Directional microphones aren’t magic.
One of my mantras over the last few years has become “directional microphones aren’t magic”. Using a cardioid lav doesn’t magically make the feedback go away unless source it’s trying to pick up and the source of any feedback are in nearly-opposite directions.
If you clip a cardioid lav onto a shirt pointing up, the source (the person’s head) will be roughly on-axis to the mic (great!) but I’m betting the speakers (which are probably mounted close to the ceiling) are also nearly on-axis (rats!). Even if (by some miracle) the speakers are 90 degrees off-axis (which would be weird!) with a cardioid pick-up pattern you’ll only see ~3 dB of attenuation.
But ~3 dB of attenuation isn’t the whole story, because with a microphone that small there’s no chance of it being an even ~3 dB. The frequency response of any directional microphone changes with angle, and the smaller (and cheaper) the microphone, the bigger the change. Even for mics with very good on-axis frequency response, off-axis frequency response can be a peaky mess. (One of the biggest differentiators between small-diaphragm condenser microphones that cost $100 and ones that cost $1000 is the off-axis frequency response.) Plus, basically all direction mics, including cardioid lavs, will exhibit some form of proximity effect, which is especially great when what you’re in proximity to is chest resonance.
For a practical example, let’s say that you’re an audio consultant for a house of worship, with a design goal of “set-and-forget” (i.e. no dedicated operator).
You clip a directional lav on the head pastor’s shirt. The pastor turns out to be a 6’8″ hulk, which means that the PA speakers are a full 90 degrees off-axis to the lav. You turn on the mic and get to work notching out frequencies that are feeding back. After a few minutes things are stable and everyone is happy. You go home and prepare your invoice.
The following week, a guest preacher comes in. They’re 5’2″. Suddenly those speakers are only 55 degrees off-axis. That ~3 dB turns into ~1 dB, but it’s worse than that, because the change in angle brings with it a change in frequency response. The preacher switches their microphone on and a whole new set of frequencies start feeding back, so a quick-thinking usher turns the level of the microphone down. The preacher is…mostly audible, but it isn’t great. Your invoice arrives at the church in the mail.
The following week, the head pastor is preaching again, but they stop to offer assistance to a broken-down motorist on the way in. As a result, they arrive at the church just as the service is about to start, and they aren’t really paying close attention as they put their microphone on. The lav ends up pointing pretty sideways. One of the speakers is now almost on-axis, while the pastor’s head is 90 degrees off. The pastor switches their microphone on and deafening feedback ensues. The quick-thinking usher tries turning the level of the microphone down, but when it’s low enough to make the feedback go away the pastor is inaudible unless they shout. Even then, they sound very strange, as if the microphone is broken.
The following week, you receive your unopened invoice in the mail marked “return to sender”.
Now let’s run the same example with an omni. Omnis are simple mics, so even cheap ones tend to measure relatively flat off-axis. All omnis exhibit an on-axis high frequency boost (with the centre frequency of the boost being determined by the size of the capsule), but for most lavs it’s high enough and tapers off gradually enough that it’s usually not a big problem.
You clip an omni lav on the head pastor’s shirt. The pastor turns out to be a 6’8″ hulk, which means that the PA speakers are a full 90 degrees off-axis to the lav, but with an omni that doesn’t really matter because you’re not going to get much in the way of axial attenuation anyway. You turn on the mic and notice that its relatively-flat frequency response means there isn’t too too much feedback to begin with. You high-pass and low-pass it, scoop out some frequencies that the P.A. is a bit peaky at, deal with chest resonance, etc. After a few minutes things are stable, the mic sounds pretty good, and everyone is happy. You go home and prepare your invoice.
The following week, a guest preacher comes in. They’re 5’2″. Suddenly those speakers are only 55 degrees off-axis, but the axial attenuation is still basically nil so it doesn’t make a difference. The preacher switches their microphone on and it works about as well as it did the week before. Your invoice arrives at the church in the mail.
The following week, the head pastor is preaching again, but they stop to offer assistance to a broken-down motorist on the way in. As a result, they arrive at the church just as the service is about to start, and they aren’t really paying close attention as they put their microphone on. The lav ends up pointing pretty sideways. One of the speakers is now almost on-axis, while the pastor’s head is 90 degrees off. The pastor switches their microphone on and the pastor sounds…mostly fine, maybe a little bit muffled? A quick-thinking usher tries turning the level of the microphone up to compensate, which results in a little bit of feedback, so they turn the level back down again. The sound isn’t as good as it was two weeks ago (or even last week), but it’s definitely intelligible. The pastor speaks up a little and the service continues basically as normal. After the service you get a phone call from the pastor, who admits upon questioning that the mic was pointed sideways (like you warned them about).
The following week, you receive a cheque for the value of the invoice.
TL;DR: directional microphones don’t magically make the feedback disappear. Directional microphones that move around a lot are often more trouble than they’re worth.