To keep episodes in a 24/7 kids’ stream from jumping noticeably in volume, measure and match each programme’s loudness before playout where you can. OBS compression and limiting can help control changing dynamics and peaks, but neither automatically normalises the perceived volume of complete prerecorded episodes.
The practical approach is to identify whether each mismatch is an overall level shift or a few loud moments, prepare the files if possible, and use OBS as a consistent final control stage. Then record and listen to transitions rather than trusting the mixer meter alone.
Episode loudness and peaks are different problems
An episode can have a safe peak level and still sound much louder than the one before it. Programme loudness describes the overall level over time; a peak meter shows brief high points. A quiet dialogue-led story may have a low average but one sudden musical sting, while a lively cartoon may stay consistently loud without a single extreme transient.
That distinction determines the remedy. If the whole episode is louder or quieter, a loudness measurement and a stable file-level adjustment are more useful than simply placing a limiter on the live output. If only effects or music jump out, dynamic control such as compression may help. A limiter is a last safeguard against peaks crossing a chosen ceiling, not a volume-equalising control for every scene.
OBS describes its mixer meter in terms of peak level, with 0 dBFS as the digital maximum. A bar that reaches a similar height in two episodes does not prove that their dialogue, music, or overall programme will seem equally loud to a child listening on a television or tablet. Use the meter to spot output risk, not as a substitute for a loudness measurement and listening check.
Children’s content often moves between speech, songs, sound effects and quiet pauses. Matching the whole episode by ear alone can miss a short loud burst; matching only peaks can leave one episode’s dialogue much harder to hear than another’s. Keep both questions in view: how loud is the programme overall, and how high do its short peaks go?
Compare the episodes before they reach OBS
Make a simple inventory of the files in the order they will play. Note the duration, format, and any known source differences, such as one series arriving from a different editor or one episode having been remastered. Then listen to the last section of one file and the first section of the next. This is where a change in programme level becomes apparent in context.
Use a loudness meter that reports integrated programme loudness, commonly expressed in LUFS, to compare complete files. Check true peak as well, because a file with a reasonable integrated loudness can still contain a sharp transient. The loudness target should come from the delivery specification for the platform or distributor you actually use. There is no universal LUFS target established here for every YouTube live stream or children’s programme.
The European Broadcasting Union’s R 128 recommendation sets programme loudness at −23.0 LUFS in its broadcast context, with a ±1.0 LU allowance where that target is not practically achievable, such as live programmes. Treat this as standards context, not as a YouTube requirement or a number OBS imposes. If you are deciding on a delivery target, check the current requirements for your distribution path rather than adopting a broadcast specification by default. See the EBU R 128 recommendation for its scope and measurement context.
For a small library, you can begin with a listening pass and a spreadsheet of the loudness readings. The purpose is not to create paperwork: it is to distinguish a consistent offset across one episode from isolated loud scenes. If episodes are numerous, a repeatable batch workflow reduces the chance of correcting one file differently from another. Keep an untouched copy of each original so you can revisit a processing choice if it sounds unnatural.
A peak-only reading is still useful. OBS warns that audio above 0 dBFS can clip, and its meter cannot establish equal perceived loudness. Look for clipping, inter-sample or true-peak concerns in a suitable analyser, and listen on the kinds of speakers your audience is likely to use. A narrow phone speaker can expose sharp effects that sound less obvious on studio headphones.
Prepare files for a consistent hand-off
If you control the episode files, correct stable overall loudness differences before they enter OBS. Choose an appropriate delivery target, measure each episode using the same method, and adjust the files consistently. Include true-peak control in the preparation workflow, then listen for pumping, distortion, or dialogue that has become tiring. Normalisation changes levels; it does not repair poor mixes or restore detail lost to clipping in the original.
A batch-capable audio or video workflow can be helpful when a library is large, but the important features are measurement, repeatability and a way to review the output. Do not select a tool just because it offers a button labelled “normalise”: confirm whether it measures programme loudness or merely scales a peak, and whether it can preserve the audio format your playout process expects. No particular paid processor or target is required by this workflow.
If an episode cannot be changed, a per-source gain adjustment in OBS may correct a stable offset. Keep a note of the source and adjustment so it can be reproduced if the playlist is rebuilt. This is different from adding a compressor to the final mix: one compressor responds to the combined signal as it plays and does not know where one file ends or what integrated loudness the next file should have.
When the content is played as a continuous file sequence, confirm that the gain is applied to the intended source and that transitions do not create a brief jump. For a larger workflow, the cartoon-stream guide to systemd and FFmpeg covers a different way to keep a prerecorded sequence running; whichever playout method you choose, prepare and compare its audio assets in advance.
Use OBS compression and limiting for what they do
OBS’s compressor reduces signal above a threshold. The threshold and ratio affect how much reduction occurs; attack and release govern how quickly it responds and recovers. Output gain can restore level after compression, but raising it also raises quieter material and any remaining peaks. OBS says compression is useful when a source sometimes spikes much louder and is typically placed at or near the beginning of a filter chain. Read the OBS compressor guide for the filter’s controls and roles.
There is no single safe preset for a playlist of different children’s shows. A setting that smooths a sudden effect in one programme could flatten a song or make dialogue sound unnatural in another. Start with modest processing, make a test recording, and compare the processed result with the original at a similar listening level. If the sound becomes dull, breathy or “pumping”, reduce the amount of compression or address the file itself instead.
A limiter constrains peaks at its threshold. Place it last in the relevant filter chain, as OBS recommends, so later filters do not undo the ceiling. Set a ceiling that leaves suitable headroom for the output path and verify the encoded stream; the limiter is not permission to drive every passage into the threshold. The OBS limiter guide explains its threshold and chain placement.
