OK. To make a “silent subliminal” like the one in your first post:
Start with a good clean vocal recording. The one in your sample sounded if it was created with a “text to speech” application. If you produce this recording in another application (rather than recording it yourself), ensure the file is saved as “WAV, 16-bit 44.1 kHz sample rate”, then import the file into Audacity. If you record the speech in Audacity, ensure that you record at 44100 Hz sample rate (the default for Audacity).
With the track selected, Apply the “Normalize” effect, and normalize to -1 dB (default setting).
With the Nyquist Prompt, apply the following code: