Livestream and broadcast
How to Mix for Church Livestream
Quick answer
Mix the stream from a dedicated aux or matrix output, not the house mix, and land it at about -16 LUFS integrated with true peak at -1 dBTP. The room supplies nothing to the camera, so kick, snare and acoustic guitar all need 6 to 10 dB more than they get in the sanctuary.
The house mix is built for a room that is already making most of the sound. The drum kit is acoustically present at 105 to 120 dBA whether the console sends it anywhere or not, the electric guitar amplifier is pointed at the third row, and the congregation is singing. The microphone on a camera, or a stereo copy of that same house mix, hears none of that in the right proportion. That is why a stream that came straight off the main outputs sounds like all vocal and no band, with a congregation that has apparently gone home.
The fix is a second mix, built from the same inputs and mixed for a listener wearing earbuds in a kitchen. It costs nothing but console resources and about twenty minutes of setup. Before you touch a fader, work out what the stream itself has to carry: the livestream bitrate calculator shows that a 1080p 30 fps stream is 4,500 kbps of video plus 160 kbps of audio, and that 160 kbps audio allowance is the entire budget your mix has to live inside.
On this page
- Why a copy of the house mix always sounds wrong online
- What the room gives you free, and what the stream must replace
- Build a dedicated stream mix
- Starting levels that actually work
- The loudness number platforms normalise to
- When the console cannot give you a separate mix
- Five things that break a stream mix every week
Why a copy of the house mix always sounds wrong online
Everything the room contributes for free is missing from the stream. In a 250 seat sanctuary the acoustic drum kit alone runs 105 to 120 dBA at three feet, so a house engineer might have the kick channel almost closed and the overheads off entirely. Send that same mix to a camera and the service has no rhythm section. The same is true of a grand piano at 85 to 100 dBA, a brass section at 90 to 105 dBA, and a congregation of two hundred people singing, which no channel on the console represents at all.
The second problem is dynamic range. A sanctuary has an acoustic noise floor around 35 to 40 dBA and a congregation that tolerates a 20 dB swing between a whispered prayer and the last chorus. A phone speaker in a moving car has a noise floor thirty decibels higher and no tolerance at all. A stream mix is therefore more compressed, more mid-forward, and far more consistent between the sermon and the songs than anything you would ever run in the room.
The third problem is that nobody is listening to the stream. The house engineer is mixing the room, and the room is the job. Unless a person or a dedicated output owns the broadcast, the broadcast gets whatever falls out. That is the actual decision this page is about, and it is an organisational decision as much as a technical one.
What the room gives you free, and what the stream must replace
Work down this list channel by channel before the first rehearsal. Every row is something the congregation hears acoustically and the stream does not, which means the stream fader has to make up the difference.
Drums, piano and congregation reach the room at 85 to 120 dBA with no help from the console, so a stream mix typically needs those channels 6 to 12 dB higher than the house mix has them.
| Source | Level in the room at 3 ft | How the house mix treats it | What the stream mix needs |
|---|---|---|---|
| Acoustic drum kit | 105 to 120 dBA | Kick and snare only, overheads often off | Full kit, overheads up, 8 to 12 dB more than the house |
| Electric guitar amplifier | 95 to 110 dBA | Barely in the mix, or not at all | The microphone or direct feed carries the whole part |
| Grand piano | 85 to 100 dBA | A little support in the far seats | 6 to 10 dB more, and pan it where it sits on stage |
| Brass section | 90 to 105 dBA | Usually unamplified | Spot microphones become the only source |
| Congregation singing | Fills the room | No channel at all | Two room microphones, 18 to 20 dB under the vocal |
| Lead vocal | 70 to 80 dBA unamplified | The loudest thing in the mix | Still loudest, but 3 to 4 dB less dominant |
| Spoken word from the pulpit | 70 to 80 dBA unamplified | Gentle support, the room carries it | The only source, compressed harder and kept level |
Room levels are the published ranges used across this site for instrument sound pressure. Your stage will differ. The point is the ordering, not the decimal place.
Build a dedicated stream mix
- Take a post-fader aux or matrix, never the main output. On a digital console, assign a stereo aux, matrix or sub-group as the broadcast bus. On the X32 family use a stereo matrix fed from the mix bus; on smaller consoles a spare stereo aux works. Set every channel send to post-fader first, so the stream inherits a sane starting balance, then break the ones that need to differ.
- Set gain structure at the interface, not at the console output. Run the broadcast bus out at line level into a USB audio interface and set the interface so the service averages about -20 dBFS with peaks no higher than -6 dBFS. That leaves headroom for the loudest chorus without ever hitting the converter ceiling. Gain structure rules are the same as anywhere else in the system.
