Download Audacity：https://www.audacityteam.org/
Download my Preset:
The tool of removing ambient noise：https://www.lalal.ai/?fp_ref=samuel76
I watched a YouTube video today and noticed the audio was really quiet, especially when watching outdoors—it was almost inaudible. It was this one:
I tweaked her audio a bit, and it turned out like this.
So, I thought, maybe a lot of you have similar needs for audio enhancement but don’t want it to be too complicated. After all, making videos is already time-consuming enough. So, after some experimenting, I came up with a super simple and beginner-friendly solution. You don’t need any audio tech background—just follow my steps, and you’ll definitely improve the sound quality of your content. Plus, all the tools are free software, and I’ll include all the tools and links on this episode’s webpage, with the URL in the description or comments.
Before we start, please hit the little bell and subscribe to my channel.
Let’s first get a quick overview of what’s going on. When you speak into a microphone, your vocal cords vibrate, causing various parts of your body to resonate, producing a continuous series of sound waves. We’ll just call this “speech” from now on. Before speech reaches the mic, it can include low-frequency sound waves starting at 80 Hz all the way up to high-frequency waves in the thousands of Hz. The energy of these sound waves varies a lot too—sometimes it’s a soft whisper, just a fraction of a decibel, and other times it’s loud, reaching 60 or 70 decibels. Quick sidenote: Hz and dB are physical units for measuring sound. Hz is the unit for frequency, which, in simple terms, affects the tone of the sound. Words like bright, deep, low, or sharp describe characteristics tied to frequency. dB measures sound pressure level, which you can think of as volume.
In short, the speech coming out of your mouth has a huge range of frequencies and volumes, also called dynamic range. Even the best microphones have limits to their sensitivity, especially in dynamic range. A mic’s sensitivity is far lower than the dynamic range your speech can produce. That’s why you see some professional singers constantly adjusting the distance between their mouth and the mic while performing.
When we record a show, it’s hard to keep adjusting the mic position like a singer. Usually, we fix the mic at a suitable distance from the mouth and lower the mic’s input level to prevent loud speech from overloading the mic or audio equipment’s dynamic limit. If the input exceeds the equipment’s dynamic limit, you get a squeaky, distorted sound—sometimes called clipping, or in layman’s terms, “audio popping.”
So, obviously, to avoid popping during recording, we lower the input overall. But then, when the audio is recorded into a computer or phone, you’ll notice the overall volume is pretty low. Some parts might be loud, but others are super quiet. No matter how much you crank up your computer or phone’s volume, the sound won’t be as rich and clear as a radio host’s.
Let me use that show I mentioned earlier as an example to show you how to achieve the final effect.
Here, I’m using a free audio software called Audacity. It’s available for both Windows and Mac, super powerful, and the download and installation are really straightforward, so I’ll skip that part. Once installed, click here to download the preset files I prepared for you. On this episode’s webpage, find “Parameter Preset Download” and download the zip file. After downloading, unzip it to get the preset files. Then open Audacity, go to the File menu, select Open, and load your audio file.
Next, we’re going to do six things:
**First thing: Filtering**
The goal here is to make the audio material clean and clear. It’s like using a juice filter—after filtering, you remove unwanted pulp and impurities, leaving just pure juice. During recording, signal amplification, and analog-to-digital conversion, the audio signal picks up some impurities, especially when recording in noisy environments. Filtering mainly reduces signals at certain frequencies.
Based on my experience, for spoken audio, signals below 100 Hz or above 4,000 Hz are generally irrelevant to speech clarity. Reducing or completely removing these signals won’t affect how clear the speech sounds.
First, select the audio segment you want to process. Click the Effect menu, go to EQ and Filters, find Filter Curve EQ, then click the Preset & Settings button and choose Import to load the preset file. From the preset files you downloaded, find “Vocal-Filter” and open it. You’ll see I completely filtered out signals below 100 Hz—this is called a low-cut. High frequencies between 4,000 and 8,000 Hz are slightly reduced. This range often includes a lot of environmental noise, sibilance, and mouth noises, but reducing it too much can dull the speech’s brightness, so I only apply a light reduction. Signals above 10,000 Hz get a high-cut, as this range includes some breathy sounds. If your show relies on a lot of breathy, whispery effects, you might want to manually boost this range, but for most speech-based content, a high-cut works fine. Click Preview to hear the processed effect. If you’re happy with it, click Apply to process the audio. Note: once you hit Apply, you can only undo it via the Edit menu’s Undo option. You can also apply the filter multiple times to the same audio for a more thorough effect.
