diff --git a/benchmarks/BM_mathfuns.cpp b/benchmarks/BM_mathfuns.cpp
index aa11b966..6cf78e9a 100644
--- a/benchmarks/BM_mathfuns.cpp
+++ b/benchmarks/BM_mathfuns.cpp
@@ -45,96 +45,6 @@ BENCHMARK_DEFINE_F(MyFixture, Dummy)
     }
 }
 
-BENCHMARK_DEFINE_F(MyFixture, ScalarExp)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::exp<float, false>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, SIMDExp)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::exp<float, true>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, ScalarExp_Unaligned)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::exp<float, false>(absl::MakeSpan(source).subspan(1), absl::MakeSpan(result).subspan(1));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, SIMDExp_Unaligned)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::exp<float, true>(absl::MakeSpan(source).subspan(1), absl::MakeSpan(result).subspan(1));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, ScalarLog)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::log<float, false>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, SIMDLog)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::log<float, true>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, ScalarSin)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::sin<float, false>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, SIMDSin)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::sin<float, true>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, ScalarCos)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::cos<float, false>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
-BENCHMARK_DEFINE_F(MyFixture, SIMDCos)
-(benchmark::State& state)
-{
-    for (auto _ : state) {
-        sfz::cos<float, true>(source, absl::MakeSpan(result));
-        benchmark::DoNotOptimize(result);
-    }
-}
-
 BENCHMARK_DEFINE_F(MyFixture, ScalarLibmFloorLog2)
 (benchmark::State& state)
 {
@@ -159,16 +69,6 @@ BENCHMARK_DEFINE_F(MyFixture, ScalarFastFloorLog2)
 }
 
 BENCHMARK_REGISTER_F(MyFixture, Dummy)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, ScalarExp)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, SIMDExp)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, ScalarExp_Unaligned)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, SIMDExp_Unaligned)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, ScalarLog)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, SIMDLog)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, ScalarSin)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, SIMDSin)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, ScalarCos)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
-BENCHMARK_REGISTER_F(MyFixture, SIMDCos)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
 BENCHMARK_REGISTER_F(MyFixture, ScalarLibmFloorLog2)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
 BENCHMARK_REGISTER_F(MyFixture, ScalarFastFloorLog2)->RangeMultiplier(4)->Range(1 << 6, 1 << 10);
 
diff --git a/benchmarks/BM_pan.cpp b/benchmarks/BM_pan.cpp
index 70707ee9..e08b447e 100644
--- a/benchmarks/BM_pan.cpp
+++ b/benchmarks/BM_pan.cpp
@@ -63,21 +63,6 @@ BENCHMARK_DEFINE_F(PanArray, SIMD)(benchmark::State& state) {
     }
 }
 
-BENCHMARK_DEFINE_F(PanArray, BlockOps)(benchmark::State& state) {
-    ScopedFTZ ftz;
-    for (auto _ : state)
-    {
-        sfz::fill(span2, 1.0f);
-        sfz::add<float>(span1, span2);
-        sfz::applyGain<float>(piFour<float>(), span2);
-        sfz::cos<float>(span2, span1);
-        sfz::sin<float>(span2, span2);
-        sfz::applyGain<float>(span1, absl::MakeSpan(left));
-        sfz::applyGain<float>(span2, absl::MakeSpan(right));
-    }
-}
-
 BENCHMARK_REGISTER_F(PanArray, Scalar)->RangeMultiplier(4)->Range(1 << 2, 1 << 12);
 BENCHMARK_REGISTER_F(PanArray, SIMD)->RangeMultiplier(4)->Range(1 << 2, 1 << 12);
-BENCHMARK_REGISTER_F(PanArray, BlockOps)->RangeMultiplier(4)->Range(1 << 2, 1 << 12);
 BENCHMARK_MAIN();
diff --git a/benchmarks/BM_ramp.cpp b/benchmarks/BM_ramp.cpp
index 3d7ee14e..2a056392 100644
--- a/benchmarks/BM_ramp.cpp
+++ b/benchmarks/BM_ramp.cpp
@@ -116,63 +116,6 @@ static void MulSIMDUnaligned(benchmark::State& state) {
     }
 }
 
