C++ template function to increase the number of bit in a number uniformly.

Increase-Bit-Depth.cpp

#include <climits>
#include <cstdint>
#include <iostream>

template <bool B, class T = void>
struct enable_if {};

template <class T>
struct enable_if<true, T> {
    typedef T type;
};

template <bool B, class T = void>
using enable_if_t = typename enable_if<B, T>::type;

template <size_t Bits_out, size_t Bits_in, class T_out, class T_in>
enable_if_t<(Bits_out <= 2 * Bits_in), T_out> increaseBitDepthImpl(T_in in);

template <size_t Bits_out, size_t Bits_in, class T_out, class T_in>
enable_if_t<(Bits_out > 2 * Bits_in), T_out> increaseBitDepthImpl(T_in in) {
    constexpr size_t leftShift = Bits_out - Bits_in;
    return ((T_out) in << leftShift) |
           increaseBitDepthImpl<leftShift, Bits_in, T_out>(in);
}

template <size_t Bits_out, size_t Bits_in, class T_out, class T_in>
enable_if_t<(Bits_out <= 2 * Bits_in), T_out> increaseBitDepthImpl(T_in in) {
    constexpr size_t leftShift  = Bits_out - Bits_in;
    constexpr size_t rightShift = Bits_in - leftShift;
    return ((T_out) in << leftShift) | (in >> rightShift);
}

template <size_t Bits_out, size_t Bits_in, class T_out, class T_in>
T_out increaseBitDepth(T_in in) {
    static_assert(Bits_in <= sizeof(T_in) * CHAR_BIT,
                  "Error: Bits_in > bits(T_in)");
    static_assert(Bits_out <= sizeof(T_out) * CHAR_BIT,
                  "Error: Bits_out > bits(T_out)");
    return increaseBitDepthImpl<Bits_out, Bits_in, T_out>(in);
}

using std::cout;
using std::endl;

int main() {
    for (int i = 0; i < 0x10; ++i) {
        cout << increaseBitDepth<4, 4, uint16_t, uint8_t>(i) << '\t'
             << increaseBitDepth<5, 4, uint16_t, uint8_t>(i) << '\t'
             << increaseBitDepth<6, 4, uint16_t, uint8_t>(i) << '\t'
             << increaseBitDepth<7, 4, uint16_t, uint8_t>(i) << '\t'
             << increaseBitDepth<8, 4, uint16_t, uint8_t>(i) << '\t'
             << increaseBitDepth<9, 4, uint16_t, uint8_t>(i) << '\t'
             << increaseBitDepth<16, 4, uint16_t, uint8_t>(i) << endl;
    }
}