Configuring RƎLAY Auto-Trim for Audio Normalisation
RƎLAY’s Auto-Trim prepares pre-recorded voice for consistent playback by trimming excess silence, matching programme loudness and controlling peaks. Used with editorial judgement, it removes repetitive adjustments without making speech sound compressed or mechanical.
The practical shift is from correcting each file by ear to applying a named, repeatable preset. That matters for Australian teams handling overnight contributions, interstate promos and material made for both broadcast and on-demand platforms.
Understanding what the tool normalises
Auto-Trim addresses two kinds of variation: structure, such as silence at the beginning, end or between phrases; and level, such as overall loudness and isolated peaks. It cannot restore a clipped consonant, remove room echo or replace a damaged recording, so source quality remains essential.
For pre-recorded material, every change is baked into the exported file. Keep the untouched master, process a working copy and confirm how the station’s playout processor will handle the result. Applying loudness control twice can flatten dynamics or make a promo jump against neighbouring content.
Preparing material for Australian workflows
Inspect files before upload and record their sample rate, bit depth and channel layout. Avoid needless conversions, and be cautious with material that has already been heavily limited; the detector needs some remaining dynamics to judge it correctly. Include a clear revision suffix and Australian Eastern Standard Time or Australian Central Standard Time in production notes.
Australian radio markets combine public services such as ABC/FM and SBS, dense metropolitan commercial networks and regional stations, so local news and traffic updates often sit beside national campaigns. Broadcasters must meet the conditions attached to licences issued under the Broadcasting Services Act 1992 and relevant ACMA codes. An EBU R128 or ITU-R BS.1770 target may support consistent operations, but it does not replace a broadcaster’s legal, platform or commissioning specifications.
For studios that combine RƎLAY with physical interfaces, playback devices or monitoring, the OEM hardware guidance can support a separate review of those requirements, keeping procurement decisions distinct from the audio preset.
Setting loudness and trim parameters
Set the integrated loudness target from the station’s written delivery specification rather than from a generic preset. Around –23 LUFS is common as a broadcast reference, while stereo podcast workflows often use –16 LUFS; these are starting points, not Australian regulatory limits. Set the loudness range so natural emphasis remains intact.
Choose a true-peak ceiling, commonly –1 dBTP, that suits the downstream encoder. Configure silence detection with enough handle time to avoid clipping soft consonants or cutting off breaths. Pause trimming should improve pacing, not create the rapid, clipped rhythm associated with over-edited speech.
Apply short, even fades at file boundaries and preview every change with meters visible. Listen to the processed file as well as the original, then save the values as a named profile so operators do not have to infer the settings from an earlier session.
The practical differences between the untouched material and the processed export are:
| Check | Unprocessed source | Auto-Trim result | Practical check |
|---|---|---|---|
| File boundaries | Unpredictable silence | Configured clean boundary | First and last words remain complete |
| Programme loudness | Source-dependent level | Consistent integrated target | Measure every export |
| True peaks | Occasional overloads | Controlled ceiling | No distortion after encoding |
| Internal pauses | Variable duration | Shortened where configured | Speech still sounds natural |
| Channel layout | As supplied | Preserved or deliberately converted | Matches the playout requirement |
Testing on representative recordings
Build a small test batch containing a quiet interview, a loud promo, a file with long pauses and a recording with a strong regional accent. Australian English varies considerably across the country, so consistency should come from the level target, not from erasing delivery character. Run every file through the same versioned profile.
Listen on studio monitors, headphones and the devices listeners use. Compare each export with its source, check integrated loudness and true peak, and note any altered consonants, abrupt edits or changes in background noise. Keep exceptions as documented overrides rather than silently altering the master preset.
Checking output across studios and devices
For teams spread across several studios, VisTool’s centralised monitoring dashboard provides a practical way to review output consistently and spot a level mismatch. The same checks used for a Sydney news file should apply to a Perth overnight package, with results recorded against the asset rather than left in a chat message.
Test the material through the actual distribution chain, including DAB+, online radio and on-demand feeds where used. Many Australians check broadcasts and podcasts during commutes, so a car system, a phone speaker and earbuds can expose problems that studio monitors conceal. A file that sounds acceptable in isolation may still need adjustment beside adjacent programme items.
RƎLAY’s processed output should also meet the same handoff expectations as content moving to Ruby consoles, Power Core nodes or VisTool monitoring. Keep pre-recorded and live sources distinguishable, and confirm that the playout system is not applying another aggressive limiter to material already controlled by Auto-Trim.
Turning settings into a repeatable standard
Store the approved profile with a version, owner, target loudness, peak ceiling and trim-handle settings. Retain the original recording, the exported master and a short quality-control record; sponsorship, news or legal material may also need retention under the broadcaster’s policies. Updates should create a new profile so past exports remain reproducible.
The immediate next step is to select one representative 30-second voice file, save the agreed values as a versioned Auto-Trim preset, and run it across a five-file test batch before station-wide use.