Summary
Highlights
The video introduces the critical role of dialogue in video production and explains the importance of mixing or processing dialogue. It highlights Adobe Audition's single track editor and spectral frequency display as powerful tools for this task.
The first step in dialogue mixing is noise reduction. The tutorial demonstrates how to capture a noise print by selecting a quiet section of the audio. After capturing, the noise reduction process is applied, with a focus on adjusting the noise reduction percentage to avoid removing too much of the actual dialogue. Noise reduction should always be the first step in the signal chain.
After noise reduction, a single-band compressor is recommended. The video explains how to set the threshold based on peak levels (around -12 dB for dialogue) and a gentle ratio (e.g., 4:1) to reduce peaks without making the audio sound too intense. Attack and release settings can usually be left at default, and output gain is adjusted later with a limiter.
A de-esser, which is a compressor specifically for sibilant frequencies (s's and ch's), is the next effect. The tutorial shows how to identify these harsh frequencies (typically 5 kHz-8 kHz for men, 7 kHz-10 kHz for women) using the spectral display and 'output sibilance only' feature, then set the threshold to subtly reduce them.
A parametric EQ is used to clean up the sound further. The video demonstrates how to use the analyzer to identify problematic low-mid frequencies (around 200-300 Hz) and pull them down before boosting the top-end for clarity. A wide Q is used for broader adjustments, and a low-pass filter can remove rumble.
Finally, a hard limiter is applied to boost the overall signal without clipping. The input boost is adjusted to bring the peaks up to a safe level (e.g., -3 dB), and the maximum amplitude is set to -1 dB. This ensures a clean, boosted signal that won't clip when mixed with other audio elements in Premiere Pro.
The tutorial concludes by comparing the original, unprocessed audio with the mixed version. The processed dialogue shows reduced noise, smoother amplitude due to compression and limiting, and a cleaner overall sound from the EQ, preparing it for integration into a larger video project.