Audio for Live Broadcast: Split, Dante and Mix-Minus.

A broadcast is more than a copy of what sounds in the venue. The audience at home listens through headphones or small speakers, without the body provided by a subwoofer system, and focuses on music and voices. Sending the venue console’s master output, or FOH, unchanged usually produces a broadcast with disproportionate bass, buried voices and ambience that makes no sense.

The solution is architectural rather than a post-production trick. You need a clean copy of the input signals, a broadcast-specific mix built from them and a return for people on stage that does not feed back. This guide follows that chain, from the FOH master to the mix leaving the encoder.

From Venue Master to Broadcast Mix

Understand why the FOH master is not usually an ideal broadcast mix. The venue console is adjusted for a particular space: it compensates for room acoustics, raises the subwoofer system and mixes for a standing listener at some distance, also hearing direct stage sound. The ideal broadcast mix is heard in very different conditions and should prioritise voice intelligibility, a stable music-to-ambience balance and a consistent average level.

There is a second, more practical reason: the venue mix changes during the show. The operator makes adjustments, opens and closes channels and touches the master. All of that transfers to the stream, often at the wrong moment. If the venue console is the only capture point, any problem affects both broadcast and venue simultaneously.

The usual approach is therefore to separate responsibilities: one person or console operates the venue, while another builds the broadcast mix from the same sources. That requires a split.

Splits: Analogue, Digital, Dante and MADI

A split obtains two independent copies of the same signal without degradation or coupling. The oldest option is an analogue split: a transformer or active distribution duplicates each channel before the console. It works with older consoles and does not depend on the network, but adds noise, can colour sound and requires separate cabling for each channel.

A digital split uses the console or stagebox’s digital outputs: the signal is copied in the digital domain, without extra conversions. Sharing the same stagebox between venue and broadcast is called stagebox sharing: a single stage rack publishes all channels and two consoles receive them over the network. It is the cleanest way to distribute signals.

Within that approach, a Dante split is now the reference for installations with an audio network: the stagebox publishes channels over Dante and each receiver subscribes to those it needs. A MADI split is the alternative when transport uses MADI: distribution of 64 channels over coaxial or fibre feeding several consoles. In both cases, plan addresses, VLANs and clocking, detailed in IP Audio Networks: Dante, AES67, AVB.

The practical rule: choose the split your infrastructure already supports, and document which channels go to broadcast. A poorly documented split is the first cause of a silent stream halfway through a show.

Broadcast Console, Submix and Multitrack

The broadcast mix is built on a broadcast console: a dedicated console with stable settings, production communications and outputs designed for broadcasting. It need not be enormous; what matters is independence from the venue console and the operator hearing exactly what goes out.

Take a submix from each console: a balanced partial mix, such as drums, vocals or ambience, that can be raised or lowered as a block. Submixes reduce the number of hands needed and prevent one adjustment from upsetting the whole balance. Alongside them, multitrack output records each channel separately: even though the live broadcast uses the main mix, multitrack allows the audio to be redone afterwards.

Ambience, Presenters and Commentary

A live mix without the audience sounds like a demo. Crowd microphones above the audience or at the sides supply the ambience that gives a sense of the venue. Mix them at a low but consistent level and compress them so they do not jump with each shout.

At the other end are voices. Handheld or lapel presenter microphones introduce the show and should always sit above the music. Event or competition commentary uses its own channels, EQ and level, allowing it to be raised or lowered for the moment. Voice microphones require knowing each capsule’s response and directivity; the guides to dynamic and condenser microphones are a good starting point.

Mix-Minus, IFB and Production Return

Everything broadcast to the audience also reaches the speaker’s ears with delay. Hearing their own voice through the return disrupts a commentator. The solution is mix-minus: a personal mix with that person’s voice subtracted, so they receive music, ambience and other voices, but not their own microphone.

Mix-minus is delivered through IFB, the interruptible program return system used by presenters and commentators. A typical IFB carries program, an instruction channel and the ability for production to interrupt. On small productions, a wireless bodypack and a clearly labelled return channel suffice; large setups centralise IFB in intercom. Documenting what each person hears prevents the classic “I hear myself with delay”.

Loudness, Limiting and Lip Sync

Broadcast level is measured against the loudness standard, ITU-R BS.1770, rather than judged by ear, and maintained within a stable range throughout the show. A broadcast starting very quiet and ending at the limit makes viewers adjust volume, and platforms penalise those jumps.

To control peaks, apply gentle limiting at the broadcast output, rather than a wall crushing the music. The aim is to prevent a shout, effect or kick hit from overloading the encoder. If broadcasting video as well, watch lip sync: audio and picture can become misaligned anywhere in the chain, from switcher to encoder.

Delay, Embedding and Audio Redundancy

With a large distance between stage and screens, or between capture and broadcast, audio delay comes into play: a deliberate delay aligning the signal with picture or direct venue sound. Set it in milliseconds and check by ear with claps or a sharp hit.

Before reaching the encoder, audio usually travels inside video through embedding; separating it at a destination is de-embedding. Know where it is embedded and extracted, because these are the points where channels are lost or a wrong signal slips in. Since losing audio hurts most, audio redundancy is mandatory: two capture routes, two independent recordings and a local multitrack copy. If the internet connection fails, the recorded audio still exists.

Next Step

With the audio chain settled, the next problem is camera placement and stage lighting that does not ruin the broadcast. The next cluster guide, Broadcasting Concerts and DJ Sets: Cameras and Lighting, covers capturing the DJ master, booth and audience cameras, lighting flicker and laser risks to sensors. Before continuing, also review Switching and Production: Cut, Mix, Keyers and Macros to understand how the program accompanying that mix is cut.