A podcast clip can sound distracting for several reasons: fan noise under the voice, a sudden bump, an uneven speaking level or a reflective room. Start by naming the problem. A noise reducer can be useful for a fairly steady background, but it is not an editor, loudness meter or mastering chain.
Does the clip need denoising, editing or a level adjustment?
Start with the clip in the context of the episode. If a steady hiss competes with an otherwise clear sentence, try a reduction preview. If a single cough or table knock interrupts it, editing that moment may be more useful. A guest who is simply quieter than the host needs level adjustment, a separate job.
A hollow voice may come from distance or room reflections; a clipped shout has distorted peaks. Neither is reliably fixed by reducing a steady noise bed.
Should you work from separate speaker tracks?
When the recording has separate tracks, keep the host and guest apart while making decisions. A guest's fan does not justify processing the host's already clean track. Save a copy of each source and note which version belongs to which speaker so that you can return to the original later.
With a mixed track, listen especially where voices overlap. A general denoiser cannot promise to separate them. An alternate take or a careful edit might be more useful than a stronger global effect.
How do you compare a cleaned voice clip?
- Use an editor to export a representative short passage as a supported audio file. Keep the original session and source files.
- Check the file against both audio limits: 25 MiB and five minutes. Supported types are WAV, MP3 and AAC in M4A; ALAC is not supported.
- Open the audio tool, select the file and create a preview of up to the first 30 seconds.
- Listen to the same words in Original and Processed at the same volume. If a longer supported clip passes that check, process the full audio and download the WAV.
Pay attention to quiet consonants and breath as well as the background. A voice that loses its character is not necessarily an improvement. The robotic voice guide covers common artifacts to listen for.
What still needs to happen before publishing an episode?
Import the downloaded WAV into your podcast editor on a new track. Align its beginning with the source selection, temporarily mute the original section and listen through both joins. If the new clip sounds abrupt, adjust the edit in your editor or keep the original; do not delete the source track until you have checked the full sequence. Save a separate project version so another edit remains possible.
Listen again with the host, music and transitions in place. You may still need to adjust levels and check the final export. This browser tool does not measure podcast loudness or master an episode.
The listening examples use documented speech and synthetic noise. They are useful practice for comparison, but they are not tests of your guest's microphone or recording environment.
When should you use a full audio editor?
Use one when you need to work across a long episode, adjust timing, remove isolated sounds, balance speakers or change the strength of a process. The browser preview has no adjustable reduction-strength control. If it makes the voice worse, keep the source or use a tool that offers the control you need.
Podcast cleanup questions
Can I upload an hour-long episode?
No. The audio input must meet both the five-minute and 25 MiB limits.
Does this remove music under speech?
It is not a music-removal or source-separation tool.
Is the downloaded WAV ready to publish?
It is a processed audio file. Check your complete edit, levels and export requirements separately.