- Rebalance drums and instruments first. Bring up overheads, hi-hat and toms until the kit sounds like a kit on headphones. Then bring the electric guitar, piano and acoustic guitar up until the arrangement makes sense without the room. Expect 6 to 12 dB of difference from the house on these channels. This single step fixes most of what people complain about.
- Add two room microphones for the congregation. Hang or stand a pair of small condensers over the congregation, not over the platform, and fly them 18 to 20 dB under the lead vocal. They are not there to be heard as microphones. They are there so the viewer at home can tell that two hundred people are in the building.
- Compress the broadcast bus, gently and always. A 2:1 or 3:1 ratio with a slow attack and automatic release, aiming for 3 to 6 dB of gain reduction on the loudest passages, keeps the sermon and the last chorus within reach of each other. Follow it with a limiter set so true peak never exceeds -1 dBTP.
- High pass everything and cut the mud. Earbuds and phone speakers reproduce almost nothing below 150 Hz, so low frequency energy is wasted power that pushes the limiter. High pass vocals at 100 to 120 Hz, most instruments at 80 Hz, and take 2 to 3 dB out of the 250 to 400 Hz range on the broadcast bus where the boxy build-up lives.
- Check it on the worst device you own. Record ten minutes of rehearsal and play it back on a phone speaker at arm’s length, then on cheap earbuds. Anything that is inaudible there will be inaudible for a third of your audience. Fix it in the stream mix, not in the house.
- Save the mix as its own scene and hand it to a person. Store the broadcast bus settings in the console scene and assign one volunteer to watch a set of headphones and a loudness meter for the whole service. A mix nobody is listening to drifts within three weeks.
Starting levels that actually work
These are starting points, not destinations. Set them at rehearsal, then mix with your ears. All figures are relative to the lead vocal sitting at 0 dB of reference on the broadcast bus.
Set the lead vocal as the reference, put kick and snare 3 to 5 dB under it, and keep the congregation room microphones 18 to 20 dB down so they add space without becoming the mix.
| Channel | Relative level | Processing that matters | Common mistake |
|---|---|---|---|
| Lead vocal | 0 dB reference | 3:1 compression, 4 to 6 dB reduction, high pass 100 Hz | Too much reverb, which turns to mush on earbuds |
| Backing vocals | -4 to -6 dB | Grouped, compressed together, panned wide | Left at house level, where the room blends them |
| Kick drum | -3 to -5 dB | High pass 40 Hz, tight gate | Missing entirely because the room provided it |
| Snare | -3 to -5 dB | Gate, short plate reverb | Gated so hard the ghost notes vanish |
| Overheads and hi-hat | -8 to -10 dB | High pass 300 Hz | Turned off, which removes the whole kit |
| Bass guitar | -4 to -6 dB | Compression, some 800 Hz to survive small speakers | All sub and no midrange, so phones hear nothing |
| Electric guitar | -6 to -8 dB | Light compression | Left off because the amplifier is loud on stage |
| Acoustic guitar | -7 to -9 dB | High pass 120 Hz, notch at 200 Hz | Boomy, eating the vocal range |
| Piano or keys | -6 to -8 dB | Gentle compression, panned as on stage | Wide stereo that collapses to mono on phones |
| Congregation room pair | -18 to -20 dB | High pass 150 Hz, no gate | Up too far, which makes everything sound distant |
| Spoken word | 0 dB reference | 4:1 compression, 6 to 8 dB reduction | Left uncompressed, so quiet passages disappear |
Relative levels, not absolute fader positions. Every console scales differently. Link the lead vocal to your loudness target first and set everything against it.
The loudness number platforms normalise to
Streaming platforms normalise audio on playback. The widely published figures are around -14 LUFS for YouTube and roughly -14 to -16 LUFS elsewhere, which means a mix delivered much louder than that is simply turned down, and all you have bought with heavy limiting is a flatter, more fatiguing sound at the same perceived volume. Mixing to about -16 LUFS integrated with true peak at -1 dBTP gives you a stream that survives normalisation on every platform without being crushed.
Put a loudness meter where the operator can see it. Most streaming software includes one, and the cheap ones are accurate enough for this purpose. Watch the integrated number across the whole service rather than the momentary reading, because the integrated figure is what the platform measures. A service that reads -16 LUFS integrated with the sermon at a similar level to the songs is the target.
The other half of the job is consistency week to week. If one Sunday lands at -20 LUFS and the next at -12 LUFS, viewers reach for the volume control every time, and that is the complaint that eventually reaches the pastor. Saving the broadcast bus in a console scene and metering every week is what stops that drift. A weekly routine belongs on your Sunday checklist.
