Configure Gain, Dynamics, Delay, Noise Suppression, and Metering

Process and measure a Signal path without hiding clipping, dropouts, or excessive latency.

Written By 4ALL.LIVE

Last updated About 24 hours ago

Build only the processing chain the program needs, preserve headroom, and use meters to expose clipping, dropouts, and loudness changes instead of hiding them.

Best for: audio operators preparing a controlled Signal path between a verified source and production output.

Choose applicable processing

  • Gain: set a usable nominal level without driving downstream stages into clipping.
  • Gate, compressor, and limiter: manage unwanted low-level content, dynamic range, and final peaks when appropriate to the source.
  • Delay: add only the alignment delay required by the production design and record the chosen value.
  • EQ and reverb: apply deliberate tonal or spatial changes and compare them against a bypassed reference.
  • Noise suppression: use only after listening for speech damage or artifacts.
  • Level Meter and Loudness Meter: observe the path before and after processing.

Build the chain

  1. Stop the pipeline and place a meter after the source.
  2. Add the approved processing nodes in signal-flow order.
  3. Add another meter before the final output.
  4. Start off air with conservative settings.
  5. Adjust one stage at a time while watching and listening to the source and processed meters.
  6. Stop, save the verified configuration, relaunch Signal, and repeat the test.

Read the meters

  • Peak shows short signal maxima; keep sufficient headroom below clipping.
  • RMS shows a longer-term signal level that helps compare average energy.
  • Clip hold preserves evidence that a clip occurred even if the instant has passed.
  • LUFS describes perceived program loudness over the meter's presented interval.
  • True peak helps identify peaks that may exceed a safe reconstructed output level.
  • Loudness range describes program variation; it is not a target by itself.

Headroom is the available level between normal operation and clipping. Preserve it across the source, effects, Signal output, interface, and downstream input.

Important: DeepFilterNet processes when the graph output rate is 48 kHz. At other graph output rates it passes audio through. Verify the actual rate before assuming suppression is active.

Verify the processed path

  • Source and output meters respond continuously without unexplained gaps.
  • No clip hold is triggered during the loudest rehearsal material.
  • Bypass comparisons confirm that each effect improves the intended issue.
  • The selected delay is documented and does not create excessive monitoring latency.
  • Speech remains intelligible and free from obvious pumping, gating, or suppression artifacts.

Troubleshoot

  • Clipping after an effect: reduce upstream gain or the effect output rather than only lowering the final hardware volume.
  • Speech cut off: relax the gate or suppression setting and compare against bypass.
  • Unexpected dry signal with DeepFilterNet: confirm a 48 kHz graph output rate.
  • Excessive latency: bypass nonessential delay and processing nodes one at a time while off air.
  • Meter gaps: check the source, TTS status, underruns, and graph edge before changing dynamics.

Operational boundary

Meter values describe points inside Signal. They do not include or guarantee upstream LL-TTS generation quality, transmitter performance, RF coverage, receiver level, or end-to-end latency.