Livestream and broadcast
Church Livestream Audio Setup
Quick answer
Feed the stream from a dedicated aux, matrix or sub-group output into a USB audio interface at line level, averaging about -20 dBFS with peaks no higher than -6 dBFS. A two input interface is enough, and it removes the camera microphone from the signal path entirely.
Viewers forgive a soft picture and leave over bad sound. That is the whole reason this page exists, and it is why the audio feed deserves the first two hundred dollars of any livestream budget rather than the last. The decision is narrow: where in the console does the stream audio come from, and what converts it to something a computer or a switcher can use.
There are exactly four answers in common use, and they are not equally good. Three of them are fine and one of them is the camera microphone, which is the source of most of the complaints churches receive about their stream. Once the feed is chosen, the remaining work is gain structure and metering, and then the balance itself, which is covered in how to mix for livestream.
On this page
The four ways to get audio to the stream
Read the last column first. Every one of these is used in real churches; the question is which failure you are willing to live with.
A dedicated aux or matrix output into a two input USB interface is the right answer for almost every church, and it is the only option that lets the stream balance differ from the room balance.
| Feed source | What it costs | Sound quality | What breaks |
|---|---|---|---|
| Camera or phone microphone | Nothing | Poor | Hears loudspeakers, air handling and the room, with automatic gain riding over everything |
| Console main output, split or copied | A cable | Fair | The stream inherits every house decision, so the drums and congregation are missing |
| Dedicated aux, matrix or sub-group | An interface | Good | Nothing, provided somebody owns the mix and the sends are set deliberately |
| Multitrack over USB into software | A console with a USB interface | Best available | Needs a computer, a second operator and real preparation each week |
Most digital consoles in the 1,000 to 4,000 dollar range provide both a spare stereo bus and multitrack USB, so the choice is about volunteer capacity rather than hardware cost.
Wire the feed and set the level
- Assign a stereo aux, matrix or sub-group as the broadcast bus. On a digital console this is a menu page and costs nothing. Name it BROADCAST so the next volunteer does not repurpose it. On an analog console, commit one stereo aux pair or a spare sub-group and label the physical knobs with tape.
- Run it at line level into a USB audio interface. Balanced XLR or TRS from the console output into the interface line inputs. If the interface only has combination inputs with microphone preamps, engage the line or pad setting so you are not driving a microphone preamp with a line level signal, which is where most distorted stream audio comes from.
- Set the interface so the service averages -20 dBFS. Play the loudest part of rehearsal and set the input control so the meter averages around -20 dBFS with peaks no higher than -6 dBFS. Digital has no forgiveness above zero, and every decibel of headroom you leave here costs nothing because the broadcast compressor and limiter will set the final loudness anyway.
- Select the interface as the audio device in the streaming software. Not the built-in microphone, which is the default on every computer and which several churches have discovered live. Disable the computer’s internal microphone in the operating system so it cannot be selected by accident.
- Insert a compressor and a limiter on the broadcast path. A 2:1 to 3:1 compressor aiming for 3 to 6 dB of gain reduction on peaks, then a limiter with true peak at -1 dBTP. Do this on the console bus if it has the processing, or in the streaming software if it does not. Both is acceptable if neither is working hard.
- Check sync against the video. Audio through a console, an interface and streaming software arrives on a different timeline from video through a camera and a capture card. Watch a recording of someone speaking. If lips and words disagree, add delay to whichever arrives first, which is usually the audio.
- Record a rehearsal and listen on a phone. Not on the booth headphones. On a phone speaker and on cheap earbuds, which is what most of the congregation at home is using. Anything inaudible there needs fixing at the mix, not at the viewer’s volume control.
What interface to buy, and why not to buy less
The requirement is small: two line inputs, USB to the computer, and clean converters. Behringer UMC204HD Audiophile 2x4, 24-Bit/192 kHz USB Audio/MIDI Interface with Midas Mic Preamplifiers at $97.98 does it, and its four outputs let the streaming machine also feed a recorder or a confidence monitor. Focusrite Scarlett 2i2 4th Gen USB-C Audio Interface at $224.99 is the tier up and is the one most churches end up with, because the drivers are dependable across operating system updates, which matters when nobody is available to troubleshoot at 9:45 on a Sunday.
Step up to Focusrite Scarlett 4i4 4th Gen USB-C Audio Interface when the stream needs its own mix sent back into the console, or a talkback path to a camera operator. Step up to Focusrite Scarlett 18i16 4th Gen USB-C Audio Interface only if you are recording the service in multitrack to remix later, which is a real workflow but a different one, covered in how to record a church service.
If the stream runs through a hardware switcher rather than a computer, you may not need an interface at all. Blackmagic Design ATEM Mini Pro HDMI Live Stream Switcher takes analog audio on a 3.5 mm input and embeds it, which is enough for a stereo broadcast feed, though the input is consumer level and usually wants a pad or an attenuating cable from a professional line output. Check the level carefully; the symptom of getting this wrong is a stream that sounds crunchy only on loud passages.
Levels, metering and the numbers to watch
Three meters matter and they measure different things. Watching the wrong one is why a stream can look fine on the console and arrive distorted.
