simd::float_4 delta_lt_0 = delta < simd::float_4::zero();
+			float_4 delta = in[c / 4] - out[c / 4];
+			float_4 delta_gt_0 = delta > float_4::zero();
+			float_4 delta_lt_0 = delta < float_4::zero();
 
-			simd::float_4 rateCV;
-			rateCV = ifelse(delta_gt_0, riseCV[c / 4], simd::float_4::zero());
+			float_4 rateCV;
+			rateCV = ifelse(delta_gt_0, riseCV[c / 4], float_4::zero());
 			rateCV = ifelse(delta_lt_0, fallCV[c / 4], rateCV) * 0.1f;
 
-			simd::float_4 pm_one = simd::sgn(delta);
-
-			simd::float_4 slew = slewMax * simd::pow(simd::float_4(slewMin / slewMax), rateCV);
+			float_4 pm_one = simd::sgn(delta);
+			float_4 slew = slewMax * simd::pow(float_4(slewMin / slewMax), rateCV);
 
 			out[c / 4] += slew * simd::crossfade(pm_one, shapeScale * delta, shape) * args.sampleTime;
 			out[c / 4] = ifelse(delta_gt_0 & (out[c / 4] > in[c / 4]), in[c / 4], out[c / 4]);
@@ -84,7 +98,7 @@ struct SlewLimiter : Module {
 
 
 struct SlewLimiterWidget : ModuleWidget {
-	SlewLimiterWidget(::SlewLimiter *module) {
+	SlewLimiterWidget(SlewLimiter* module) {
 		setModule(module);
 		setPanel(APP->window->loadSvg(asset::plugin(pluginInstance, "res/SlewLimiter.svg")));
 
@@ -105,4 +119,4 @@ struct SlewLimiterWidget : ModuleWidget {
 };
 
 
-Model *modelSlewLimiter = createModel<::SlewLimiter, SlewLimiterWidget>("SlewLimiter");
+Model* modelSlewLimiter = createModel<::SlewLimiter, SlewLimiterWidget>("SlewLimiter");
diff --git a/src/simd_input.hpp b/src/simd_input.hpp
deleted file mode 100644
index 4b1265f..0000000
--- a/src/simd_input.hpp
+++ /dev/null
@@ -1,28 +0,0 @@
-#pragma once
-
-#include "rack.hpp"
-
-
-inline void load_input(Input &in, simd::float_4 *v, int numChannels) {
-	int inChannels = in.getChannels();
-	if (inChannels == 1) {
-		for (int i = 0; i < numChannels; i++)
-			v[i] = simd::float_4(in.getVoltage());
-	}
-	else {
-		for (int c = 0; c < inChannels; c += 4)
-			v[c / 4] = simd::float_4::load(in.getVoltages(c));
-	}
-}
-
-inline void add_input(Input &in, simd::float_4 *v, int numChannels) {
-	int inChannels = in.getChannels();
-	if (inChannels == 1) {
-		for (int i = 0; i < numChannels; i++)
-			v[i] += simd::float_4(in.getVoltage());
-	}
-	else {
-		for (int c = 0; c < inChannels; c += 4)
-			v[c / 4] += simd::float_4::load(in.getVoltages(c));
-	}
-}