-static void LogDomainScalar(benchmark::State& state) {
-    sfz::Buffer<float> output(state.range(0));
-    std::random_device rd { };
-    std::mt19937 gen { rd() };
-    std::uniform_real_distribution<float> dist { 1, 2 };
-    for (auto _ : state)
-    {
-        auto value = dist(gen);
-        sfz::linearRamp<float, false>(absl::MakeSpan(output), 1.0f, value);
-        sfz::applyGain<float, false>(std::log(2.0f), absl::MakeSpan(output));
-        sfz::exp<float, false>(output, absl::MakeSpan(output));
-    }
-}
-
-static void LogDomainSIMD(benchmark::State& state) {
-    sfz::Buffer<float> output(state.range(0));
-    std::random_device rd { };
-    std::mt19937 gen { rd() };
-    std::uniform_real_distribution<float> dist { 1, 2 };
-    for (auto _ : state)
-    {
-        auto value = dist(gen);
-        sfz::linearRamp<float, true>(absl::MakeSpan(output), 1.0f, value);
-        sfz::applyGain<float, true>(std::log(2.0f), absl::MakeSpan(output));
-        sfz::exp<float, true>(output, absl::MakeSpan(output));
-    }
-}
-static void LogDomainScalarUnaligned(benchmark::State& state) {
-    sfz::Buffer<float> output(state.range(0));
-    std::random_device rd { };
-    std::mt19937 gen { rd() };
-    std::uniform_real_distribution<float> dist { 1, 2 };
-    for (auto _ : state)
-    {
-        auto value = dist(gen);
-        auto outputSpan = absl::MakeSpan(output).subspan(1);
-        sfz::linearRamp<float, false>(outputSpan, 1.0f, value);
-        sfz::applyGain<float, false>(std::log(2.0f), outputSpan);
-        sfz::exp<float, false>(outputSpan, outputSpan);
-    }
-}
-
-static void LogDomainSIMDUnaligned(benchmark::State& state) {
-    sfz::Buffer<float> output(state.range(0));
-    std::random_device rd { };
-    std::mt19937 gen { rd() };
-    std::uniform_real_distribution<float> dist { 1, 2 };
-    for (auto _ : state)
-    {
-        auto value = dist(gen);
-        auto outputSpan = absl::MakeSpan(output).subspan(1);
-        sfz::linearRamp<float, true>(outputSpan, 1.0f, value);
-        sfz::applyGain<float, true>(std::log(2.0f), outputSpan);
-        sfz::exp<float, true>(outputSpan, outputSpan);
-    }
-}
-
 // Register the function as a benchmark
 BENCHMARK(Dummy)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
 BENCHMARK(LinearScalar)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
@@ -183,8 +126,4 @@ BENCHMARK(MulScalar)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
 BENCHMARK(MulSIMD)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
 BENCHMARK(MulScalarUnaligned)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
 BENCHMARK(MulSIMDUnaligned)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
-BENCHMARK(LogDomainScalar)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
-BENCHMARK(LogDomainSIMD)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
-BENCHMARK(LogDomainScalarUnaligned)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
-BENCHMARK(LogDomainSIMDUnaligned)->RangeMultiplier(4)->Range((1 << 2), (1 << 12));
 BENCHMARK_MAIN();
diff --git a/benchmarks/BM_readChunk.cpp b/benchmarks/BM_readChunk.cpp
index 4cdc2068..01b93aaa 100644
--- a/benchmarks/BM_readChunk.cpp
+++ b/benchmarks/BM_readChunk.cpp
@@ -61,7 +61,7 @@ BENCHMARK_DEFINE_F(FileFixture, JustRead)(benchmark::State& state) {
     {
         sfz::Buffer<float> buffer { numFrames * sndfile.channels() };
         sndfile.readf(buffer.data(), numFrames);
-        sfz::readInterleaved<float>(buffer, output->getSpan(0), output->getSpan(1));
+        sfz::readInterleaved(buffer, output->getSpan(0), output->getSpan(1));
     }
 }
 
