Radio and podcasts for DJs: programmes, jingles, beds and loudness.

A radio broadcast is not a live set with a different poster. Timing changes: scheduled blocks, spots, an opening and closing voice, and a sound level lasting a full hour without listeners adjusting volume. Podcasts take that logic further: episodes are distributed through a feed and heard at any time rather than a fixed hour.

This guide covers broadcasting music with voice: what a Radio DJ does, constructing a programme with a running order, the roles of jingles, beds and talkover, publishing a DJ/music podcast through RSS and measuring broadcast loudness.

What a Radio DJ does today

A Radio DJ programmes and presents: selecting and ordering music, speaking over it and, above all, maintaining a recognisable thread. Four specific responsibilities:

• Timing. Always know which minute you are at and what follows. Broadcasting without a clock quickly lets a block overrun and push back the next spot.
• Identity. A voice, coherent musical treatment and signature listeners recognise by the third second.
• Level. Voice, music and stings sounding at the same level throughout.
• Technical consistency. The same broadcast chain weekly: server, encoder and metering. For your own broadcast, server documentation such as Icecast’s helps explain the audio’s journey from your equipment to the listener’s player.

The difference from a club is discipline: in the booth you improvise around the dance floor; in the programme, improvisation happens within a predetermined structure.

Building a programme: running order and structure

Building a programme starts with three decisions made before recording:

1. Concept and duration. An hour of music with commentary, fifteen minutes of new releases, a monthly special. Duration determines block count and supporting material.
2. Running order. Divide time into blocks: opening, music block one, speech, block two, spot and closing. Note each element’s start and end minute. This document saves you when live broadcasting goes off course.
3. Supporting material. A script with time marks, track introductions, artist details and, of course, a labelled, accessible folder of jingles, beds and clips.

Repeatable programmes use templates: the same blocks, opening/closing order and level treatment. Music selection supplies variety rather than the skeleton. A useful production rule: always record the entire broadcast even if publishing only part, avoiding a block remake after a timing error.

Jingles: the programme’s identity in thirty seconds

Jingles are identity clips: programme name, musical signature and sometimes presenter voice. They mark block changes and orient listeners who have just tuned in.

Work with them well:

• Usage order. Opening jingle, transitions between blocks and closing jingle. No new one is needed every five minutes; three well-produced pieces cover a broadcast.
• Level. A louder jingle tires listeners and forces live master adjustments. Treat it as part of the same running order and use the same levelling criteria.
• Origin. Produce them yourself (your own voice and music) or buy from an imaging library. In the latter case, check public-broadcast coverage: a studio purchase does not always include broadcasting rights. The distinction between use and copyright is addressed in the guide to collecting societies and licences.
• Consistency. The same voice, music treatment and dynamic range as the programme.

Beds and talkover: music beneath speech

Beds are background tracks under your speech. A good bed has little percussion, occupies little of the voice range and develops slowly: a chorus with sung vocals is unsuitable beneath an explanation.

Talkover means speaking while music plays. Three keys:

• Distance and angle. Work slightly off-axis with gain set in the booth; adjust level by singing rather than raising the microphone fader mid-sentence.
• Relative level. A common working reference is 10–15 dB below voice, accompanying without competing. Lower the bed with its fader rather than EQ: cutting background bass leaves the music bland.
• Timing. Speak in track gaps and finish before the drop. If music intensifies, lower your voice or stop.

Voice technique is the same as for any live address: microphone technique and talkover covers microphone practice and control of the room if you need a refresher.

DJ/music podcasts and RSS distribution

A DJ/music podcast turns programme logic into on-demand episodes: edited sets, monthly selections, producer interviews or new-release reviews. Distribution is the technical difference.

• The RSS feed is the channel. It holds each episode’s titles, descriptions, audio links and artwork. Created on publishing the first episode, Spotify documentation says to enable it for other platforms.
• Hosting publishes; apps listen. Upload the file once and the feed distributes it to Apple Podcasts, Spotify and other readers.
• Technical requirements. Apple requires, among other things, a unique enclosure per episode with URL, length and type, recommending RSS audio in MP3 or AAC at 128–256 kbps stereo and 44.1 or 48 kHz.
• Cadence and duration. A weekly 45–90-minute episode suits journeys; publishing consistently on the same day matters.
• Rights. Prerecorded music requires covered rights: your own material, authorised sets or an agreement with the relevant parties. Rules vary by country; assume nothing.

Broadcast loudness

Broadcast loudness keeps the whole programme equally loud, without jumps between songs, spots and voice. Three ideas organise this section:

LUFS and LKFS measure the same thing. The same scale under different names: a perceptually weighted average of the whole content, rather than peaks. Peak measurement says little about comfortable listening.

Every destination has a target. Europe’s EBU reference is -23 LUFS with ±0.5 LU tolerance and true peak below -1 dBTP. US ATSC A/85 specifies -24 LKFS without metadata and recommends -23 to -27 LKFS for streaming. For podcasts, Apple recommends -16 dB LKFS with ±1 dB tolerance and true peak below -1 dB FS.

The process is always the same. Measure the full programme, correct gain, limit only true peak and measure again. Do not chain limiters to get louder: playback normalisation leaves an overcompressed master with less dynamic range. Metering and gain staging are explained in dBFS and LUFS meters.

The table above summarises destinations; use the row matching your broadcast rather than mixing references.

Next step

With the programme built and level measured, move to operations: platform choice, chat, rights blocks, recording and backups for each broadcast in platforms, copyright, ISO recording and backup. If the technical chain is unresolved, return to the pillar guide on live broadcasting with OBS, latency and synchronisation.