Help with compression for single-mic live audio
January 19, 2021 1:24 AM   Subscribe

Audio engineers, a little help? I'm doing tech support for a small folk music organization that's running some live streamed concerts, and I'm trying to figure out if there's a magic compression setting that will help with the scenario where a band has a single microphone, most likely a LDC, that they're using both for their music and for introducing songs. Is there anything I can do to make their speech louder without making the music sound awful?

If I set levels for speech, the music is too loud, and if I set them for music, their speech is nearly inaudible.

I kind of understand how compression works, but I don't have a firm grasp on all the variables, and all the guides I'm finding online are assuming that any given microphone is being used for a narrower range of input than this. Is there a way to set a compressor so that the volume for both a person speaking and a trio singing and playing acoustic instruments is reasonable? Or should I just ride the gain slider manually for the whole show, boosting it between songs?

Everyone involved is in their own space -- musicians are in one location, I'm in another, with them sending an a/v feed to me to send back out through OBS. The only control I have over the levels they send me is during sound check, when I can make sure that they aren't clipping when playing at their loudest. They don't have an engineer on their end while performing; they have to be able to set their levels once and not touch them for the rest of the show.

The only live audio work I've done before now has been very simple (contra/square dances, where the only vocals are from a caller with their own mic), so I'm kind of at sea here. Thanks for any guidance you can give me!
posted by hades to Media & Arts (7 answers total)
 
Best answer: Compression isn't usually used to boost levels, but instead to rein them in. Compressing all the music to bring up the volume of a few words of intro vocal isn't a great solution.

I'd recommend just manually riding the gain.
posted by Spacelegoman at 1:58 AM on January 19, 2021 [2 favorites]


Best answer: Aside from running the board live - ideally with tools that allow for push-button presets - another option might be the old-timey gramophone approach: put the mic far from the loud stuff and have the announcer walk up very close to it when talking.
posted by eotvos at 2:17 AM on January 19, 2021 [2 favorites]


Best answer: Running the board live with fingers on the faders is the ordinary approach. In sound check, you should be making sure that no mics clip at the peak input they'll experience over the course of the whole show, so if someone's coming right up to the mic and introducing songs, you should be setting the gain appropriately for that, and then boosting the level in the quieter sections when the bands are playing.

I think if you want to stick with your one-setting-fits-all hands-off approach, eotvos has the best idea - figure out how much physically closer to the mic people need to be for the speaking levels to be right and mark the floor there, so people can walk up to the marked line and speak. The only caveat is that you need to be able to trust people not to ignore it/forget. Setting a brickwall compression level, so that all dynamics between quiet band-playing and close-miked talking are flattened out, will make the music (particularly gentle folk) sound worse.
posted by spielzebub at 2:44 AM on January 19, 2021 [4 favorites]


Best answer: Nthing pretty much all of the above.

any given microphone is being used for a narrower range of input than this.

Yeah. IME compression is very useful but it can only do so much while still remaining musical.

Is there a way to set a compressor so that the volume for both a person speaking and a trio singing and playing acoustic instruments is reasonable

Even for us pros there's a certain amount of real world trial and error here - "maybe, depending on the difference in levels" is about all I can say.

FWIW my sort of default starting point settings for compressing anything are: 2:1 ratio (I rarely go above 3 or 4:1); fast attack; moderate release; hard knee if I'm compressing something percussive (snare drum or a keyboard player with a heavy hand), moderate to soft knee otherwise, set the threshold initially so the compressor kicks in at the loudest parts. Then adjust to taste.

Or should I just ride the gain slider manually for the whole show, boosting it between songs?

One thing to note is that even mild compression can help with not having to make drastic fader changes - you don't have to suddenly leap for the fader and shoot it up 20 db, you can smoothly raise it 5 db over a second or two, which is generally unnoticeable to the listener.

The fact that the band is not already doing the "get closer to get louder" thing suggests that they're not used to performing with just one mic (or with mics at all)- acts that do this regularly will not only step forward to speak but will actually move around to mix/blend themselves as needed (the fiddle player steps forward for a solo, for example.) Getting them to try this could be worth a shot, along with reminding them that they need to speak up/project a little more when talking.

How are they getting the mic signal to you? If they're using an LDC they need phantom power, which suggests they have a small mixer or audio-to-usb interface on their end, and tons of those have at least 2 channels of input. Maybe they could add a talking mic in a second channel. They would have to set that mic's level as part of soundcheck, but if they get a mic with a switch they can just turn it on when needed and not have to adjust levels during the show.
posted by soundguy99 at 5:04 AM on January 19, 2021 [6 favorites]


Best answer: A little compression on the music will help smooth it out and bring down its volume relative to the speaking voice. I'm not sure there's a magic setting that would make them equal volume without totally destroying the musicality of the performance though. I'd encourage you to experiment a bit with compression, because good compression doesn't just reduce dynamic range but can help bring a mix together (even if that mix is being done based on their positioning to the mic).

You might be able to find a VST plugin that does automatic gain adjustment well enough for the purpose, though there's always the risk that it'll mangle the sound. Your best bet is applying a little compression and riding the fader. If this is something you're going to be doing a lot, investing in an external controller with a 100mm fader might give you a better experience than trying to do it with a mouse.

If they do have a second channel and mic, a separate mic just for speaking like soundguy99 recommended is a good idea.
posted by Candleman at 8:37 AM on January 19, 2021 [2 favorites]


Response by poster: It's a real grab bag on the musicians' side. Some have good interfaces/mixers and could probably be encouraged to hook up another mic to use just for spoken sections, others have USB microphones ranging from no-name $30 mics from Amazon to Blue Yeti. One was running everything from their iPhone, both audio and video (and that's probably going to be a common setup if we do a lot more of these).

It sounds like applying some light compression and picking up a hardware control surface for manually adjusting the levels live is going to be my best bet, which is what I figured. Good to have confirmation, though; thanks!
posted by hades at 11:33 AM on January 19, 2021 [1 favorite]


Response by poster: Update: I picked up a Behringer X-Touch Mini for $70 and mapped its fader to the volume control of the audio source they'll be coming in on. Even with a 60mm throw, it's so much nicer than trying to control it with the mouse. And now I have a bunch of other knobs and buttons I can use to automate other things. Thanks!
posted by hades at 11:30 AM on January 26, 2021 [2 favorites]


« Older The ancient world on two dollars a day   |   What is the green creature on the Wellington... Newer »
This thread is closed to new comments.