Podcast episodes sound more consistent on a YouTube stream when you even out their perceived loudness before encoding. YouTube’s guidance reviewed here gives audio encoding recommendations, but it does not establish a podcast-specific LUFS target or require a particular normalisation level.
Treat processing and encoding as separate jobs: first prepare and listen to the episode, then choose the format and settings for the upload or live encoder. This gives you a repeatable workflow without mistaking bitrate for a volume control.
Why podcast episodes sound uneven
A podcast can vary in loudness within a single episode and from one episode to the next. One guest may speak close to a microphone while another sits further away; a quiet introduction may be followed by a forceful advertisement; or an older recording may have been mixed differently from the current one. A player or stream can reproduce these differences, but it cannot infer which variation you meant to keep.
A single gain adjustment raises or lowers the whole file. If the host is too quiet throughout, that may help. If the host is quiet but a guest is already loud, raising the whole programme makes the imbalance worse. Whole-programme loudness normalisation likewise cannot repair every uneven passage on its own: it measures and adjusts the overall programme level, not the meaning of each moment.
Start with the mix. Check dialogue levels between speakers and sections, and correct obvious imbalances in the editor before applying overall loudness processing. If an advert, music bed or intro is deliberately louder, you may decide to retain that contrast. The aim is consistent listening, not flattening every expressive change.
The issue becomes more noticeable when episodes run one after another. A quiet episode followed by a louder one can make a listener reach for the volume control at each transition. For a continuous music or spoken-word channel, the same principle applies to joins between pieces; the practical checks in this guide complement advice on switching songs smoothly in an always-on stream.
Loudness and encoding are different jobs
Loudness processing changes the perceived level of the programme. Encoding settings describe how audio is represented and delivered as part of a file or live signal. Changing a bitrate does not normalise volume, and choosing a sample rate does not make a quiet speaker louder.
YouTube’s upload troubleshooting guidance discusses audio format, bitrate and sample rate, and advises editing audio and video tracks in editing software before uploading when troubleshooting issues. Its live encoder guidance likewise lists settings for the incoming audio stream. Neither page establishes a podcast LUFS target. In particular, do not present a commonly repeated LUFS figure as a YouTube requirement unless current official guidance actually says so.
| Task | What it changes | Where to handle it |
|---|---|---|
| Correct speaker or section imbalance | Relative level within the episode | Your audio or video editor |
| Normalise programme loudness | Overall perceived programme level | An editor or loudness-processing tool |
| Control peaks | Brief maximum levels after gain changes | A limiter or other peak-control process, used carefully |
| Choose codec, sample rate and bitrate | How audio is encoded for delivery | Export settings or the live encoder |
Keeping these stages distinct makes troubleshooting clearer. If a processed file still sounds too quiet, changing AAC bitrate is not the first fix. Revisit the mix and loudness processing, then check that the correct file and audio track are being sent to the stream.
Choose a repeatable normalisation workflow
Work from the finished episode mix, rather than trying to normalise the raw recordings as a substitute for mixing. First balance the dialogue and other important elements. Then measure integrated loudness with a capable editor or loudness meter. Choose a consistent workflow target that suits your show and listening context; do not call it an official YouTube target. The important thing for a series is to apply the same decision method to each episode and listen to the results.
Some editors offer a loudness-normalisation function, while dedicated tools may provide metering and batch processing. You do not need a particular product to begin. Check whether the tool measures integrated loudness, whether it lets you set a repeatable target, and whether peak control is a separate step. Batch processing can save time for a backlog, but it is not a substitute for checking episodes with unusual dynamics or a different recording chain.
A practical sequence is:
- Finish the mix and correct obvious speaker or section differences.
- Measure the full programme, not just a short loud passage.
- Apply your chosen loudness process consistently across episodes.
- Inspect peaks and use restrained peak control if needed.
- Export, then listen to the rendered file and its joins.
If you use a recorder or dedicated appliance that offers optimisation, treat it as a convenience rather than a requirement. For example, Zoom’s PodTrak P8 documentation describes analysing and automatically optimising a file’s loudness for podcasting. That feature does not make the device necessary: software you already use may be enough for a finished episode.
Write down the workflow decisions that matter: which mix version is final, which loudness measurement you use, what processing you apply, and which export preset serves the destination. This short record helps you identify why two episodes differ without relying on memory. Avoid changing the target simply because one episode contains a naturally quiet interview; first decide whether the mix needs correction or whether the variation is intentional.
Check peaks and listen for artefacts
Raising overall loudness can bring peaks closer to the maximum level your file or signal can represent. If a peak needs control, a limiter can reduce it, but aggressive limiting may make speech sound harsh, flattened or tiring. Pumping, where the level seems to dip and swell around loud words or music, is another warning. The reviewed YouTube sources do not set a mandatory true-peak ceiling or a required limiting method, so do not invent one and attribute it to the platform.
Listen to the rendered file, not only the editor preview. Check the opening, a quiet section, the loudest passage, and any music or advert transitions. Listen for clipping, distortion, breaths or room noise made conspicuous by processing, and abrupt changes between speakers. A loudness meter helps make the process repeatable; it cannot tell you whether a guest suddenly sounds unnatural.