When the console cannot give you a separate mix
Plenty of churches run a small analog console with four aux sends already committed to monitors. You have three honest options and none of them involve pretending the house mix is fine.
The first is to use one aux as a mono broadcast send and accept mono. Mono is not a compromise on a phone speaker, and a mono stream with the right balance beats a stereo stream with the wrong one every time. The second is to add a small mixer fed from direct outputs or insert sends, mixed by a second volunteer. The third, and the one most churches eventually take, is to move to a digital console where a stereo matrix costs nothing but a menu page. If you are already planning that purchase, the console selection guide covers what channel count and bus count you need.
What you should not do is take the stream from a camera microphone. A camera at the back of a 53 foot room is hearing the loudspeakers, the air handling and the congregation in whatever proportion physics decides, with automatic gain control riding on top of it. The cheapest interface in the catalogue removes that problem for less than the cost of one wireless battery order, and the options are laid out in the livestream audio setup guide and the audio interface roundup.
Five things that break a stream mix every week
Post-fader sends nobody remembered. If the broadcast aux is post-fader and the house engineer pulls the acoustic guitar down because it is feeding back in the room, the stream loses it too. Decide deliberately which channels are pre-fader on the broadcast bus, usually the ones that exist only for the stream.
Effects returns not assigned to the broadcast bus. Vocals arrive at home bone dry while the room hears reverb. Check every effects return, not just the channels.
Mute groups. A mute group built for the room takes channels off the stream as well. Digital consoles let you exclude the broadcast bus from mute group behaviour. Do it, and test it.
Phantom power dropped on a room microphone. Nobody notices, because the room microphones are not in the house mix. A silent congregation is the symptom.
Gain changes during the sermon. A wireless lavalier that gets gain-ridden during the sermon will be riding the stream compressor too. Get the gain right at the head amplifier once, using the approach in the gain structure guide, and leave it alone. If the band is fighting you on stage level, the deeper fix is in reducing stage volume, because a loud stage bleeds into every open microphone the stream is using.
Sources
- Manufacturer published specifications for the consoles and interfaces named
- Published streaming platform encoder and loudness normalisation guidance
- Instrument sound pressure figures as used across this site
Frequently asked questions
Should the livestream mix be stereo or mono?
Mono is a perfectly good answer and it is the safer one. A large share of viewers listen on a single phone speaker, and a wide stereo mix collapses unpredictably to mono, which can partially cancel doubled instruments. If you mix in stereo, keep the important elements near the centre, pan only for width on overheads, keys and backing vocals, and check the mix in mono before you commit to it.
Do we really need room microphones for the congregation?
Yes, and they are the cheapest improvement available. Without them the stream sounds like a rehearsal in an empty building, because every channel on the console is a platform channel. Two small condensers hung over the congregation, eight to twelve rows back, blended 18 to 20 dB under the lead vocal, restore the sense that a room full of people is singing. They live on the broadcast bus only, so they never cause feedback in the house.
What loudness should a church livestream be?
Aim for about -16 LUFS integrated across the whole service with true peak at -1 dBTP. Platforms normalise playback to roughly -14 LUFS, so delivering much hotter than that gains nothing and costs you dynamic range. The integrated figure matters more than any momentary reading, and consistency between the sermon and the songs, and between one week and the next, matters more than hitting the number exactly.
Can one person mix the room and the stream at the same time?
For a spoken word service, yes. For a band, it is a stretch. The two mixes want opposite things at the same moments, and the operator can only hear one of them. Most churches that care about the stream set the broadcast mix at rehearsal, save it in a scene, and give one volunteer a set of closed headphones and a loudness meter to ride it during the service.
Why does our stream sound thin on phones but fine in the booth?
Studio headphones reproduce down to 20 Hz and a phone speaker starts rolling off around 400 Hz, so everything you are hearing in the low end is invisible to a large part of your audience. Fix it by adding midrange rather than more bass: bring 800 Hz to 2 kHz up on bass guitar and kick, high pass aggressively elsewhere, and check every mix on the cheapest earbuds in the building.
How much audio bitrate does the stream actually get?
At 1080p 30 fps the standard profile allocates 160 kbps to audio inside a 4,660 kbps total, and at 720p 30 fps it is 128 kbps inside 3,128 kbps. That is enough for a well controlled mix and unforgiving of a noisy one, because the encoder spends bits on hiss and room rumble just as readily as on music. A clean, compressed, high passed mix encodes better at the same bitrate.
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.