Now, the audio no longer has that background hum or popping mic sounds, and mouth noises like sibilance are reduced.
**Second thing: Dynamic Compression**
Dynamic refers to the range between the quietest and loudest parts of the audio. Compressing the dynamics means narrowing the gap between the loudest and quietest volumes. The main goal is to make the audio’s loudness more consistent, so it doesn’t feel like it’s jumping from loud to soft or far to near.
How do we do this? One way is to select the audio segment you want to adjust, go to the Effect menu, choose Volume and Compression, open Amplify, and enter a negative value to reduce volume or a positive value to increase it, then hit Apply. This method is flexible and lets you control every detail, but if the audio is long with lots of variations, it can be time-consuming.
In Audacity, we can use the Compressor for this. Select the audio to process, go to the Effect menu, choose Volume and Compression, find Compressor, and open it. Click the Presets & Settings button, then Import, and from the downloaded preset files, find “VocalStartUp,” open it, and click Apply.
After processing, the audio is noticeably louder, and details are slightly more pronounced. If some parts are still too loud, you can use the first method to adjust them individually.
When using the Compressor, besides reducing the dynamic range, you’ll also get some enhanced details and a slight change in sound quality. Applying the Compressor multiple times might produce unexpected effects, so listen carefully to the changes before deciding whether to keep the multi-processed result.
**Third thing: Distortion**
You might have heard of electric guitar distortion effects. Here, we’re using a distortion effect on the voice—not to mimic a guitar, but to add harmonic components with a subtle distortion effect, enhancing the audio’s brightness and fullness.
In Audacity, go to the Effect menu, find Distortion and Modulation, select Distortion, and open it. Click Presets & Settings, then Import. From the downloaded preset files, find “Voice Exciton” and open it. The parameters are: Distortion Type set to Hard Clipping, Clipping Level set between -6 and -10, Drive can be adjusted but defaults to 50, and Make-up Gain should be set to 0 to avoid extra volume boost. Then click Apply.
Listen to it now. The audio sounds quite different.
**Fifth thing: Loudness Normalization**
Many platforms and industries have their own loudness standards. For YouTube, the recommended loudness is no lower than -14 dB LUFS. I won’t dive into what that means here—just follow the steps.
Select the audio to process, go to the Effect menu, choose Volume and Compression, find Loudness Normalization, and open it. Click Presets & Settings, then Import. From the downloaded preset files, find “YouTube,” open it, and click Apply.
**Sixth thing: Limiting**
This is a standard safety process for audio. We’ll use Audacity’s Limiter. Select the audio, go to the Effect menu, choose Volume and Compression, find Limiter, open it, set the Threshold (dB) to -1.0, and click Apply.
Go to the File menu, choose Export Audio, select Export to Computer, and set the audio file format as recommended here.
Then click Export.
At this point, you’ve completed all the speech processing. Doesn’t it sound much louder and clearer than before?
Although the overall audio is louder, the background noise might get amplified too. What to do? Thanks to AI advancements, I recommend a super practical online tool for removing ambient noise called LALAL.AI.
Open this website, click here to switch to Chinese, then click Log In. For first-time users, register with your email or another login method. After registering and logging in, go to Products, select Voice Remover, drag your audio file into the page, wait a moment, and it’s done.
How’s that? Doesn’t it feel like something from a big professional media outlet?
Alright, that’s it for today’s video. All the tools and links are on this episode’s webpage. If you run into any issues while processing your audio, feel free to ask in the comments, and I’ll reply as soon as I can. If you found this helpful, don’t forget to subscribe, like, and save. Thanks for watching!
—
This translation maintains the original structure and intent, using conversational English suitable for video narration. Let me know if you need further adjustments!