@@ -75,7 +75,7 @@ BENCHMARK_DEFINE_F(FileFixture, AllocInside)(benchmark::State& state) {
         {
             sfz::Buffer<float> buffer { chunkSize * sndfile.channels() };
             auto read = sndfile.readf(buffer.data(), chunkSize);
-            sfz::readInterleaved<float>(
+            sfz::readInterleaved(
                 absl::MakeSpan(buffer).first(read),
                 output->getSpan(0).subspan(framesRead),
                 output->getSpan(1).subspan(framesRead)
@@ -95,7 +95,7 @@ BENCHMARK_DEFINE_F(FileFixture, AllocOutside)(benchmark::State& state) {
         while(framesRead < numFrames)
         {
             auto read = sndfile.readf(buffer.data(), chunkSize);
-            sfz::readInterleaved<float>(
+            sfz::readInterleaved(
                 absl::MakeSpan(buffer).first(read),
                 output->getSpan(0).subspan(framesRead),
                 output->getSpan(1).subspan(framesRead)
@@ -124,7 +124,7 @@ BENCHMARK_DEFINE_F(FileFixture, DrWavChunked)(benchmark::State& state) {
         while(framesRead < numFrames)
         {
             auto read = drwav_read_pcm_frames_f32(&wav, chunkSize, buffer.data());
-            sfz::readInterleaved<float>(
+            sfz::readInterleaved(
                 absl::MakeSpan(buffer).first(read),
                 output->getSpan(0).subspan(framesRead),
                 output->getSpan(1).subspan(framesRead)
diff --git a/benchmarks/BM_readChunkFlac.cpp b/benchmarks/BM_readChunkFlac.cpp
index 048ae23d..df77f637 100644
--- a/benchmarks/BM_readChunkFlac.cpp
+++ b/benchmarks/BM_readChunkFlac.cpp
@@ -61,7 +61,7 @@ BENCHMARK_DEFINE_F(FileFixture, SndFileOnce)(benchmark::State& state) {
     {
         sfz::Buffer<float> buffer { numFrames * sndfile.channels() };
         sndfile.readf(buffer.data(), numFrames);
-        sfz::readInterleaved<float>(buffer, output->getSpan(0), output->getSpan(1));
+        sfz::readInterleaved(buffer, output->getSpan(0), output->getSpan(1));
     }
 }
 