Set the interface so the service averages -20 dBFS with peaks under -6 dBFS, and let the broadcast limiter hold true peak at -1 dBTP while the integrated loudness lands near -16 LUFS.
| Point in the chain | Meter | Target | Why |
|---|---|---|---|
| Microphone preamp at the console | Channel meter | Average -18 dBFS, peaks -10 dBFS | Sets the noise floor for everything downstream |
| Broadcast bus output | Bus meter | Average -20 dBFS, peaks -6 dBFS | Leaves converter headroom for the loudest chorus |
| Audio interface input | Interface meter | Never lights the clip indicator | Clipping here cannot be undone anywhere later |
| Streaming software input | Software meter | Average -20 dBFS | Confirms the right device is actually selected |
| Final encoded stream | Loudness meter | -16 LUFS integrated, -1 dBTP | What the platform measures and normalises against |
dBFS and LUFS are different scales and are not convertible. A mix can sit at -20 dBFS average and still be far from -16 LUFS depending on how dense it is.
Room microphones, and where they belong in the signal path
Congregation microphones are part of the audio setup, not an afterthought, and their routing is the whole reason a separate broadcast bus is worth building. Two small condensers hung over the congregation, eight to twelve rows back and aimed away from the loudspeakers, go to the broadcast bus only. They never reach the house system, so they cannot cause feedback in the room no matter how much gain they have.
This is a genuinely useful thing to understand. Gain before feedback is a closed loop problem: a microphone can only ring if its output eventually reaches a loudspeaker that the microphone can hear. Take the loudspeaker out of that loop and the constraint disappears. The feedback guide covers the room side of that equation, and a hanging pair such as Audio-Technica PRO 45W ProPoint Cardioid Condenser Hanging Microphone or Audio-Technica U853R Cardioid Condenser Hanging Microphone 250 Ohms, Low Profile Design is the usual hardware.
Blend them 18 to 20 dB under the lead vocal. At that level they are not audible as microphones, and the stream simply stops sounding like an empty building. High pass them at 150 Hz to keep air handling rumble out, and never gate them, because a gate on a room pair breathes audibly every time the congregation stops singing.
Overflow, hearing assistance and the other feeds off the same bus
Once a broadcast bus exists, several other problems solve themselves from the same output. An overflow room, a nursery feed, a cry room and a lobby speaker all want approximately the same thing the stream wants: a mix balanced for a space with no acoustic contribution of its own. Feeding them from the broadcast bus rather than the main output is usually an improvement, and the routing is covered in overflow room audio and video.
Assistive listening is the exception. An assistive listening transmitter should be fed from a mix optimised for speech intelligibility, which is often the main output with music pulled down slightly rather than the broadcast mix, and the receiver count is set by the 2010 ADA table rather than by preference. The hearing assistance guide and the receiver calculator cover the counts.
The recording feed is the other one to plan now rather than later. If the same bus feeds the stream and the archive recording, you have one mix to get right and one thing to check. If the archive matters more than the stream, a multitrack recording lets you fix it afterwards instead, at the cost of somebody actually doing that work each week.
Sources
- Manufacturer published specifications for the interfaces, switchers and microphones named
- Published streaming platform loudness normalisation guidance
- 2010 ADA Standards for Accessible Design, section 219.3, for assistive listening counts
Frequently asked questions
Can we just take the stream feed from the console main output?
You can, and it is far better than a camera microphone, but it inherits every decision the house engineer makes. The drums, piano and congregation that the room supplies acoustically will be missing or quiet, and when the house engineer pulls a channel down for the room, the stream loses it too. A dedicated aux or matrix costs nothing on a digital console and removes that coupling entirely.
What level should the livestream audio feed be?
Set the interface so the service averages about -20 dBFS with peaks no higher than -6 dBFS. That headroom costs nothing because the broadcast compressor and limiter set the final loudness. Watch the interface meter rather than the console meter, because clipping at the converter cannot be repaired anywhere downstream and is the most common cause of a stream that sounds harsh only during loud songs.
Why is there a hum on the stream but not in the room?
Almost always a ground loop created when the console and the streaming computer are grounded through different circuits. Fix it with balanced connections throughout, an audio isolation transformer, or by putting both devices on the same circuit. If the hum appears only when the video cable is connected, use an HDMI isolator. Never lift a safety ground pin with an adapter; that is dangerous and it is not a fix.
Do we need a separate audio interface if we use a hardware switcher?
Often not. A switcher such as the ATEM Mini Pro accepts analog audio and embeds it into the stream, which covers a stereo broadcast feed. The catch is that the input is consumer level, so a professional line output usually needs a pad or an attenuating cable. Get that wrong and the stream sounds fine when quiet and crunchy when loud, which is a distortion problem rather than a mix problem.
How do we fix audio that is out of sync with the video?
Record a few minutes of someone speaking and watch their lips. Audio through a console and an interface usually arrives earlier than video through a camera and a capture card, so you add delay to the audio, in milliseconds, until they agree. Most streaming software and most switchers have an audio delay setting. Check it again whenever a camera or a capture device changes.
Where do congregation microphones plug in?
Into normal console channels, routed to the broadcast bus only and not to the main output. Because they never reach the house loudspeakers, they cannot feed back in the room regardless of their gain, which is the main advantage of a separate broadcast mix. Hang them over the congregation rather than the platform, high pass at 150 Hz, and blend them 18 to 20 dB below the lead vocal.
Researched, not professional advice. This page is compiled from published manufacturer specifications, operator manuals, FCC rules, published standards and owner-review consensus, not hands-on testing. Sound system design, rigging loudspeakers overhead, and any electrical work are jobs for a qualified professional: have flown loudspeakers and their attachment points signed off by a structural engineer or a certified rigger, and have all wiring done by a licensed electrician to your local code. Wireless microphone rules change, so confirm the current FCC position before buying. As an Amazon Associate we earn from qualifying purchases.