The reliable way to keep a continuous YouTube podcast stream at a consistent volume is to measure each finished episode with the same integrated-loudness method, then adjust it to one documented house target. Do not rely on YouTube’s playback features to repair differences between episodes.
YouTube can change how a viewer experiences dynamics through controls such as Stable volume and Voice boost. Those controls are useful, but they vary by viewer, device and video, so you still need to prepare the episodes consistently before putting them into the stream.
Why episode-to-episode loudness varies
An episode can sound louder or quieter than the one before it even when both files have the same peak level. Peak level describes the highest moment in the waveform. It does not describe how loud the whole programme feels over time.
A quiet interview with long pauses may have a lower average level than a fast conversation with little space between sentences. A guest recorded close to a microphone may sound much louder than a guest sitting further away. Intro music, telephone recordings, room noise and deliberate pauses add further differences.
There can also be a production-history problem. One episode may have been edited by raising the final gain, while another was compressed more heavily during mixing. If you normalise only by looking at the waveform or setting the highest peak to a particular value, the episodes may still feel mismatched.
The continuous stream makes this more noticeable. A viewer might listen through an hour of one episode, hear a loud musical introduction, then encounter a quieter guest recording in the next episode. On a normal podcast feed, the listener may choose another episode or adjust the player. On an always-on channel, the change happens while the viewer is already listening.
You do not need to make every second equally loud. Speech should retain pauses and emphasis, and music should still rise and fall. The practical aim is for complete episodes to sit in a similar loudness range when played at the same device volume.
This is separate from the question of how to deliver the files continuously. For the playlist and stream side, see the guide to streaming podcast MP3 files to YouTube Live using FFmpeg. First make the source episodes consistent, then choose the mechanism that keeps them playing.
Measure and adjust finished episodes consistently
Measure the final programme after editing and mixing, not an early voice track or an unfinished sequence. A finished episode includes the speech, music, silence, adverts or sponsorship messages, intro, outro and any transition you intend to publish.
Use integrated loudness as the main comparison. Integrated measurement considers the programme over its full duration, so it is more useful for matching complete episodes than a peak meter or a reading taken from one sentence. Keep the measurement method unchanged between episodes.
A repeatable workflow looks like this:
- Export or open the completed episode as one file.
- Measure its integrated loudness in LUFS.
- Record the result, along with true peak and loudness range if your meter provides them.
- Apply the same house target used for other episodes.
- Export a new master without changing the stereo relationship.
- Listen to the adjusted file for clipping, pumping, harsh speech and missing detail.
- Compare it with a representative earlier episode at matched playback volume.
The target is a production decision, not a number that YouTube has universally required for podcasts. The official YouTube guidance reviewed for this subject gives upload encoding recommendations and playback information, but it does not establish one LUFS target for creator-made continuous podcast episodes.
The European Broadcasting Union’s EBU R 128 recommendation gives a broadcast reference of −23 LUFS average programme loudness and includes Loudness Range and Maximum True Peak Level as useful descriptors. That is broadcast guidance, not a YouTube podcast mandate. You can use it as a reference when designing your process, but do not present it as a platform rule.
Audacity’s documentation describes a direct loudness normalisation function under Effect, Volume and compression, Loudness Normalization. Its manual uses −23 LUFS as a reasonable spoken-word broadcast default and notes that streaming targets are often a few LU higher. Again, treat that as software guidance and an editorial reference, not evidence of a required YouTube setting. The Audacity loudness normalisation documentation explains the operation and its controls.
Choose one house target for your show, write it in the production template, and use it for every episode unless the format changes. If you later decide that the target does not suit your speech, music or audience, change the documented process deliberately rather than allowing each editor to make a separate judgement.
Keep ordinary stereo channels linked when applying gain. Independently normalising the left and right channels can change the stereo balance. Separate-channel correction is appropriate only when the two channels are genuinely separate recordings and you intend to correct them independently.
Do not stop at integrated loudness. Check true peak so that the adjusted file does not create clipping during conversion or playback. Check loudness range so you know whether the programme contains unusually wide changes between quiet and loud sections. Then listen. A meter can tell you that two episodes are close on paper while one still has distorted consonants or an over-compressed music bed.
Compare episodes using the same listening setup
Metering makes the process repeatable, but listening tells you whether the result works for the actual programme. Use the same headphones or speakers, the same device, the same player and the same volume position when comparing a new episode with an earlier representative one.
