Loudness Targets for Podcasts and Streaming

Understand the signal and the goal

Loudness measurement estimates perceived program level over time, while peak measurement describes signal maxima according to the meter used. A delivery target is a recipient's specification, not a universal creative setting. Podcast hosts and music services may measure, normalize, encode, or present audio differently.

Integrated loudness summarizes a measured program, short-term and momentary views show changing sections, and true-peak estimation looks for reconstructed peaks between stored samples. Normalization changes playback or file gain according to a system's policy; it does not repair clipping, balance, dynamics, or intelligibility.

Prepare a controlled session

Find the current specification for the exact host, distributor, file type, channel layout, and submission path. Record the source and date because requirements can change. Complete editing and mixing first, keep a lossless master, and choose a meter whose method matches the requested measurement.

Measure the complete program from its intended start through all tails, inspect loud and quiet passages, and adjust with mix balance, automation, or controlled limiting only as needed. Remeasure after every material change. Create separate deliveries when recipients genuinely require different specifications rather than repeatedly overwriting one master.

Make and judge the change

Check integrated measurement, local intelligibility, peak behavior, distortion, pumping, quiet sections, transitions, and codec results. Listen at a stable monitor setting and compare with appropriate references without forcing identical dynamics. A meter pass is necessary for a stated requirement but does not prove a good mix.

Use output gain when only overall level needs correction and dynamics remain suitable; revisit the mix when one section dominates the measurement or speech is unclear. Every numeric target, limiter ceiling, release, and gain value must be labeled as a starting point or recipient requirement, then checked by listening and metering.

Check, document, and deliver

Render the requested format, channels, loudness method, peak constraint, metadata, and naming, then reopen and remeasure the delivered file. Listen after lossy encoding when applicable. Archive the lossless master, the exact measured delivery, the specification source, meter version, and measurement date.

There is no single permanent loudness target for every podcast, song, platform, region, or workflow. Services may change policies and may normalize playback without changing uploaded files. Matching a number cannot guarantee equal perceived loudness, translation, quality, listener comfort, or freedom from codec artifacts.

Worked example: deliver measured loudness

An editor must deliver the same spoken program to two publishers whose current submission documents specify different file and measurement requirements.

1. Save both current publisher documents with dates, identify their loudness method, peak rule, channels, codec, and metadata requirements, and flag any ambiguity.
2. Finish one balanced lossless master, measure the complete program, and repair any local speech or dynamics issue before changing overall output gain.
3. Create separate named deliveries according to each verified specification, then reopen, remeasure, and listen to the encoded files for distortion and transition problems.
4. Archive the common master, both deliveries, meter report, specification sources, and delivery notes without calling either requirement a universal standard.

Loudness delivery record

Use this record to repeat and review loudness targets for podcasts and streaming without losing the source, context, or reason for each choice.

• Recipient, specification URL or document, access date, program type, region, submission path, and contact.
• Required format, codec, sample rate, bit depth where applicable, channels, loudness method, peak rule, and metadata.
• Meter and version, complete-program range, integrated result, local ranges, peak result, and mix issues found.
• Gain, automation, dynamics, or limiter change, starting value label, listening result, and remeasurement.
• Lossless master, encoded delivery, reopened-file result, codec audition, report, filename, and archive.

Common mistakes

• Publishing one remembered podcast or streaming number as a permanent target for every service and program.
• Normalizing an unclear or clipped mix and assuming a measured result repairs intelligibility or distortion.
• Measuring only a loud excerpt instead of the complete intended program and its full beginning and ending.

Try one

Sources

• Audacity loudness normalization guideOfficial Audacity documentation for measured loudness normalization and its controls.
• Audacity Export Audio guideOfficial Audacity guidance for file formats, export ranges, audio options, metadata, and mixing on export.