In the previous article on sound source equalization,
it was noted that the use of an anechoic room is required for accurate results.
Although that statement is true as a general principle,
it may not hold for every individual measurement setup.
As shown by Alastair Gurtner, in a follow up to his earlier work, using carefully created outdoor free-field measurements may give better results than those from a sub-optimal anechoic room.
Not all anechoic rooms absorb all sound at all frequencies, and depending on the size of the room and the absorbers that are used, deviations from the ideal may be quite substantial. Particularly for smaller rooms. It is therefore important to qualify the anechoic rooms one intends to use for sound source equalization measurements. In addition to the quality, the accessibility of anechoic rooms may also be an issue to contend with. Many rooms are not available for third parties, or rather expensive to use.
As an alternative to sound source equalization in an anechoic room, one could use outdoor free-field measurements, as demonstrated by Alastair Gurtner.
A good outdoor free-field measurement is not always easy to perform, and there are many factors that need to be taken into account. The measurement setup should be far removed from any reflecting surfaces, which means it should be relatively high above the ground. Climate conditions - temperature and humidity - may have an impact on the measurements, while these conditions cannot be controlled in a typical setup. Most important though is the influence of environmental noise - wind, traffic, animals, etc. All of these issues severely limit the times at which measurements can be performed.
The crux of the proposed free-field calibration method is to take the coarse (one-third octave) level spectrum of the sound source, and use it to shape an ideal noise-free impulse response (a dirac pulse). This shaped impulse response will then have the same third-octave spectrum as the sound source, and can subsequently be used in Dirac's SFD to create an equalized stimulus for the measured sound source.
A very detailed discussion of the use of this method, but also of the problems associated with anechoic room measurements, are discussed in a document provided by Alastair Gurtner. This download not only contains the document but also an Excel template designed for use with the new method.