Editing Spoken Audio for Malaysian Podcasts

Master at -16 LUFS integrated, keep your loudest 30-second window below -12 LUFS short-term, and always structure the conversation before you touch a fader. That is the sequence. A producer at BFM 89.9 trims a three-minute interview to 90 seconds, cutting a half-sentence from a CEO's answer without leaving a click. The edit serves the listener's time, not the host's ego. Build the story from your raw conversation first. Clean up breath noise and filler words next. Adjust level last. Editing in that order stops you polishing a section you later cut entirely. The goal is information density, not a tidy waveform.

Make structural decisions before sonic ones. For a Malaysian talk show the sequence is structure, then clarity, then level. Structural work means selecting and arranging sections. Find the one idea in each that advances the episode. Everything else goes. This is how BFM 89.9 production material gets its density: a 45-minute raw interview might yield 12 minutes of air because the editor prioritises the information the listener came for.

Mark your in and out points on every section. Do not delete anything yet. Lay them out in a rough narrative order, then listen through as a listener would. Cut any section that repeats an earlier point, any tangent that does not pay off, and any setup that runs past 20 seconds. A business show cannot afford rambling intros. The Malaysian audience expects you to respect their commute time. Lock the structure, then bounce a rough cut and step away for an hour. Come back and listen in a single pass. If your attention drifts, cut that section. No structural edit survives the first cold listen unchanged.

Malaysian talk formats edit for information density and listener time. Conversational formats edit for rhythm. The distinction changes where you cut. In a business format like BFM 89.9, cut every sentence that does not carry a number, a fact, a named source, or a decision. If an answer can be paraphrased in the host's summary, cut the answer and let the host bridge it. Listeners to this format are often driving or on a train. They cannot rewind. One idea per section is the rule. Two ideas in one sentence? Split them with a sharp transition, even if it sounds slightly unnatural. Clarity over smoothness.

Conversational formats, including the true-crime and interview shows that dominate Malaysian post-production work, edit for rhythm. Here the host's personality is part of the content. Keep a laugh, a hesitation, a reaction noise because it carries emotional information. The edit point stays clean, but the section length follows the natural rhythm of the exchange. A good conversational edit removes dead air and leaves the back-and-forth pace intact. The trade-off is between tightness and warmth. Err toward tightness for business. Err toward warmth for conversation.

Level comes last, and it must follow platform variation. No single LUFS target works across Spotify, Apple Podcasts, and direct download. The common advice to master at -16 LUFS for stereo material is a starting point, not a rule. Spotify normalises playback to roughly -14 LUFS integrated. Apple Podcasts uses -16 LUFS. Master at -14 LUFS, and your episode gets turned down on Apple Podcasts, sounding quieter relative to other shows. Master at -19 LUFS, and your episode gets turned up on both platforms. The normalisation process may introduce audible limiting you did not intend.

Setting Your Integrated and Short-Term Levels

Set your integrated loudness to -16 LUFS. Measure your short-term loudness during the loudest 30-second window. That window must not exceed -12 LUFS short-term. This gives you a file that survives platform normalisation without distortion. Use a meter plugin on your master bus. Youlean Loudness Meter or the built-in meter in Reaper are both adequate. Do not mix by ear and check the level at the end. Level is a measurement step, not an aesthetic one.

Gain Staging Before the Limiter

Gain staging through your signal flow is why most local mixes fail before they reach the limiter. If your raw recording peaks at -12 dBFS, you have enough headroom to edit, process, and land at your target without heavy limiting. If your recording peaks at -3 dBFS, you will be pulling levels down at every stage and fighting headroom. Set your interface preamp so the speaker's average level sits around -18 dBFS with peaks at -12 dBFS. This is not the broadcast standard. It is the right starting point for spoken word with dynamic range.

Your monitoring setup decides whether your level decisions are real. Mix on headphones if you must, but check your final episode on a single small speaker at low volume. This is the car test. It matters more in Malaysia than in markets where most listening happens on earbuds. A large share of listening here still happens in cars via Bluetooth. If your voice level drops below the road noise at 70 km/h on a Malaysian highway, your loudness is wrong regardless of what the meter says.

The RTM broadcast standard is a useful reference for dialogue clarity, not level. RTM speech material is mixed to a higher average level with less dynamic range than a show. Do not target RTM levels for an episode. The podcast standard is lower and more dynamic. What you can borrow from the RTM and BFM 89.9 production standard is the treatment of sibilance. De-ess aggressively on any voice that sounds harsh on a small speaker. A de-esser set to reduce 5 kHz to 8 kHz by 3 dB to 6 dB on the worst sibilants is your starting point. Listen for the "s" and "t" sounds. Malay and Mandarin speakers often produce sibilance at slightly different frequencies from English speakers. Sweep the de-esser frequency until you hear the harshness drop, then back off 1 dB.

A transition between two sections should be inaudible in a business format. Use a 2ms to 5ms crossfade at the edit point. If the background noise changes between the two clips, the transition will still be audible as a noise floor jump. Fix this by adding a short room tone clip under the transition. Record 30 seconds of room tone at every session. Without it, you have nothing to bridge a noise floor shift. A transition that clicks or jumps in background noise signals amateur post-production faster than any other single mistake.

The trade-off every editor faces on a deadline is between a perfect edit and a finished episode. On a weekly show, you cannot spend four hours on a 30-minute episode. Set a time budget: 90 minutes of editing per 30 minutes of raw conversation for a business format, and 60 minutes for a conversational format. If you exceed that, your recording quality or your editing technique needs work. A good monitoring setup and clean gain staging at the recording stage will save you more time in post-production than any editing trick.