For a long episode, you can inspect sections that are likely to reveal problems: a guest change, a change from speech to music, or a passage that was particularly quiet in the original mix. Keep the original export or project available until the processed version has passed these checks. If a section sounds wrong, return to the mix or adjust the processing rather than adding more gain indiscriminately.
Headphones can help reveal noise and harshness, but there is no required model in the YouTube guidance cited here. Check on speakers or another ordinary listening setup too, if available; the point is to hear the programme at a sensible level in more than one context, not to buy specific equipment.
Set audio encoding separately
Choose encoding settings for the route the audio takes. For uploaded video, YouTube’s troubleshooting page lists AAC-LC, a sample rate of 44.1 or 48 kHz, and a bitrate of 128–256 kbps. For live stereo audio, its encoder guidance lists AAC or MP3, 44.1 kHz and 128 kbps. These figures describe encoding recommendations, not loudness targets. Use the appropriate path rather than combining upload advice and live-encoder advice as if they were a single preset.
| Delivery path | Audio guidance in the cited YouTube documentation | What it is not |
|---|---|---|
| Uploaded video | AAC-LC; 44.1 or 48 kHz; 128–256 kbps | A loudness or LUFS instruction |
| Live stereo encoder input | AAC or MP3; 44.1 kHz; 128 kbps | A volume-normalisation process |
The practical options and trade-offs are straightforward. An upload preset in your editor keeps export choices alongside the finished episode. A live encoder preset applies to the audio you send during a broadcast. If the episode is rendered into a video before upload, use the upload guidance; if audio enters a live encoder, consult the live guidance. Do not change encoding figures to solve uneven speech levels.
These recommendations are from YouTube’s upload troubleshooting guidance and its live encoder settings page. Check those pages for current details before changing a production preset. They describe delivery settings; they do not say that every podcast must use a particular loudness, codec for every workflow, or mastering algorithm.
If you maintain a separate preset for a recurring stream, label it clearly and note whether it is for a pre-rendered upload or live encoder input. That avoids an easy hand-off mistake when an episode is prepared by one person and streamed by another. You can also compare the encoding choices with a bitrate and audio settings guide for a 24/7 music stream, while keeping this episode’s loudness decisions in the audio workflow.
Test the processed episode in the stream
A file that sounds correct in an editor can behave differently once routed through the stream, particularly if the wrong audio track is selected or another gain control is active. Run a test before relying on a recurring schedule. YouTube recommends a pre-stream test with audio and movement similar to the planned broadcast, and recommends monitoring stream health during the event. Its live streaming tips are useful when planning that check.
Use a representative passage rather than a silent waiting screen. Include ordinary dialogue, a louder passage, and a transition into the next episode if your channel plays episodes continuously. Listen at the receiving end where practical, and watch for audio dropouts or a level change that was not present in the rendered file. This is an application of YouTube’s testing and monitoring guidance, not a separate platform loudness rule.
When an episode is part of a long-running channel, test the joins as well as the episode itself. A clean ending followed by a much louder opening can still unsettle a listener even if both files were processed consistently in isolation. Compare the last speech or music in one file with the first moments of the next, then adjust the mix or transition if the contrast is distracting. For a channel built around repeated content, a reusable stream key setup can simplify the broadcast arrangement, but it does not replace audio checks.
If the stream sounds different from the rendered file, trace the signal path one step at a time: confirm the intended media file, the selected audio source, and any encoder or player gain setting. Do not respond by altering bitrate unless there is evidence of an encoding issue. Keep notes on the test result so the next episode uses the same known-good path.
A cloud-run broadcast can remove the need to keep your own computer on while episodes play continuously. For that specific operational problem, StreamNeo turns an uploaded video into a YouTube live stream, so you can prepare the file and channel without leaving a local computer running overnight. The loudness work still belongs in your episode preparation: listen to the processed file and test the stream before committing to a repeat schedule.
Before committing, compare the operating options on the pricing page. When the file and channel are ready, start free — 24-hour trial, no card.
FAQ
How do I normalise podcast audio before uploading it to YouTube?
Correct obvious imbalances in the mix, measure the finished programme, and apply a consistent loudness workflow before export. Check peaks, listen to the rendered file, and use an upload encoding preset separately from that processing.
What LUFS should a podcast be for YouTube?
The official YouTube guidance reviewed for this article establishes no podcast-specific LUFS target. Choose a consistent production target for your own audience and distribution plan, and do not describe it as a YouTube requirement.
Does YouTube normalise podcast volume for live streams?
The cited live encoder guidance covers audio encoding settings, testing and monitoring, not a podcast loudness-normalisation target. Prepare and listen to the audio before sending it to the live stream, then check how it sounds in a test.
What audio bitrate and sample rate should I use for a YouTube stream?
For live stereo encoder input, YouTube lists AAC or MP3, 44.1 kHz and 128 kbps; for uploaded video, its troubleshooting page lists AAC-LC, 44.1 or 48 kHz, and 128–256 kbps. Select the recommendation for your delivery path, and remember that none of these settings normalises volume.