Do not compare one file on studio headphones with another through a phone speaker. The difference in playback equipment can be larger than the difference you are trying to judge. You do not need expensive equipment for this check. You need a repeatable setup and enough attention to hear speech clarity, music level and changes in density.
Choose comparison points rather than listening only to the first minute. Check an ordinary conversation, a quiet section, the loudest natural speech, the intro music and the outro. If the show includes different types of guests, use a sample that represents the range of recordings you normally publish.
A simple comparison record can contain:
| Check | What to compare | What to do if it differs |
|---|---|---|
| Integrated loudness | Whole-episode perceived level | Recheck the measurement and house target |
| True peak | Highest reconstructed peak | Reduce gain or revise limiting if it is unsafe |
| Loudness range | Quiet and loud programme sections | Decide whether the contrast is editorially intentional |
| Speech clarity | Words, consonants and pauses | Revisit compression, noise reduction or source gain |
| Stereo balance | Left and right relationship | Keep channels linked unless correction is intentional |
| Mono fold-down | Speech and music after summing | Fix phase or channel-specific problems before export |
The last row matters for viewers using phones or small devices. YouTube notes that stereo audio may be converted to mono on many mobile devices with a single speaker. A file that sounds acceptable in stereo can lose part of its speech or music when folded down if the channels have phase or balance problems.
Use the same comparison method for every batch. If you are preparing ten older episodes, do not adjust the first one by ear, the next one by peak level and the rest by LUFS. That creates a new inconsistency while attempting to remove the old one.
For a channel that publishes from a small computer, the audio workflow and the delivery workflow are separate jobs. The article on streaming a podcast playlist from a Mac mini may help with the delivery arrangement, but it does not replace measuring the finished files.
Check transitions in the continuous stream
An episode can pass its individual loudness check and still create an awkward stream when it follows another episode. The hand-off is where listeners notice sudden silence, a loud jingle, a duplicate intro or a change in room tone.
Build a short test sequence containing the end of one episode, the planned gap or transition, and the beginning of the next. Listen through it at the same level you used for episode comparisons. Check both a normal transition and the most difficult pairing in your queue, such as a quiet interview followed by a music-heavy introduction.
Look for these specific problems:
- The first words of the next episode arrive much louder than the preceding speech.
- A fade ends before the next file begins, leaving an unintended silent gap.
- An outro and the next intro overlap when the playback tool is configured to crossfade.
- A normalised episode has a much louder advert or sponsor message inside it.
- The first seconds of an episode are quiet because the editor measured the whole file but did not check the opening.
- The stream repeats an intro or a file because the playlist order does not match the intended schedule.
Avoid correcting a transition by turning up one entire episode without measuring it again. If only the intro is too loud, repair the intro in the source project and remeasure the finished episode. If the gap is the problem, change the playlist or transition settings rather than altering programme loudness.
A short test stream can reveal issues that a file-by-file review misses. If your channel uses a static image with audio, the visual side may be simple, but the sequence still needs to be checked from the viewer’s position. The guide to setting up a YouTube 24/7 stream with a static image and audio covers that broader arrangement.
If the stream is assembled on a computer or virtual machine, also test what happens after a file ends, the playback application restarts or the connection returns. Volume consistency is valuable only if the intended files actually play in the intended order.
Understand YouTube viewer-side volume controls
Creator-side mastering is not the same as playback-side processing. YouTube Help states: “By default, stable volume is turned on when you’re watching videos on YouTube.” Stable volume balances quieter and louder parts where it is available, and viewers can turn it off in the player.
YouTube also documents Voice boost, which reduces background sounds and highlights vocals. When Voice boost is on, Stable volume is automatically off. These controls can alter perceived dynamics, particularly in a spoken podcast with music or room sound beneath the voice.
Read the current YouTube Help guidance on controlling video volume before explaining these controls to your audience. Availability can depend on the video, device or account experience. You should not assume every listener has the same option, or that the option processes every part of a continuous stream in the same way.
This is why viewer-side normalisation cannot replace a consistent master. A viewer may disable Stable volume, use Voice boost, listen through a mono phone speaker, use a television sound mode or change the device volume manually. Another viewer may hear the same episode through headphones with all of those features unavailable.
You also cannot use the presence of Stable volume as a reason to make poorly matched files. It may reduce some variation for some viewers, but it does not guarantee that two episodes will sound identical. The creator remains responsible for preparing a coherent programme.
