class SpeechTools < Formula desc "C++ general speech software library from the University of Edinburgh" homepage "http://festvox.org/docs/speech_tools-2.4.0/" url "http://festvox.org/packed/festival/2.4/speech_tools-2.4-release.tar.gz" sha1 "5b0ebd39bb7afa33e4093724a2123bdc62a6aebc" bottle do cellar :any sha1 "861a04fc277c90118cf88641c2ea2d07e7afa636" => :yosemite sha1 "c4b33389281c09184012fb79f53cc1112c8e11ce" => :mavericks sha1 "2fe64f901cd995effbcb26edf0ebce2b0a1141fd" => :mountain_lion end def install ENV.deparallelize system "./configure" system "make" # install all executable files in "main" directory bin.install Dir["main/*"].select { |f| File.file?(f) && File.executable?(f) } end test do rate_hz = 16000 frequency_hz = 100 duration_secs = 5 basename = "sine" txtfile = "#{basename}.txt" wavfile = "#{basename}.wav" ptcfile = "#{basename}.ptc" File.open(txtfile, 'w') do |f| scale = 2 ** 15 - 1 f.puts Array.new(duration_secs * rate_hz) { |i| (scale * Math.sin(frequency_hz * 2 * Math::PI * i / rate_hz)).to_i } end # convert to wav format using ch_wave system "ch_wave", txtfile, "-itype", "raw", "-istype", "ascii", "-f", rate_hz.to_s, "-o", wavfile, "-otype", "riff" # pitch tracking to est format using pda system "pda", wavfile, "-shift", (1 / frequency_hz.to_f).to_s, "-o", ptcfile, "-otype", "est" # extract one frame from the middle using ch_track, capturing stdout pitch = `ch_track #{ptcfile} -from #{frequency_hz * duration_secs / 2} -to #{frequency_hz * duration_secs / 2}`.strip # should be 100 (Hz) assert_equal frequency_hz, pitch.to_i end end