Optimize dBFS/amplitude conversion functions (#7535)

* Optimize dBFS <-> amplitude functions

include/lmms_math.h

@@ -164,37 +164,22 @@ inline float linearToLogScale(float min, float max, float value)
 	return std::isnan( result ) ? 0 : result;
 }
 
-
-
-
-//! @brief Converts linear amplitude (0-1.0) to dBFS scale. Handles zeroes as -inf.
-//! @param amp Linear amplitude, where 1.0 = 0dBFS. 
-//! @return Amplitude in dBFS. -inf for 0 amplitude.
-inline float safeAmpToDbfs(float amp)
+inline float fastPow10f(float x)
 {
-	return amp == 0.0f
-		? -INFINITY
-		: log10f( amp ) * 20.0f;
+	return std::exp(2.302585092994046f * x);
 }
 
-
-//! @brief Converts dBFS-scale to linear amplitude with 0dBFS = 1.0. Handles infinity as zero.
-//! @param dbfs The dBFS value to convert: all infinites are treated as -inf and result in 0
-//! @return Linear amplitude
-inline float safeDbfsToAmp(float dbfs)
+inline float fastLog10f(float x)
 {
-	return std::isinf( dbfs )
-		? 0.0f
-		: std::pow(10.f, dbfs * 0.05f );
+	return std::log(x) * 0.4342944819032518f;
 }
 
-
 //! @brief Converts linear amplitude (>0-1.0) to dBFS scale. 
 //! @param amp Linear amplitude, where 1.0 = 0dBFS. ** Must be larger than zero! **
 //! @return Amplitude in dBFS. 
 inline float ampToDbfs(float amp)
 {
-	return log10f(amp) * 20.0f;
+	return fastLog10f(amp) * 20.0f;
 }
 
 
@@ -203,10 +188,29 @@ inline float ampToDbfs(float amp)
 //! @return Linear amplitude
 inline float dbfsToAmp(float dbfs)
 {
-	return std::pow(10.f, dbfs * 0.05f);
+	return fastPow10f(dbfs * 0.05f);
+}
+
+
+//! @brief Converts linear amplitude (0-1.0) to dBFS scale. Handles zeroes as -inf.
+//! @param amp Linear amplitude, where 1.0 = 0dBFS. 
+//! @return Amplitude in dBFS. -inf for 0 amplitude.
+inline float safeAmpToDbfs(float amp)
+{
+	return amp == 0.0f ? -INFINITY : ampToDbfs(amp);
 }
 
 
+//! @brief Converts dBFS-scale to linear amplitude with 0dBFS = 1.0. Handles infinity as zero.
+//! @param dbfs The dBFS value to convert: all infinites are treated as -inf and result in 0
+//! @return Linear amplitude
+inline float safeDbfsToAmp(float dbfs)
+{
+	return std::isinf(dbfs) ? 0.0f : dbfsToAmp(dbfs);
+}
+
+
+
 //! Returns the linear interpolation of the two values
 template<class T, class F>
 constexpr T lerp(T a, T b, F t)