The analyzer is designed to handle many simultaneous noise and background signals at once, and can concentrate on whichever one you specify, including its harmonics. To do that, you need to specify where on the spectrogram you are interested in. You can look on the spectrogram (drag the magnifying glass over a region to zoom in) and find the approximate region of the spectrogram where your waveform of interest is located (specifically, the approximate time and approximate frequency of its fundamental). Assuming you already have your file open and the spectrogram is visible, then you can press "7" to specify the boundaries of a spiral surrounding your region of interest.
Freq. at centre = your best guess of the fundamental frequency. The centre of the spiral will be placed here. The program asks for this in case there are many interfering signals happening at the same time. This way, it knows which fundamental frequency to consider, and which to reject. You can use the magnifying glass to examine approximately where you would like to place the centre.
Freq. width = how broadly the analysis will span across the spectrogram vertically. Make this number approximately 2 or 3 times the vertical fatness of the spectral line. (This number also needs to be wider than at least 10 of the spectrogram grid elements.) Your goal is to cause several of the trials (or at least one) to surround and completely enclose the event on the spectrogram, rather than the boxes simply overlapping onto part of the event. If any/all of the trials have badly-shaped boxes and do not enclose the waveform well, then the program generates three warnings: (1) on the Confidence graph, the "F enclosure quality" for that trial is low; (2) the white oval for the particular trial has 'X' crossed through the top and bottom of the oval, indicating the true vertical span is not known because the event was not vertically enclosed properly; (3) the spectrogram text readout for Frequency is appended with an 'X'.
Time at centre = your best guess of the time at which the event occurs. The centre of the spiral is placed here. You can use the magnifying glass to examine approximately where you would like to place the centre.
Time width = how broadly the analysis will span across the spectrogram horizontally. If you intend to enclose the entire duration of the event on the spectrogram, make this approximately 2 or 3 times the horizontal fatness of the event. (This number also needs to be wider than at least 10 of the spectrogram grid elements.) If any/all of the trials do not enclose the duration of the waveform, then the program generates three warnings: (1) on the Confidence graph, the "T enclosure quality" for that trial is low; (2) the white oval for the particular trial has 'X' crossed through the sides of the oval, indicating the true horizontal span is not known because the event was not fully enclosed horizontally; (3) the spectrogram text readout for Time is appended with an 'X'.
Spiral: Number of loops around = how intricately you would like the algorithm to explore around your region. "2" is a good value to try.
Harmonics: Number to examine = how many boxes should be created to enclose harmonics. Do not specify so many boxes that they would spill over the top of the spectrogram.
Viewing pause time between each trial = number of seconds to slow down the analysis
Finally, the analysis will take a short time to execute. When complete, the markers will sit dormant at the last trial, waiting for you to cause them to snap back to the best trial.
Once multiple trials have been run through, the program waits for you to request data from a specific one of the trials.
For example, from the main menu you can enter "1" if you are trying to determine the fundamental frequency. Enter "2" if you were trying to determine the time of a percussive event (spread vertically across many frequencies). Enter "3" if you want to know the frequency and time of a short tone that was completely enclosed by your region-of-interest.
If you are analyzing a long-time steady signal and the frequencies of the spectrogram are too blurry (not enough frequency resolution), then you can press 8 to create a new spectrogram with better frequency resolution.
If you are analyzing a quick signal and are interested in exactly what time it occurred, but the spectogram is too blurry in time (not enough time resolution), then you can press 8 to create a new spectrogram with better time resolution.