Livestream and broadcast
How to Record a Church Service
Quick answer
A 32 channel multitrack recording of a 90 minute service at 24-bit 48 kHz is about 25 GB, or 16.6 GB per hour. A stereo board mix of the same service is 1.6 GB, and a 1080p 30 fps video file at 4,660 kbps adds 2.1 GB per hour.
Recording decisions are storage decisions, and storage is the thing churches discover they got wrong in year two when the drive is full and nobody wants to delete anything. So start with the arithmetic. Audio at 24-bit and 48 kHz uses 144 kilobytes per second per channel, which is 0.52 GB per channel per hour. Multiply by the channel count and you have the whole answer.
That gives a stereo board mix at 1.04 GB per hour, a 16 channel multitrack at 8.29 GB per hour and a 32 channel multitrack at 16.59 GB per hour. Video is comparatively small: the same 1080p 30 fps encode a livestream uses, at 4,660 kbps, is 2.1 GB per hour, which you can check in the bitrate calculator. The expensive thing to store is not the video. It is the multitrack.
On this page
What each recording method costs to store
Pick the row that matches what you will actually do with the recording afterwards. The last column is the real decision: a multitrack you never remix is 25 GB a week of nothing.
A 32 channel multitrack of a 90 minute service is about 25 GB, sixteen times the 1.6 GB a stereo board mix of the same service uses.
| Method | Per hour | 90 minute service | Per year, 52 services | What you can fix afterwards |
|---|---|---|---|---|
| Stereo board mix, 24-bit 48 kHz | 1.04 GB | 1.6 GB | About 83 GB | Overall level and simple edits only |
| Stereo board mix, compressed | About 0.06 GB | About 0.09 GB | About 5 GB | Nothing, but it is a fine archive format |
| 16 channel multitrack | 8.29 GB | 12.4 GB | About 646 GB | Most balance problems, one instrument at a time |
| 24 channel multitrack | 12.44 GB | 18.7 GB | About 971 GB | The same, with the drum kit properly separated |
| 32 channel multitrack | 16.59 GB | 24.9 GB | About 1.3 TB | Everything except what the microphones never captured |
| 1080p 30 fps video | 2.10 GB | 3.2 GB | About 164 GB | Edits and cuts, not framing or focus |
| Isolated recording of four camera inputs | About 8.4 GB | About 12.6 GB | About 655 GB | Bad camera cuts, re-edited after the fact |
Audio figures are calculated from 24-bit 48 kHz uncompressed, at 0.52 GB per channel per hour. Video figures use the standard 1080p 30 fps profile at 4,660 kbps. Compressed audio depends heavily on the encoder setting chosen.
Decide what the recording is for before choosing the method
Three different jobs get called recording the service, and they want different things.
An archive. Somebody was ill, somebody wants the sermon again, the church needs a record. A stereo board mix is completely adequate, compressed audio is adequate, and the whole year fits in a few gigabytes. Do not build a multitrack workflow for this.
A podcast or a sermon feed. This wants the spoken word clean and consistent, which means a feed that is mostly the pulpit microphone with the music mixed sensibly around it. A stereo record of the broadcast bus is usually right, with light processing afterwards.
Music you intend to release, or a mix you intend to fix. Only this one needs multitrack, and only if a person is genuinely going to sit down and remix it each week. That is several hours of work per service. Churches consistently overestimate how much of this they will do, and 1.3 TB a year of unopened multitrack is the evidence.
Being honest about which one you are doing saves both money and disappointment. If the answer is the first, the entire recording setup is one output from the console and a device to catch it, and the mix that feeds it is the one described in how to mix for livestream.
Set up the recording
- Choose the source: broadcast bus or multitrack. For an archive or a podcast, record the same dedicated broadcast bus that feeds the stream, so there is one mix to get right rather than two. For a remix workflow, record every channel pre-fader and pre-processing, so nothing the operator does during the service is baked into the file.
