Microphone positions for system alignment
“Where do I put the mic?” is probably the single most frequently asked question whenever system alignment topics are discussed. Ask fifteen sound engineers and you’ll get twenty opinions – here we will instead explore a reductionist approach to the question.
A measurement microphone can be considered a test probe – its job is to capture an acoustic signal at a point in space. So if we’re asking “where do we put the mic?,” we can reframe this as “what am I trying to learn with my measurement?” and then “where do I need to put the mic to learn that?”
If we think about the act of measurement as asking and answering a question, the path forward becomes a little more clear. If we want to know how long an object is, we measure its length. If we want to know how heavy an object is, we measure its weight. We have to ask the question and then choose the appropriate tool and measurement procedure to give us the answer.
This sounds so simple as to be almost silly, but let’s apply the same logic to a system alignment workflow. If we want to know what the level or tonality of a loudspeaker is, it follows that we should measure it from within the bulk of its coverage area (“on axis” or ONAX for short). If we want to match the level and tonality between two loudspeakers covering different areas, then we would want to measure each loudspeaker within its respective coverage area and then we can directly compare the level and tonality by observing the magnitude data. (Or to make the point a little more clearly, if you’re wondering what a certain loudspeaker sounds like or how loud it is, where would you stand to listen to it?)
Besides making decisions about tonality and level, another popular “question” we ask in the process of aligning a system has to do with relative timing – or in other words – who is arriving first? Since arrival times between different sources can shift quite dramatically as we move around the space, timing decisions have a specific location attached (“Who is arriving first here?). Since systems are most interactive where they’re equal in level, these “seams” between different sources are where it’s most critical to get the timing relationship correct. So now we know that we want to make level (gain) and tonality (EQ) decisions on-axis to a source, and timing (delay) decisions at the seams where those sources meet at equal level. How can we leverage these two “first principles” into practical microphone locations?
If you want a full rundown on my “order of operations” for sound system alignment, visit this post and then come back here after :). So “where do the mics go?” If we’re starting with the mains – I typically start with Main Left – and it’s a point source, you measure a few positions through its coverage area and set EQ based on the average (this could be a spatial average calculated by the analyzer from the individual measurements, or an “optical average” where you simply view all the measurements overlaid and visually evaluate for trend.).
If it’s an array with some granularity of processing, we can pay a closer eye to characterizing and then addressing variance by measuring on-axis to each zone and then balancing the HF response per zone as appropriate to improve consistency. So the measurement locations align with the physical zoning of the array (which hopefully also aligns with subdivision within the processing). If an array has four zones, you might then take four measurements at the proper locations ONAX to each zone, and then deal with HF tapering from there.
Should I average them all? Or should I not?
Do you want to see the trend? Or do you want to see the variance?
A spatial average shows you what’s common between measurement locations – so use it when you are asking questions about the overall response of a system in its coverage area. Example: a single loudspeaker covering half of the house. Measure from several locations in that coverage area and average them to view the response trend for making tonal decisions.
If, on the other hand, we are interested in what is different from location to location, then averaging is not helpful. We want to measure the system response at multiple locations throughout its coverage and compare how they differ to evaluate variance.
Head height, or on the ground?
Again, it depends what question we’re asking – listener ear height is an important choice for obvious reasons, but varying the stand height a bit can help randomize comb filtering in the data caused by floor bounce, if that’s the type of thing that you find bothersome. Since the ground bounce tends to present itself most severely in the 100 – 200 Hz range, it can sometimes complicate seeing clean data in that region, particularly for main/sub alignment, an alternative is sometimes in order.
A “boundary” or “ground plane” mic position, in which the microphone is placed directly on the ground, is helpful to eliminate the floor bounce from the measurement, which provides a cleaner impulse response due to the removal of the strong second arrival – and the frequency domain equivalent (lack of comb filtering and clearer phase trace). However the tonal response doesn’t correlate well with what’s going on at head height (put your head down near the floor sometime and take a listen). So a boundary mic placement is helpful for timing decisions but distinctly less so for tonal ones.
Mics pointing upwards, or at the PA?
Measurement mics are omnidirectional, so to a first approximation it shouldn’t matter which way we point them. However – that’s an important thing to prove to yourself. There are two main types of measurement microphones – free field and diffuse field (also called “random incidence”). In simple terms, the difference is that free field mics have their response corrected for the small HF buildup that occurs when placed in a sound field, and random incidence mics do not (they are designed for environments where sound energy is approaching from all directions, such as highly reverberant spaces or for measuring HVAC or environmental noise).
For this reason, the general rule of thumb here is to point free field mics at the source, and point random incidence mics away from it (70 to 90°, or “straight up” is a convenient and common approach). In practice, the difference in response due to orientation tends to be constrained to the top octave (above 8 kHz) and a dB or two, and is most pronounced when up close to a source where direct to reverberant ratio is high. Further back and in more reverberant spaces the difference is much less pronounced. If you know you’re using a random incidence mic (common models include the dbx RTA-M, the Rational Acoustics RTA-420, and the Beyerdynamic MM-1), it’s probably good practice to point it upwards. However, when in doubt just keep your microphone orientation consistent to eliminate that variable entirely.