Sonible smart:chain: podcast mixing tutorial and comparison

This walkthrough uses Sonible smart:chain to mix a two-person podcast in Hindenburg. The speaker selects a speech source for each voice and plays audio to train the EQ and compressor. He treats the resulting settings as a starting point, then adjusts the EQ below 100 Hz and adds a high-pass filter for rumble. The input rider levels the signal entering the compressor, while separate controls bias the EQ toward brightness or warmth and compression toward a dynamic or dense sound.

The gate demonstration compares context-aware speech detection with traditional level-dependent gating. Training produces threshold and impact values of 50, but further adjustments do not remove the breaths the speaker wants to suppress. He reports better results on quieter breaths with the dedicated smart:gate plugin. For de-essing, he auditions balanced and soft modes, preferring balanced because stronger soft processing changes the high end to his ears.

The tutorial also covers oversampling and latency settings. Low-latency mode disables the gate and de-esser. The speaker considers the channel strip useful for faster dialogue mixing, but says its simplified modules do not fully replace the dedicated plugins. In particular, he prefers smart:deess for client work because it also handles plosives. He skips saturation and autoleveling in this dialogue workflow.