- Set levels for the loudest moment, not the sermon. Aim for an average around -20 dBFS with peaks no higher than -6 dBFS, set during the loudest song at rehearsal. Digital clipping cannot be repaired afterwards, and at 24-bit there is no penalty at all for recording conservatively. The headroom is free.
- Record to a dedicated drive, never the system drive. A 32 channel multitrack writes about 4.6 megabytes every second for the whole service. Give it its own solid state or fast external drive with plenty of free space, and check the free space as part of the weekly routine rather than discovering it during the offering.
- Start the recording before the countdown and stop it after the room clears. Storage is cheaper than a missing first line. Start it at the beginning of the pre-service countdown and leave it running through the closing announcements. Trimming a long file is trivial and recovering a missed take is impossible.
- Name the file the moment the service ends. A consistent scheme with the date in year, month and day number order, the service name and the type, such as 2026-09-16-1030-multitrack. Files named untitled_003 are indistinguishable within a month and are the main reason church archives become unusable.
- Copy it off the recording machine the same day. Follow the three, two, one rule: three copies of anything worth keeping, on two different kinds of storage, with one of them off site or in cloud storage. A single external drive in the booth is not a backup, it is one flood or one theft away from nothing.
- Check the recording actually captured before anyone leaves. Play the last thirty seconds back. Every church with an archive has at least one week where the meter was moving and the file was empty, or the wrong device was selected. A thirty second check catches it while the platform is still set up.
- Log it and hand it to whoever publishes it. One shared sheet: date, what was recorded, where the file is, whether it has been backed up and whether it has been published. This is the step that collapses first when a volunteer is away, so write it down rather than keeping it in one person’s head.
What to record with
For a stereo archive, the cheapest honest answer is the interface already feeding the stream. Behringer UMC204HD Audiophile 2x4, 24-Bit/192 kHz USB Audio/MIDI Interface with Midas Mic Preamplifiers at $97.98 or Focusrite Scarlett 2i2 4th Gen USB-C Audio Interface captures the broadcast bus into any computer, and Focusrite Scarlett 4i4 4th Gen USB-C Audio Interface adds outputs for a confidence feed. Many digital consoles will also record a stereo file to a USB stick with no computer at all, which is the most volunteer-proof option available and worth checking for on the console you already own.
For multitrack, the console usually is the interface. An Behringer 32 40-Input 25-Bus Digital Mixing Console, Black (X32) carries 32 channels over USB to a computer, and Allen & Heath CQ-18T Digital Mixer with 7" Touchscreen WiFi and Bluetooth Connectivity and PreSonus StudioLive Series III 16R Digital Rack Mixer with Studio One Pro Software record multitrack directly to an attached drive. If the console cannot, Focusrite Scarlett 18i16 4th Gen USB-C Audio Interface at $569.99 gives eighteen inputs, which covers a typical church input list. Which console you have shapes this decision more than anything else, and if you are still choosing, the console guide covers recording capability as a selection criterion.
For video, Blackmagic Design ATEM Mini Pro HDMI Live Stream Switcher records the programme output to a USB drive, and Blackmagic Design ATEM Mini Pro ISO HDMI Switcher at $585.00 records all four inputs separately plus an editing project file. That isolated recording is the feature worth paying for if anyone re-edits the service, because it turns a bad camera cut from a permanent feature of the archive into something fixable on a Tuesday.
Levels and file settings that hold up later
Getting these right at the recorder is the difference between a file you can work with and one you cannot.
