What Are DSP Effects?
DSP (Digital Signal Processing) effects are audio transformations applied to your voice in real-time. Unlike AI voice conversion which changes your voice identity, DSP effects shape the tone, dynamics, and spatial characteristics of your audio. Think of DSP as the audio equivalent of photo filters — they enhance and refine the raw signal. In Echo, DSP effects are applied after AI voice conversion, allowing you to polish the output to studio quality.
Noise Gate — Eliminate Background Noise
A noise gate is the first effect in the chain. It works like an automatic mute switch: when your voice is below a set threshold (typically -40dB to -30dB), the gate closes and no audio passes through. When you speak, the gate opens. This eliminates keyboard clicks, fan noise, room hum, and any ambient sound that would otherwise be transmitted. Set the threshold just above your room's noise floor for best results. If set too high, the gate will cut off the beginnings of quiet words.
Equalizer (EQ) — Shape Your Tone
An equalizer lets you boost or cut specific frequency ranges. For voice, the key ranges are: 80-250 Hz (warmth and body), 250-500 Hz (muddiness — often needs cutting), 1-4 kHz (presence and clarity), and 4-8 kHz (air and sibilance). A common voice EQ recipe: cut 2-3dB around 300 Hz to reduce muddiness, boost 2dB around 3 kHz for clarity, and apply a gentle high-pass filter at 80 Hz to remove rumble. Echo includes a 10-band graphic EQ with presets for common voice types.
Compressor — Consistent Volume
A compressor reduces the dynamic range of your voice, making quiet parts louder and loud parts quieter. This is essential for gaming and streaming where you might whisper tactical callouts one moment and shout the next. Key parameters: threshold (the level above which compression kicks in), ratio (how much to reduce the signal), attack (how fast compression engages), and release (how fast it stops). For voice, start with: threshold -20dB, ratio 3:1, attack 10ms, release 100ms.
Reverb — Add Space and Depth
Reverb simulates the acoustic reflections of a physical space. A small amount of reverb makes voices sound more natural and less "dry." Too much reverb makes speech unintelligible. For gaming and streaming, use a short room reverb with a decay time under 0.5 seconds and mix below 15%. For creative effects (announcer voice, cave echo, cathedral), increase decay time and mix level. Echo includes preset reverbs for Room, Hall, Plate, and Cathedral.
Effect Chain Order
The order of effects matters. The standard chain is: Input → Noise Gate → EQ → Compressor → AI Voice Conversion → Reverb → Output. The noise gate goes first to prevent noise from being processed by later effects. EQ shapes the tone before compression. The compressor evens out dynamics. AI conversion changes the voice identity. Reverb goes last to add natural space to the final output.
- ●1. Noise Gate — clean the input
- ●2. High-Pass Filter — remove rumble below 80 Hz
- ●3. Equalizer — shape tone and clarity
- ●4. Compressor — even out dynamics
- ●5. AI Voice Conversion — change voice identity
- ●6. De-esser — tame harsh sibilance (optional)
- ●7. Reverb — add spatial depth
- ●8. Limiter — prevent clipping