Docs

README.md

@@ -114,7 +114,6 @@ actual bootleneck of the emulation.
 Also, sample rate conversions add delays to their outputs, which are not
 welcome to realtime processing.
 
-
 ### Libraries
 
 This library does not provide sample rate conversion itself, because proper
@@ -143,7 +142,6 @@ Suitable for realtime, and with small footprint.
 A nice modern C++ library, licensed as *GPL3*.
 Suitable for realtime, and with small footprint.
 
-
 ### Analog filtering
 
 The datasheet suggests some input and output analog filtering, to reduce
@@ -169,7 +167,6 @@ _______________________________________________________________________________
 Here you can find some descriptions and discussions about implementation
 details.
 
-
 ### Language
 
 I chose *C89* with a bit of *C99*. I was going to use *C++20* for my own
@@ -178,7 +175,6 @@ tiny library. C is also easier to integrate with other languages, and the
 mighty features of a colossal language like C++ are more of a burden than for
 actual use in such case.
 
-
 ### Cross-platform support
 
 The code itself should be cross-platform and clean enough not to give
@@ -188,7 +184,6 @@ Currently the library is developed and tested under *Windows* with *MSVC++ 14*.
 Of course, I will at least provide support for *gcc* and *clang* under *Linux*.
 I do not have access to *macOS*, but I guess the *Linux* support should fit.
 
-
 ### Code style
 
 I chose the path of verbosity for variable declaration, to help debugging fixed
@@ -201,7 +196,6 @@ at this stage of development. Actually I am not satisfied about how the
 code for vectoring should improve the performance by some margin, especially
 the parts for parallel 8-tap delay and output oversampling.
 
-
 ### Sample format
 
 The datasheet claims 14-bit *floating point* sampling for both input and
@@ -223,7 +217,6 @@ am emulating the system right now.
 Tests with 16-bit sample emulation sound less worse, so I am wondering whether
 signals are actually processed as 14-bit or more.
 
-
 ### Multiplication
 
 Being a DSP, one of the most common operations is multiplication. Since this
@@ -276,7 +269,6 @@ I guess the negative coefficients are actually loaded as the one-complement
 (bit flip) instead of the two-complement, as often seen in Yamaha's
 synthesizers of the same age, to save silicon area despite a tiny gain error.
 
-
 ### Digital delay line
 
 The digital delay line allows up to 100 ms of delay at the nominal input rate.
@@ -286,7 +278,6 @@ Such buffer should be a shifting FIFO with 32 pre-determined tap positions.
 The delay line emulation is actually implemented as a random-access ring
 buffer, as shifting the whole buffer at each input sample is a waste of time.
 
-
 ### Oversampler
 
 The chip includes a 2x oversampling interpolator at the output stage, to help
@@ -308,7 +299,23 @@ I also think that the kernel is not *minimum-phase*, to save further silicon
 area, thanks to the mirrored coefficient values.
 I am no expert, but it looks like minimum-phase is also not welcome to audio,
 because of phase-incoherence among frequencies, despite having shorter delays.
-I left the possibility to choose the minimum-phase feature by configuration.
+I left the possibility to choose the minimum-phase feature by configuring the
+`YM7128B_USE_MINPHASE` preprocessor symbol.
+
+### Floating-point
+
+It is possible to configure the floating point data type used for processing,
+via the `YM7128B_FLOAT` preprocessor symbol. It defaults to `double`
+(*double precision*).
+
+Please note that, contrary to common beliefs, the `double` data type is
+actually very fast on machines with hardware support for it.
+
+I think that you should switch to `float` (*single precision*) only if:
+
+- the hardware is limited to it, or
+- if machine code vectorization gets faster, or
+- conversion from/to buffer data to/from double precision is slower.
 
 _______________________________________________________________________________
 

src/YM7128B_emu.c

@@ -369,7 +369,7 @@ YM7128B_Tap const YM7128B_Tap_Table[YM7128B_Tap_Value_Count] =
 
 YM7128B_Fixed const YM7128B_OversamplerFixed_Kernel[YM7128B_Oversampler_Length] =
 {
-#if YM7128B_MINPHASE
+#if YM7128B_USE_MINPHASE
     // minimum phase
     KERNEL(+0.073585247514714749),
     KERNEL(+0.269340051166713890),
@@ -448,7 +448,7 @@ YM7128B_Fixed YM7128B_OversamplerFixed_Process(
 
 YM7128B_Float const YM7128B_OversamplerFloat_Kernel[YM7128B_Oversampler_Length] =
 {
-#if YM7128B_MINPHASE
+#if YM7128B_USE_MINPHASE
     // minimum phase
     KERNEL(+0.073585247514714749),
     KERNEL(+0.269340051166713890),

src/YM7128B_emu.h

@@ -48,8 +48,8 @@ extern "C" {
 #define YM7128B_Float_Max   (+1)    //!< Minimum floating point value
 #endif
 
-#ifndef YM7128B_MINPHASE
-#define YM7128B_MINPHASE 0          //!< Enables minimum-phase oversampler kernel
+#ifndef YM7128B_USE_MINPHASE
+#define YM7128B_USE_MINPHASE 0          //!< Enables minimum-phase oversampler kernel
 #endif
 
 // ============================================================================