Kjetil Matheussen <k.s.matheussen@gmail.com> writes:
>> Mario Lang:
>
>> #include <jack.hpp>
>>
>> #include <boost/accumulators/accumulators.hpp>
>> #include <boost/accumulators/statistics.hpp>
>>
>> template<typename... Features>
>> using AudioAccumulatorSet = boost::accumulators::accumulator_set<
>> float, boost::accumulators::features<Features...>
>> >;
>>
>> using Count = boost::accumulators::tag::count;
>> using Max = boost::accumulators::tag::max;
>> using Min = boost::accumulators::tag::min;
>> using Mean = boost::accumulators::tag::mean;
>> using Variance = boost::accumulators::tag::variance;
>>
>> class Statistics final : public JACK::Client {
>> JACK::AudioIn In;
>> AudioAccumulatorSet<Count, Max, Mean, Min, Variance> Accumulator;
>>
>> public:
>> Statistics() : JACK::Client("Statistics"), In(createAudioIn("In")) {}
>> int process(std::uint32_t FrameCount) override {
>> for (auto &Value: In.buffer(FrameCount)) Accumulator(Value);
>> return 0;
>> }
>> auto max() const { return boost::accumulators::max(Accumulator); }
>> auto mean() const { return boost::accumulators::mean(Accumulator); }
>> auto min() const { return boost::accumulators::min(Accumulator); }
>> auto sampleCount() const { return boost::accumulators::count(Accumulator); }
>> auto variance() const { return boost::accumulators::variance(Accumulator); }
>> };
>
> Nice code. But I wonder about one small thing related to C++.
> Couldn't these max/mean/etc. methods in the Statististics class
> be written shorter like this?:
>
> auto max() const { return Max(Accumulator); }
> auto mean() const { return Mean(Accumulator); }
> auto min() const { return Min(Accumulator); }
> auto sampleCount() const { return Count(Accumulator); }
> auto variance() const { return Variance(Accumulator); }
No, because boost::accumulators::tag::mean is not the same as
boost::accumulators::mean. The first is a "tag" to let the accumulator_set
know (at compile time) which statistics it is supposed to collect,
and the second is a so-called "extractor" which is used to retrieve the result of a
particular statistics (obviously, at runtime).
> Sorry if it's a stupid question, but I haven't used "using" in C++ yet. :-)
Sorry for sort of overusing 'using' here, I was trying to make a point
(to me, mostly) of abstracting away overly long names without the big
hammer of just including a whole namespace. I tend to prefer knowing
what parts of a namespace are actually actively called/used in my programs.
If you are into long lines, you could of course drop all the using
directives, and eliminate the convenience template AudioAccumulatorSet altogether,
which would give you:
boost::accumulators::accumulator_set<float, boost::accumulators::features<boost::accumulators::tag::count, boost::accumulators::tag::mean, boost::accumulators::tag::min, boost::accumulators::tag::max, boost::accumulators::tag::variance>> Accumulator;
But now it doesn't look like an ad gig for C++ anymore :-)
But the point of that code was really just to make JACK::Client do
something.
-- CYa, ⡍⠁⠗⠊⠕ _______________________________________________ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org https://lists.linuxaudio.org/listinfo/linux-audio-devReceived on Sat Jul 29 16:15:01 2017
This archive was generated by hypermail 2.1.8 : Sat Jul 29 2017 - 16:15:02 EEST