'cpps' measures cepstral peak promience (with or without smoothing across frames and/or quefrency) from a time signal in RIFF (.wav) format. 'cpps.exe' is a stand-alone console application that runs out of Windows Command Prompt or a Cygwin (Unix) console window (www.cygwin.com). Copy the '.exe' file to any directory that is in your path (from Windows command prompt, type 'path ' to see what your path looks like). To get a usage message, type: cpps A full description of the algorithm can be found in Hillenbrand & Houde (JSLHR, 1996). The most straightforward uses of the program are: cpps ah.wav vowel (sets the defaults for sustained vowels) cpps sentence.wav speech (sets the defaults for connected speech) The program supports quite a few options. A brief description of each option is given in the program's usage message. The default for each option is given in square brackets following the description. Option names can be abbreviated; e.g., 'cpps sig.wav vowel' and 'cpps sig.wav vow' both work. The shortest unambiguous abbreviation is given in upper case letters in the usage message; e.g., 'OUtfile=FILE' in the description of the 'outfile' option means that something like 'cpps ou=sig.f0' will work, but 'cpps o=sig.f0' will not. Options that are described with something like 'Zcross=#' require a numeric argument (e.g., 'zcross=20'). Options that are described with something like 'Vowel=ON/OFF' are switches that are to be turned on or off. The switch can be turned on with either 'cpps sig.wav vowel=on' or just 'cpps sig.wav vowel'; the switch is turned off with something like 'cpps sig.wav vowel=off'. The default (on or off) is given in square brackets. After the program is run, 'cpps' will report mean fundamental frequency and mean CPP for the file. If smoothing is used, the program will also report mean CPPS. Measurements for each frame are written to 'inname.f0c' (e.g., for the command 'cpps sig.wav', the frame-by-frame output file will default to 'sig.f0c', unless the default is changed with something like 'cpps sig.wav out=sig.pitch'). The format of this output file is: col1 = time col2 = F0 col3 = intensity in dB for a 10 ms interval beginning at the time reported in col1 col4 = number of zero crossings for the same 10 ms interval col5 = cepstral peak amplitude (unnormalized) col6 = cepstral peak prominence Signals can be analyzed one at a time or they can be analyzed in batch mode using what we have called a 'stack' file. A stack file is just a plain text file consisting of a list of the files that are to be analyzed; for example: sig1.wav sig2.wav sig3.wav sig4.wav The stack file can be created with any editor that produces a plain text file (i.e., a Word '.doc' file will not work). The DOS 'dir' command can also be used to create the stack file; e.g., dir/b sig*.wav > signals.nam The stack file must have the extension '.nam'. Once the stack file is made, the signals can be analyzed using the same kinds of commands that are used to analyze single files. A few examples: cpps sentences.nam speech cpps vowels.nam vowel shift=5 upper=400 lower=150 CPP and/or CPPS measurements are written to the file specified with the 'cpp=FILE' and 'cps=FILE' options. If these options are not set on the command line, the filenames default to 'STACKNAME.cpp' (unsmoothed CPP) and 'STACKNAME.cps' (CPP with smoothing). For example, for the commands above, CPP measurments will be written to 'sentences.cpp' and 'vowels.cpp' and CPPS measurments will be written to 'sentences.cps' and 'vowels.cps'. cpps sentences.nam speech cps=cps.dat cpp=cpp.dat cpps vowels.nam vowel shift=5 upper=400 lower=150 cps=cpsv.dat cpp=cppv.dat The formant of these summary output files is simple: col1 = filename col2 = mean F0 col3 = mean CPP or CPPS