Think of the two filters as a boundary around dynamic behaviour, not a loudness target. Compression can reduce the difference between a loud transient and nearby material. Limiting can prevent an excessive peak from passing its set ceiling. Neither scans the next episode, calculates its programme loudness, and adjusts it to match the previous one. Keep file preparation and per-episode gain correction separate from live peak control.
Set a repeatable chain and test transitions
A practical OBS chain begins with consistent source routing and gain, followed by compression where it is useful, then a limiter as the final filter in that chain. The exact chain depends on where the audio enters OBS and whether the filter is on an individual media source or a shared output. Avoid applying the same strong processing at multiple stages without checking the combined effect.
Set a baseline using representative content, not silence or a single loud jingle. Record an ordinary section, a busy section with music and effects, and a quieter dialogue section. OBS’s mixer guidance recommends recording a normal session and listening before going live; see the OBS audio mixer guide. Play the test back, because watching meter bars while you adjust is not the same as hearing whether words remain clear.
Then test the actual boundary between episodes. A useful test recording includes the tail of the first episode, any gap, and the opening of the next. Check whether the opening feels like a jump, whether the outgoing audio is cut off, and whether a transition effect introduces a new peak. If your playlist uses fades, test the exact fade duration and overlapping audio: overlapping music and dialogue can sum to a louder moment even when each file was prepared sensibly.
YouTube’s live encoder guidance recommends 128 kbps for stereo audio. That is an audio bitrate recommendation, not a loudness target. Its guidance also recommends testing with audio and movement similar to the real event. Check the current YouTube Live encoder settings before configuring the broadcast, and keep the distinction between transmission settings and programme loudness clear.
If your current setup is causing overnight computer or playout interruptions in addition to audio variation, stabilise the operating method separately from loudness work. The guide to running a YouTube stream from a Synology NAS discusses an alternative operating arrangement; it does not replace the need to prepare and test episode audio.
Monitor clipping and uneven levels during playout
During a live run, watch OBS’s audio meters for unexpected peaks and listen periodically to the output. A meter reaching the top or audio sounding distorted is a prompt to investigate the source, gain, or processing. A quiet-looking meter is not evidence by itself that the audience cannot hear dialogue; check actual playback on representative devices and at a sensible listening level.
If only one source is consistently too loud, adjust that source rather than changing the entire mix. If all sources are peaking together, review the output gain and the final limiter threshold. Avoid chasing every meter movement with the fader while live: note the episode and timestamp, then make a controlled change during a test or between runs. Keep a record of filter settings so a successful baseline is not lost.
Monitor the encoded stream where possible as well as the local OBS mix. Encoding and platform playback can alter what you hear, and a local preview does not show whether the outgoing broadcast has the same balance. Use a private or unlisted test when suitable, and confirm that the audio reaching the destination is not muted, delayed, or significantly different from the recording.
A 24/7 channel also needs an archive plan independent of the live mix. YouTube Help says an automatically archived live stream longer than 12 hours may not be captured at all, so check current destination behaviour and maintain an independent recording if you need a continuous copy. The cloud recording storage guide can help you think through where to keep those recordings. Do not assume the platform archive is a complete backup.
Recheck the sequence as a whole
After adjusting files or filters, listen through a representative sequence rather than sampling only one episode. Include material from different programmes, the transitions between them, and a section with the loudest expected music or effects. Check the beginning and end of the day’s sequence too if those points involve a different intro, ident, or schedule change.
Use a simple review sheet: episode name, integrated loudness reading, true-peak reading, gain adjustment if any, and a note about the transition before and after it. The numbers help you find outliers; the notes from listening help you decide whether the numbers matter. If one episode is measurably different but sounds natural in context, avoid processing it solely to make the spreadsheet uniform.
Recheck after replacing or editing an asset. A new version may have different dialogue levels, a different duration, or altered music. Likewise, if you change the OBS filter order, output gain, or playlist behaviour, make another recording of transitions. In a continuous channel, a small change can affect every episode, so do not rely on a test made before that change.
If the library is already running and the mismatch is obvious, use the least disruptive correction first. A stable per-source gain change can address a known file offset; a carefully adjusted compressor may tame occasional dynamics; and the limiter can catch peaks. Schedule a fuller file-level measurement pass when practical. StreamNeo can remove the need to leave a computer running for file-based continuous playout: upload the video, provide the YouTube stream key, and it runs the broadcast with monitoring and automatic restart if it drops. It does not replace the work of matching the audio assets themselves.
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
Can OBS normalise every prerecorded episode automatically?
Do not rely on OBS’s compressor to do this. It responds to signal level above its threshold while audio plays; it is not documented as measuring each complete episode’s integrated loudness and setting a matching level. Prepare files or apply a deliberate source gain adjustment for stable offsets.
Should I use LUFS or the OBS meter?
Use integrated loudness measurement, often expressed in LUFS, to compare whole programmes, and use peak or true-peak checks to identify peak risk. OBS’s meter is useful for monitoring signal peaks but cannot show that two episodes will sound equally loud overall. Confirm the target with the destination’s current delivery guidance.
Does a limiter make quiet episodes louder?
No. A limiter constrains peaks at its threshold; it does not lift quiet passages to a common programme loudness. Raising the output gain may make the whole signal louder, but it can also push peaks into the limiter and change the sound.
Is −23 LUFS the right target for a YouTube kids’ stream?
Not automatically. −23.0 LUFS is the EBU R 128 broadcast recommendation in its context, not a universal OBS or YouTube Live requirement. Check the current specification for your distribution path and use that, alongside true-peak checks and listening tests.