Below is a screen display of the impulse and the reflections in the sampled data for the '3&SQ Left' baffle using full parametric equalization. The trace represents 3770 points of data and includes all the wall reflections and other environment problems therein. Much of the 'time of flight' of the impulse has been edited out of the plot, as it contains no useful information. The full data set is 4096 points in length.
Data aquired in a normal room is filled with echos and reflections due to the hard surfaces that do not absorb sound. These intrusions into the data represent contamination, at least, in-so-far as we desire to produce 'Free Field' or semi-anechoic data. Hence the requirement that the data be truncated to exclude the undesired reflections.
Below is the plot of the truncated data that excludes major reflections. Down from over 3700 points of data, to 432 points! The meaning of this in regards Fast Fourier Transform conversion of the data from a time series (the impulse) to frequency response, is that the low frequency limit of the conversion is increased. Also, resolution in the frequency response is lost.
Finally, here is the same plot as above, but with a few more points added to
show the first major reflection. Can you find that reflection in the first
3770 point plot? Rudi Blondia has suggested a good name for the noise that
appears between the impulse and the first reflection, to wit, "grit".
Measurement Issues - The Meaning of the Measurements.
Measurement Issues - Comparison: MLS Averages.
Acoustic Line Source Research - Table of Contents.
