Best Practices for Live Audio Mixing

Frequency Clipping

Another thing I need to consider is that clipping can occur where any particular frequency hits maximum level. But clipping LEDs don’t indicate which frequency is clipping. There might be bass frequencies from a body mic that clip well before the upper frequencies simply because of microphone placement, or handling noise, or even wind. This is so common that mixers often have a “high pass” filter, also known as a “bass roll off,” to lower the handling (wind noise) in outdoor events.

Wind noise can be an animal unto itself. How do they get such clean audio in such windy situations in the movies? They redub everything. But in a live stream, we don’t have that luxury. So those audience mics in the stands at the outdoor game could become nothing but noise as the winds of a cool front move in. Or a poorly placed room mic just might be directly under the HVAC vent, and thus it becomes unusable every time the air conditioner kicks on. Or the lav mic is placed high on the chest of a noisy breather, etc.

There are windscreens (Figure 4, below) available for all kinds of mics, from little lavaliere windscreens and handheld mic windscreens to blimps for shotgun mics. But I’ve been in situations where some nylon mesh on the mic, under the foam, inside the blimp, with a dead cat on it, was still getting intermittent wind noise. So sometimes in addition to accessories, some crafty placement (behind a pole, under an overhang, etc.) is required.

Figure 4. A handheld mic windscreen

Often wind noise is so broad-spectrum (spanning all the frequencies), there’s no simple equalizer adjustment or filter to get rid of it completely and give me clean, pristine audio. I can only protect the mic better from the wind, use the noisy audio at a lower level, or not use it at all.

However, I need to consider that, if my live stream involves a weathercaster in hurricane winds, some wind noise is actually desired. If the audio is too clean, it can sound fake. So I need to think of the overall situation and adjust accordingly.

Feedback

One of the biggest challenges for a single operator mixing for both the room and for the stream is understanding and handling the different needs for each. The remote audience benefits from hearing the audience in the live stream mix, just as the audience in the room hears everyone around them. So to get everyone laughing at a joke, I need microphones in the audience.

However, what I want to avoid at all costs is feedback. That (usually) high-pitched squeal kills us when the audio from a microphone in a room is taken in, amplified, and played back into the room, where that same microphone picks it up, amplifies it again, until (sometimes in less than a second) it becomes ear-splittingly loud.

I never want to use any mics in the room—or in front of the speakers—in the mix that is feeding the room. But that never stops someone with a handheld mic from walking into the audience or getting too close to the speakers in the room. When this happens, it can become a real challenge to manage the audio levels without feedback.

This is why there are dedicated professionals for live audio. They can use specialized equipment and processing to “sound out” a room and find the frequencies waiting to cause problems (Figure 5, below). Then, when the room is filling up with an audience, they do it again. They make live stage shows a pleasure to watch. If there’s a way they can give the streaming crew a “board feed” of the live show, that’s best. Meanwhile, we let them do the job they’re already great at.

Figure 5. Identifying problematic frequencies in a waveform can help prevent feedback.

Crewing for Live-Stream Audio

Then we have a second audio person for our live stream. This person captures additional audio for the viewers not in the room. This division of labor is the best way to do it. But if the labor can’t be divided, I recommend dividing the mixing boards. I’ll have two smaller mixers, right next to each other to make sure things stay separate. If you're trying to do it all on one mixer, I feel it's too easy to turn the wrong knob or press a channel assignment button and immediately have everyone in a room covering their ears. Keeping these two signal paths separate is the easiest way to avoid that.

Even within the room mixer, getting sound right for the stream can still be a challenge. Let’s say you’re producing a corporate webcast announcement for a company with a new glass building with a marble floor. There are processors the audio engineer can put in line like the Behringer Shark (Figure 6, below), to digitally notch those frequencies during the event before we even hear them. The other solution is to manage overall audio level in the space, which means working with the AV provider, so that many small speakers at low volume are spread out through the venue instead of a couple speakers on the stage at a very high volume.

Figure 6. The Behringer Shark FBQ100 Automatic Feedback Destroyer

Additionally, there are digital mixers that can “gate” microphones with audio below a threshold I set so they don’t add to the total overall audio level. These mixers quickly un-mute themselves automatically when each person speaks. It’s semi-automatic audio mixing (Figure 7, below). Couple this with some compression on the output, and after I tweak it a couple times, the mixer can deliver great audio without me having to adjust each microphone's fader up and down as each person speaks.

Figure 7. The auto-mixing capable Roland V-60HD

Sometimes, budgets don't allow for as many people as we’d like, and often audio is one of the first line items to get cut. We’ve all been there. I flip a couple mics on, set the levels. We’re good, right? But without a dedicated person listening, and focusing on, the audio, it's easy to miss when things start going downhill.

Let’s say we’re streaming a conference presentation, a question comes from the audience. We can hear it because we’re in the room, but we forget to pay attention to the audio going out to the streaming audience. Before we know it, we’re getting messages that the stream has no audio. Or, one of the wireless mics starts to quietly drop out intermittently as the batteries fade. Audio can be deceptively easy, making us forget to keep our attention on it.

If I can’t have a person dedicated to audio, then making sure I have ears, or at least one ear, on it at all times is key. Second to that is a good set of audio meters. And I don’t mean the tiny little audio meters tucked away in the corner of the window of an external monitor/recorder; I mean a dedicated set of audio meters I can see at a glance from several feet away.

Better than that is, of course, listening. The challenge comes as we whittle down the crew list and the director/tech director is also running sound while calling the shots over coms. Some people do one ear on each—a good set of headphones on one ear, with the com headset on the other ear. Some people use good audio earbuds underneath the com headsets, and then adjust the levels of both.

Neither is really a great solution. They are both compromises, in terms of both attention and the quality of listening. If I’m wearing com headsets over earbuds, without a bit of headset-swapping, I can’t be sure if the source of buzz in the audio is from coms or program audio. I don’t care if the com headsets buzz, but I do care if the program audio has a buzz.

So I tend to be a single-ear-per-use kind of person when forced to pare down the crew. If at all possible, I make sure there’s an audio person so I can hand off all the potential problems in this entire article, and more, to a dedicated person. Let them set it up, figure it out, worry about it, fix problems, etc. Let me focus on the stream.