Best Audacity export settings for short audio previews on a website?

Feedback and Discussion Forum
1

Hi, I’m working on a [link removed] and I’m preparing short spoken-word audio previews in Audacity. I’d like to keep the files lightweight, clear, and consistent in volume across multiple clips.

At the moment I’m testing noise reduction, compression, and normalization, but I’m not sure what settings work best for short web audio.

My main questions are:

  1. What export format is usually best for short voice previews on a website?

  2. Is it better to normalize peaks or target loudness for this use case?

  3. Are there recommended Audacity settings to reduce background noise without making speech sound distorted?

Thanks.

2

Loudness for consistency, as one rogue loud peak will make your audio too quiet when you normalize peaks.
If any peaks are clipping after loudness-normalization, use (soft) limiter (-1dB) without make-up gain to knock then down.

3
  1. For maximum compatibility among different web browser programs, I would recommend the MP3 format.
  2. For mainly voice audio, I use loudness normalization as the last step before exporting.
  3. Any efforts to minimize background noise should be done during recording, rather than afterwards in editing. As people read your post, you will get plenty of advice for doing that.

I will post more on website HTML markup for audio soon.

4

This is a template for a web page with an audio file.

<!DOCTYPE html>

<html lang="en" xmlns="http://www.w3.org/1999/xhtml">
<head>
    <meta charset="utf-8" />
    <title>Web page tab text goes here</title>
</head>
<body>
    <br />
    <br />
    <video controls style="width: 400px; height: 500px; background-color: black;" poster="bc000.png" preload="metadata">
        <source src="bc000.mp3" type="audio/mpeg" />
        <track default src="bc000.vtt" kind="captions" srclang="en" label="English">
        <ul>
            <li>
                <a href="bc000.mp3">Listen to episode 000 </a>
            </li>
        </ul>
    </video>
    <br />
    <br />
    <ul>
        <li>
            <a href="#bcTranscript">Transcript</a>
        </li>
    </ul>
    <br />
    Text of web page goes here.
    <br />
    <br />
    <iframe id="bcTranscript" src="bc000.vtt" frameborder="1" style="border: 1px solid white; width: 800px; height: 400px;" height="400" width="800" name="bcTranscript" scrolling="auto" title="Transcript"></iframe>
    <br />
    <br />
</body>
</html>

The <video></video> element and the tags between them make up the audio player with controls the user can click on to start and stop playback of the media.

I highly recommend providing a transcript which will play along with the audio in the control. You can use the OpenVINO Whisper AI to generate the transcript inside Audacity. When it is done, it will create a series of labels in a Label track of the Audacity project. The main advantage is that you can listen along to the audio in Audacity, and correct the text in the labels (with the hidden timestamps embedded) all in one place instead of switching between an audio player and a text editor.

Once the text in the labels is correct, extract the labels to a .VTT file. The <video> element of the web page can use the text and timestamps in the VTT file to synchronize the text with the audio as it is playing in the browser program.

The <iframe> element at the bottom of the web page will allow people to read the transcript and search for text to locate a particular section of the audio.

Mix and render the audio tracts of the Audacity project to an exported .MP3 file that the <video> control can play. Create a .PNG file as cover art for the <video> control as a background for the transcript to display over. It is easier to organize things if you give these three files the same filename with their different extensions.

I would also recommend that you use a meta tag editor to add the cover art to the MP3 file as well as title, track, and copyright information.

1 Like
What do you actually -do- with your music?
5

Hi!

Here is the fastest way to prep those clips:

1. Format: Export as MP3 (96-128kbps) in Mono. Switching to Mono cuts the file size in half but keeps voice perfectly clear for the web.

2. Volume: Use Loudness Normalization (-19 LUFS), not Peak. Peak normalization gets thrown off by a single loud sound (like a laugh). LUFS ensures every single clip sounds equally loud to the human ear.

3. Noise Reduction: To avoid that “underwater” robotic voice, use light passes. Grab your noise profile, then set Reduction to 6-10 dB, Sensitivity 6, and Smoothing 3. If it still hisses, run a second light pass instead of one heavy one.

(Side note: If you’re on a Mac and tweaking dozens of clips manually gets tedious, I actually built a local app called AudioClean Pro that automates this noise removal and LUFS leveling in one click. Just thought I’d mention it if you want to speed up your workflow!)