@@ -75,7 +75,7 @@ BENCHMARK_DEFINE_F(FileFixture, SndFileChunked)(benchmark::State& state) {
         {
             sfz::Buffer<float> buffer { chunkSize * sndfile.channels() };
             auto read = sndfile.readf(buffer.data(), chunkSize);
-            sfz::readInterleaved<float>(
+            sfz::readInterleaved(
                 absl::MakeSpan(buffer).first(read),
                 output->getSpan(0).subspan(framesRead),
                 output->getSpan(1).subspan(framesRead)
@@ -104,7 +104,7 @@ BENCHMARK_DEFINE_F(FileFixture, DrWavChunked)(benchmark::State& state) {
         while(framesRead < numFrames)
         {
             auto read = drflac_read_pcm_frames_f32(flac, chunkSize, buffer.data());
-            sfz::readInterleaved<float>(
+            sfz::readInterleaved(
                 absl::MakeSpan(buffer).first(read),
                 output->getSpan(0).subspan(framesRead),
                 output->getSpan(1).subspan(framesRead)
diff --git a/benchmarks/BM_resample.cpp b/benchmarks/BM_resample.cpp
index 75101ac3..ffd7f902 100644
--- a/benchmarks/BM_resample.cpp
+++ b/benchmarks/BM_resample.cpp
@@ -232,7 +232,7 @@ BENCHMARK_DEFINE_F(SndFile, HIIR2X_scalar)(benchmark::State& state)
 {
     for (auto _ : state) {
         auto baseBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, numFrames);
-        sfz::readInterleaved<float>(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
+        sfz::readInterleaved(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
         auto outBuffer = upsample2x<float, false>(*baseBuffer);
         benchmark::DoNotOptimize(outBuffer);
     }
@@ -242,7 +242,7 @@ BENCHMARK_DEFINE_F(SndFile, HIIR4X_scalar)(benchmark::State& state)
 {
     for (auto _ : state) {
         auto baseBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, numFrames);
-        sfz::readInterleaved<float>(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
+        sfz::readInterleaved(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
         auto outBuffer = upsample4x<float, false>(*baseBuffer);
         benchmark::DoNotOptimize(outBuffer);
     }
@@ -252,7 +252,7 @@ BENCHMARK_DEFINE_F(SndFile, HIIR8X_scalar)(benchmark::State& state)
 {
     for (auto _ : state) {
         auto baseBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, numFrames);
-        sfz::readInterleaved<float>(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
+        sfz::readInterleaved(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
         auto outBuffer = upsample8x<float, false>(*baseBuffer);
         benchmark::DoNotOptimize(outBuffer);
     }
@@ -262,7 +262,7 @@ BENCHMARK_DEFINE_F(SndFile, HIIR2X_vector)(benchmark::State& state)
 {
     for (auto _ : state) {
         auto baseBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, numFrames);
-        sfz::readInterleaved<float>(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
+        sfz::readInterleaved(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
         auto outBuffer = upsample2x<float, true>(*baseBuffer);
         benchmark::DoNotOptimize(outBuffer);
     }
@@ -272,7 +272,7 @@ BENCHMARK_DEFINE_F(SndFile, HIIR4X_vector)(benchmark::State& state)
 {
     for (auto _ : state) {
         auto baseBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, numFrames);
-        sfz::readInterleaved<float>(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
+        sfz::readInterleaved(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
         auto outBuffer = upsample4x<float, true>(*baseBuffer);
         benchmark::DoNotOptimize(outBuffer);
     }
@@ -282,7 +282,7 @@ BENCHMARK_DEFINE_F(SndFile, HIIR8X_vector)(benchmark::State& state)
 {
     for (auto _ : state) {
         auto baseBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, numFrames);
-        sfz::readInterleaved<float>(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
+        sfz::readInterleaved(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
         auto outBuffer = upsample8x<float, true>(*baseBuffer);
         benchmark::DoNotOptimize(outBuffer);
     }
@@ -300,7 +300,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC2x_BEST)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_BEST_QUALITY, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -317,7 +317,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC2x_MEDIUM)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_MEDIUM_QUALITY, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -334,7 +334,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC2x_FASTEST)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_FASTEST, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -352,7 +352,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC4x_BEST)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_BEST_QUALITY, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -369,7 +369,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC4x_MEDIUM)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_MEDIUM_QUALITY, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -386,7 +386,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC4x_FASTEST)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_FASTEST, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -403,7 +403,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC8x_BEST)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_BEST_QUALITY, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -420,7 +420,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC8x_MEDIUM)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_MEDIUM_QUALITY, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -437,7 +437,7 @@ BENCHMARK_DEFINE_F(SndFile, SRC8x_FASTEST)(benchmark::State& state)
         srcData.output_frames = static_cast<long>(2 * numFrames);
         src_simple(&srcData, SRC_SINC_FASTEST, static_cast<int>(numChannels));
         auto outBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, 2 * numFrames);
-        sfz::readInterleaved<float>(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
+        sfz::readInterleaved(*intermediateBuffer, outBuffer->getSpan(0), outBuffer->getSpan(1));
         benchmark::DoNotOptimize(outBuffer);
     }
 }
@@ -446,7 +446,7 @@ BENCHMARK_DEFINE_F(SndFile, HIIR8X_default)(benchmark::State& state)
 {
     for (auto _ : state) {
         auto baseBuffer = absl::make_unique<sfz::AudioBuffer<float>>(numChannels, numFrames);
-        sfz::readInterleaved<float>(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
+        sfz::readInterleaved(*interleavedBuffer, baseBuffer->getSpan(0), baseBuffer->getSpan(1));
         auto outBuffer = upsample8x<float>(*baseBuffer);
         benchmark::DoNotOptimize(outBuffer);
     }
diff --git a/benchmarks/BM_resampleChunk.cpp b/benchmarks/BM_resampleChunk.cpp
index a75c06c5..59ba3434 100644
--- a/benchmarks/BM_resampleChunk.cpp
+++ b/benchmarks/BM_resampleChunk.cpp
@@ -105,7 +105,7 @@ BENCHMARK_DEFINE_F(FileFixture, NoResampling)(benchmark::State& state) {
     {
         sfz::Buffer<float> buffer { numFrames * sndfile.channels() };
         sndfile.readf(buffer.data(), sndfile.frames());
-        sfz::readInterleaved<float>(buffer, output->getSpan(0), output->getSpan(1));
+        sfz::readInterleaved(buffer, output->getSpan(0), output->getSpan(1));
     }
 }
 