Record at 24-bit 48 kHz with an average around -20 dBFS and peaks no higher than -6 dBFS, which costs 0.52 GB per channel per hour and leaves every repair option open.
| Setting | Choose | Why |
|---|---|---|
| Bit depth | 24-bit | Gives enormous headroom at no practical cost, so conservative levels are free |
| Sample rate | 48 kHz | The video standard, so audio and video stay in step with no conversion |
| Average level | About -20 dBFS | Leaves room for the loudest chorus without ever reaching the ceiling |
| Peak level | No higher than -6 dBFS | Digital clipping is permanent and cannot be repaired afterwards |
| Multitrack tap point | Pre-fader, pre-processing | Nothing the operator does live is baked into the file |
| File format | Uncompressed WAV for work, compressed for the archive | Edit from the uncompressed file, publish and store the compressed one |
| Backup | Three copies, two media types, one off site | A single drive in the booth is not a backup |
Storage is calculated from these settings throughout this page. Recording at 96 kHz doubles every storage figure and gains a church nothing that anyone will hear.
The archive problem nobody plans for
At 32 channels a church generates about 1.3 TB of multitrack a year, plus 164 GB of video, and neither of those numbers goes down. Within three years a single drive is full and the decision about what to keep gets made by whoever is standing there with a full drive, which is the worst possible way to make it.
Decide a retention policy early and write it down. A reasonable one keeps the compressed stereo archive of everything permanently, because it is tiny, keeps the video for a defined period, and keeps multitrack only for services somebody has specifically flagged. That converts an unbounded storage problem into a small predictable one, and it means the recordings that matter are the ones that survive.
Label the drives, keep one copy off site, and check once a year that an old file still opens. Storage fails silently, and the moment you discover the archive is corrupt is always the moment somebody asks for a recording from four years ago. Put that annual check on the maintenance schedule alongside the other things that quietly rot.
Finally, remember that the recording is only ever as good as the mix feeding it and the microphones feeding that. A multitrack of a badly captured service is a high resolution copy of a problem. If the recordings are disappointing, the fix is usually upstream, in the feed described in livestream audio setup or in the microphone choices behind it, not in the recorder.
Sources
- Storage figures calculated from 24-bit 48 kHz uncompressed audio at 0.52 GB per channel per hour
- Manufacturer published specifications for the consoles, interfaces and switchers named
- Published streaming platform encoder guidance for the video bitrate figures
Frequently asked questions
How much storage does recording a church service need?
At 24-bit 48 kHz, audio uses 0.52 GB per channel per hour. A stereo board mix of a 90 minute service is 1.6 GB, a 16 channel multitrack is 12.4 GB and a 32 channel multitrack is 24.9 GB. Adding a 1080p video file at the standard streaming bitrate contributes 2.1 GB per hour. Over a year of weekly services, 32 channel multitrack alone reaches about 1.3 TB.
Should we record multitrack or just a stereo mix?
Record multitrack only if somebody is genuinely going to remix the service afterwards, which is several hours of work each week. For an archive, a sermon podcast or a recording of a service someone missed, a stereo record of the broadcast bus is completely adequate and uses a sixteenth of the storage. Most churches that build multitrack workflows stop using them within a year.
What levels should we record at?
Aim for an average around -20 dBFS with peaks no higher than -6 dBFS, set during the loudest part of rehearsal rather than during the sermon. At 24-bit there is no penalty for recording conservatively, and the headroom protects you from the one chorus that is louder than expected. Digital clipping is permanent, so err downward every time.
Can we record the service on the mixing console itself?
Often yes, and it is the most reliable option because no computer is involved. Many digital consoles record a stereo file straight to a USB stick, and several record multitrack to an attached drive. Check what your console already does before buying an interface. The main limitation is that console recorders rarely let you monitor the recording while it happens, so the verification check matters more.
What sample rate should a church record at?
Forty eight kilohertz, because it is the video standard and keeps audio and video in step with no conversion. Recording at 96 kHz doubles every storage figure on this page and delivers nothing a congregation will ever hear. Bit depth is where the real benefit is: 24-bit gives enough headroom that conservative recording levels cost you nothing at all.
How should we back up service recordings?
Three copies, on two different kinds of storage, with one copy off site or in cloud storage. A single external drive in the booth is not a backup. Copy files off the recording machine the same day, name them consistently with the date first, and check once a year that an older file still opens, because storage media fail silently and you only find out when somebody asks.
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.