Do not describe −14 LUFS as a universal YouTube podcast requirement. It is often repeated in creator discussions, but the official pages reviewed here do not state it as a mandatory target for continuous podcast uploads. Choose and document a house target based on your programme, measurement method and listening checks.
Listen to the complete stream before publishing
A test of individual episodes is necessary but not sufficient. Before publishing, listen to the complete sequence that will be sent to YouTube, or at least a representative run that includes every type of transition and source recording.
The purpose is not to sit beside the stream for an entire day waiting for a fault. It is to verify the assembled result rather than assuming that correct source files guarantee a correct broadcast. Confirm that the episode order, gaps, fades, audio routing and loop behaviour are what you intended.
Use a checklist with named points in the schedule. Include the start of the stream, the first transition, the loudest music section, a quiet conversation, a sponsor or advert if present, a mono playback check and a later transition after the system has been running.
Export settings are part of the check. YouTube’s recommended upload encoding settings list audio bitrates of 128 kbps for mono, 384 kbps for stereo and 512 kbps for 5.1. These are upload audio recommendations, not loudness targets. Select the setting that matches the channel format and verify the final file rather than changing a loudness decision to meet a bitrate number.
If your stream is expected to run while the main computer is switched off, remove that computer from the final test only after the cloud or remote arrangement has been checked. StreamNeo removes the specific burden of leaving your own computer running by taking an uploaded video, your YouTube stream key and the continuous broadcast into one monitored workflow that can restart the stream if it drops.
That does not make an episode mix correct, and it does not alter YouTube’s viewer-side controls. It simply separates the finished-file check from the need to keep a local machine awake. You still need to verify the uploaded file, stream order and transitions before relying on the channel overnight.
Set expectations for perceived loudness
Normalisation improves consistency; it does not make every listener hear the same volume. Perception changes with headphones, speakers, room noise, hearing preference, phone settings, mono playback and YouTube features enabled on the viewer’s device.
A useful publishing standard is therefore a process rather than a promise:
- Every finished episode is measured with the same integrated-loudness method.
- One house target is recorded and applied consistently.
- True peak and loudness range are checked alongside integrated loudness.
- Stereo channels remain linked unless separate correction is deliberate.
- The final export is tested in stereo and mono.
- Transitions are checked as they will occur in the continuous stream.
- The complete sequence is reviewed before publishing.
The EBU’s −23 LUFS figure remains useful as a clearly labelled broadcast reference. It should not be turned into a claim that YouTube requires it, nor should a different house target be presented as an official platform rule. The important distinction is between a standard you choose for your production and a specification YouTube has actually published.
If you work with several editors, put the target, meter mode, channel handling and export settings in one shared template. Ask editors to include the measured result with each master. This makes a quiet episode easier to investigate because you can tell whether the difference came from the source recording, the mix, the normalisation step or the playback chain.
Keep the original mixed file and the normalised master separately. If you later change the house target, you can return to the finished mix instead of repeatedly processing an already processed export. Repeated gain changes and compression can make speech harsher without solving the underlying mismatch.
A VPS or local computer can deliver the stream, but the choice of delivery method does not decide the right loudness target. If you are weighing those operational choices, the comparison of a VPS and cloud streaming service for a 24/7 lecture channel is relevant to continuity and maintenance, not a substitute for audio mastering.
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
Should I set every podcast episode to −23 LUFS?
−23 LUFS is the EBU R 128 broadcast recommendation for average programme loudness, and Audacity describes it as a reasonable spoken-word broadcast default. It is not an official universal YouTube target for continuous podcast episodes. Choose a house target for your show, document it and apply it consistently.
Does YouTube Stable volume replace loudness normalisation?
No. Stable volume is a viewer-side feature that balances quiet and loud parts when available, and viewers can turn it off. Other viewers may use Voice boost, mono playback or different device processing, so the finished episodes still need consistent creator-side preparation.
Should I normalise the left and right channels separately?
Usually not for an ordinary stereo programme. Keep the channels linked so the same gain is applied to both sides; independent correction can change stereo balance. Separate treatment is appropriate when the channels are genuinely separate recordings and that correction is intentional.
Why do matched episodes still sound different on a phone?
A phone may fold stereo into mono, apply its own speaker processing or be used in a noisy environment. YouTube playback features can also change perceived dynamics. Check a mono fold-down and compare episodes on the same phone or speaker, but do not promise identical loudness for every listener.