@@ -121,7 +121,7 @@ BENCHMARK_DEFINE_F(FileFixture, ResampleAtOnce)(benchmark::State& state) {
         upsampler4x.set_coefs(coeffsStage4x.data());
 
         sndfile.readf(buffer.data(), numFrames);
-        sfz::readInterleaved<float>(buffer, output->getSpan(0), output->getSpan(1));
+        sfz::readInterleaved(buffer, output->getSpan(0), output->getSpan(1));
 
         upsampler2x.process_block(temp.data(), output->channelReader(0), static_cast<long>(numFrames));
         upsampler4x.process_block(output->channelWriter(0), temp.data(), static_cast<long>(numFrames * 2));
@@ -171,7 +171,7 @@ BENCHMARK_DEFINE_F(FileFixture, ResampleInChunks)(benchmark::State& state) {
                 thisChunkSize * sndfile.channels()
             );
 
-            sfz::readInterleaved<float>(bufferChunk, leftSpan, rightSpan);
+            sfz::readInterleaved(bufferChunk, leftSpan, rightSpan);
 
             upsampler2xLeft.process_block(chunkSpan.data(), leftSpan.data(), static_cast<long>(thisChunkSize));
             upsampler4xLeft.process_block(output->channelWriter(0) + outputFrameCounter, chunkSpan.data(), static_cast<long>(thisChunkSize * 2));
diff --git a/benchmarks/CMakeLists.txt b/benchmarks/CMakeLists.txt
index 3f3ca5b2..b2fa4195 100644
--- a/benchmarks/CMakeLists.txt
+++ b/benchmarks/CMakeLists.txt
@@ -70,7 +70,7 @@ target_link_libraries(bm_logger PRIVATE sfizz::sfizz)
 
 if (TARGET sfizz-samplerate)
 sfizz_add_benchmark(bm_resample BM_resample.cpp ${BENCHMARK_SIMD_SOURCES})
-target_link_libraries(bm_resample PRIVATE sfizz-samplerate sfizz-sndfile)
+target_link_libraries(bm_resample PRIVATE sfizz-samplerate sfizz-sndfile sfizz-cpuid)
 endif()
 
 sfizz_add_benchmark(bm_envelopes BM_envelopes.cpp)
diff --git a/src/sfizz/SIMDHelpers.h b/src/sfizz/SIMDHelpers.h
index d22ef484..840853a3 100644
--- a/src/sfizz/SIMDHelpers.h
+++ b/src/sfizz/SIMDHelpers.h
@@ -103,12 +103,10 @@ bool unaligned(const T* ptr1, Args... rest)
 /**
  * @brief Read interleaved stereo data from a buffer and separate it in a left/right pair of buffers.
  *
- * The output size will be the minimum of the input span and output spans size.
- *
- * @tparam T the underlying type
  * @param input
  * @param outputLeft
  * @param outputRight
+ * @param inputSize
  */
 void readInterleaved(const float* input, float* outputLeft, float* outputRight, unsigned inputSize) noexcept;
 
@@ -125,12 +123,10 @@ inline void readInterleaved(absl::Span<const float> input, absl::Span<float> out
 /**
  * @brief Write a pair of left and right stereo input into a single buffer interleaved.
  *
- * The output size will be the minimum of the input spans and output span size.
- *
- * @tparam T the underlying type
  * @param inputLeft
  * @param inputRight
  * @param output
+ * @param outputSize
  */
 void writeInterleaved(const float* inputLeft, const float* inputRight, float* output, unsigned outputSize) noexcept;
 
@@ -144,10 +140,9 @@ inline void writeInterleaved(absl::Span<const float> inputLeft, absl::Span<const
 }
 
 /**
- * @brief Fill a buffer with a value; comparable to std::fill in essence.
+ * @brief Fill a buffer with a value
  *
  * @tparam T the underlying type
- * @tparam SIMD use the SIMD version or the scalar version
  * @param output
  * @param value
  */
@@ -157,92 +152,6 @@ void fill(absl::Span<T> output, T value) noexcept
     absl::c_fill(output, value);
 }
 
-/**
- * @brief Exp math function
- *
- * @tparam T the underlying type
- * @tparam SIMD use the SIMD version or the scalar version
- * @param input
- * @param output
- */
-template <class Type, bool SIMD = SIMDConfig::mathfuns>
-void exp(absl::Span<const Type> input, absl::Span<Type> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto sentinel = std::min(input.size(), output.size());
-    for (decltype(sentinel) i = 0; i < sentinel; ++i)
-        output[i] = std::exp(input[i]);
-}
-
-template <>
-void exp<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept;
-
-/**
- * @brief Log math function
- *
- * The output size will be the minimum of the input span and output span size.
- *
- * @tparam T the underlying type
- * @tparam SIMD use the SIMD version or the scalar version
- * @param input
- * @param output
- */
-template <class Type, bool SIMD = SIMDConfig::mathfuns>
-void log(absl::Span<const Type> input, absl::Span<Type> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto sentinel = std::min(input.size(), output.size());
-    for (decltype(sentinel) i = 0; i < sentinel; ++i)
-        output[i] = std::log(input[i]);
-}
-
-template <>
-void log<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept;
-
-/**
- * @brief sin math function
- *
- * The output size will be the minimum of the input span and output span size.
- *
- * @tparam T the underlying type
- * @tparam SIMD use the SIMD version or the scalar version
- * @param input
- * @param output
- */
-template <class Type, bool SIMD = SIMDConfig::mathfuns>
-void sin(absl::Span<const Type> input, absl::Span<Type> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto sentinel = std::min(input.size(), output.size());
-    for (decltype(sentinel) i = 0; i < sentinel; ++i)
-        output[i] = std::sin(input[i]);
-}
-
-template <>
-void sin<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept;
-
-/**
- * @brief cos math function
- *
- * The output size will be the minimum of the input span and output span size.
- *
- * @tparam T the underlying type
- * @tparam SIMD use the SIMD version or the scalar version
- * @param input
- * @param output
- */
-template <class Type, bool SIMD = SIMDConfig::mathfuns>
-void cos(absl::Span<const Type> input, absl::Span<Type> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto sentinel = std::min(input.size(), output.size());
-    for (decltype(sentinel) i = 0; i < sentinel; ++i)
-        output[i] = std::cos(input[i]);
-}
-
-template <>
-void cos<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept;
-
 namespace _internals {
     template <class T>
     inline void snippetSaturatingIndex(const T*& jump, T*& leftCoeff, T*& rightCoeff, int*& index, T& floatIndex, T loopEnd)
diff --git a/src/sfizz/SIMDSSE.cpp b/src/sfizz/SIMDSSE.cpp
index 3c9fdb8f..607ae2e7 100644
--- a/src/sfizz/SIMDSSE.cpp
+++ b/src/sfizz/SIMDSSE.cpp
@@ -16,90 +16,6 @@
 
 constexpr uintptr_t TypeAlignment = 4;
 
-template <>
-void sfz::exp<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto* in = input.begin();
-    auto* out = output.begin();
-    auto* sentinel = in + std::min(input.size(), output.size());
-    const auto* lastAligned = prevAligned(sentinel);
-
-    while (unaligned(in, out) && in < lastAligned)
-        *out++ = std::exp(*in++);
-
-    while (in < lastAligned) {
-        _mm_store_ps(out, exp_ps(_mm_load_ps(in)));
-        incrementAll<TypeAlignment>(out, in);
-    }
-
-    while (in < sentinel)
-        *out++ = std::exp(*in++);
-}
-
-template <>
-void sfz::cos<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto* in = input.begin();
-    auto* out = output.begin();
-    auto* sentinel = in + std::min(input.size(), output.size());
-    const auto* lastAligned = prevAligned(sentinel);
-
-    while (unaligned(in, out) && in < lastAligned)
-        *out++ = std::exp(*in++);
-
-    while (in < lastAligned) {
-        _mm_store_ps(out, cos_ps(_mm_load_ps(in)));
-        incrementAll<TypeAlignment>(out, in);
-    }
-
-    while (in < sentinel)
-        *out++ = std::exp(*in++);
-}
-
-template <>
-void sfz::log<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto* in = input.begin();
-    auto* out = output.begin();
-    auto* sentinel = in + std::min(input.size(), output.size());
-    const auto* lastAligned = prevAligned(sentinel);
-
-    while (unaligned(in, out) && in < lastAligned)
-        *out++ = std::exp(*in++);
-
-    while (in < lastAligned) {
-        _mm_store_ps(out, log_ps(_mm_load_ps(in)));
-        incrementAll<TypeAlignment>(out, in);
-    }
-
-    while (in < sentinel)
-        *out++ = std::exp(*in++);
-}
-
-template <>
-void sfz::sin<float, true>(absl::Span<const float> input, absl::Span<float> output) noexcept
-{
-    CHECK(output.size() >= input.size());
-    auto* in = input.begin();
-    auto* out = output.begin();
-    auto* sentinel = in + std::min(input.size(), output.size());
-    const auto* lastAligned = prevAligned(sentinel);
-
-    while (unaligned(in, out) && in < lastAligned)
-        *out++ = std::exp(*in++);
-
-    while (in < lastAligned) {
-        _mm_store_ps(out, sin_ps(_mm_load_ps(in)));
-        incrementAll<TypeAlignment>(out, in);
-    }
-
-    while (in < sentinel)
-        *out++ = std::exp(*in++);
-}
-
 template <>
 void sfz::applyGain<float, true>(float gain, absl::Span<const float> input, absl::Span<float> output) noexcept
 {
diff --git a/tests/EQ.cpp b/tests/EQ.cpp
index a75eea55..23082efe 100644
--- a/tests/EQ.cpp
+++ b/tests/EQ.cpp
@@ -68,7 +68,7 @@ int main(int argc, char** argv)
         sfz::Buffer<float> buffer { numFrames * 2 };
         sfz::Buffer<float> right { numFrames };
         sndfile.readf(buffer.data(), numFrames * 2 );
-        sfz::readInterleaved<float>(buffer, absl::MakeSpan(left), absl::MakeSpan(right));
+        sfz::readInterleaved(buffer, absl::MakeSpan(left), absl::MakeSpan(right));
     } else if (sndfile.channels() == 1) {
         sndfile.readf(left.data(), numFrames);
     } else {
diff --git a/tests/Filter.cpp b/tests/Filter.cpp
index 32c7bd2a..8104aa0b 100644
--- a/tests/Filter.cpp
+++ b/tests/Filter.cpp
@@ -71,7 +71,7 @@ int main(int argc, char** argv)
         sfz::Buffer<float> buffer { numFrames * 2 };
         sfz::Buffer<float> right { numFrames };
         sndfile.readf(buffer.data(), numFrames * 2 );
-        sfz::readInterleaved<float>(buffer, absl::MakeSpan(left), absl::MakeSpan(right));
+        sfz::readInterleaved(buffer, absl::MakeSpan(left), absl::MakeSpan(right));
     } else if (sndfile.channels() == 1) {
         sndfile.readf(left.data(), numFrames);
     } else {