Video summary

Audio Post Production: Dialogue and Voice EQ

Main summary

Key takeaways

Technology

Summary: Audio post-production — Dialogue/Voice EQ (live stream + tutorial concepts)

This video is a structured guide (with live examples) on EQ’ing spoken dialogue/voice for content such as movies, podcasts, streaming (OBS), and TV. It also notes that many of the same concepts apply to singing/rapping.

Key themes include:

  • Systematic ear training
  • Corrective EQ
  • Avoiding limitations/overuse
  • Using complementary EQ so dialogue remains clear inside a full mix

Core idea: EQ can only change what’s already present in the recording—it can’t create clarity that isn’t there.

Course / tutorial outline (what’s covered)

1. Overview of EQ (equalization)

  • Why EQ is needed for speech
    • Boost/add clarity
    • Fix missing bass/treble
    • Reduce midrange buildup
    • Remove problem frequencies
    • Tame sharpness or excess “base” (low-end emphasis)
  • Beginner warning: too much low-end can flatten perception:
    • “If everything has base, nothing feels like base.”
  • EQ for noise/rumble reduction
    • Including effects like “phone” audio using high/low-pass and notches
  • Limitation principle
    • EQ cannot create clarity that isn’t present in the source (mic/recording quality and capture matter).

2. Corrective EQ in OBS (live)

  • Demonstrates adding EQ to an audio source in OBS using:
    • Filters → Add → VST2.x plugin
    • Example: FabFilter Pro-Q3
  • Emphasizes live listening and adjustment to reduce rumble and other problematic bands.

3. Overusing EQ + the importance of monitoring

  • Explains why heavy EQ corrections can sound unnatural:
    • Phase/comb-filter and related artifacts
  • Guidance for proper evaluation:
    • Monitor using flat headphones/speakers
    • Test rooms with tools like Room EQ Wizard
    • Turn EQ off when judging issues to avoid “mixing through” processing
  • Mentions acoustic treatment concepts:
    • Bass traps, absorption, diffusion

4. Complementary EQ in the mix

  • When music/game SFX are present, EQ’ing dialogue alone isn’t enough.
  • Recommended approach:
    • Reserve space for voice by reducing competing music frequencies in the ~200–800 Hz region.
  • Automation notes:
    • Possible on a submaster, track, or bus
  • “Confusion points”:
    • Often in the midrange; sometimes also near ~100 Hz in busy scenes.

Key technological / product details and workflow

OBS EQ workflow

  • Add a 3-band EQ or (preferably) a VST2.x EQ plugin such as:
    • FabFilter Pro-Q3
  • Apply it to the audio source (example: DJI Mic 2).
  • Uses OBS with multiple mics at once for A/B comparison between raw capture and EQ needs.

Ear training + frequency “adjectives” chart

  • References a chart (from “DIY Audio Heaven”) that maps descriptors to frequency behavior:
    • muddy, warm, honky, nasal, bright, sibilant
  • Encourages learning by listening rather than staring at the graph first.
  • Breaks ranges into rough areas:
    • ~50–100 Hz: “muddy”
    • ~300 Hz: “puppy”
    • ~1 kHz: “nasal”
    • ~3–6 kHz: “sibilant”
    • ~5–10 kHz: “piercing” region
    • >10 kHz: “air”

Recommended free EQ plugin

  • Mentions Tokyo Dawn Free EQ (referred to as “NVR” in subtitles, but contextual meaning is Tokyo Dawn Free EQ).

Pro-Q3 / FabFilter usage (visual EQ)

  • Recommends using visual EQ to spot resonance/peaks.
    • Example reference: resonance around ~400 Hz
    • Rumble typically below ~50 Hz

Corrective EQ examples (what they actually change)

Rumble removal

  • Start with a high-pass filter to remove sub-bass/rumble
    • Example: car starting noise
  • Often follow with a low-pass filter
    • Helps prevent unwanted high-frequency clutter.

Notching / cutting problem bands

  • Example:
    • Reduce a band near ~450 Hz
  • Notes how Q width changes the character:
    • Narrow vs. broad cuts
  • Practical guidance:
    • Prefer smaller corrective dips (“carving wood”) rather than extreme cuts (“chainsaw”).

Sibilance

  • Example approach:
    • A small dip around ~4 kHz.

Handling reverb (question answered)

  • Suggests EQ’ing reverbs professionally by high-passing around ~200 Hz
    • Goal: reduce muddiness.

Microphone selection/product comparisons emphasized

The demo uses multiple mics recorded simultaneously, comparing both raw tone and how much EQ is required.

Sennheiser MKH60

  • Described as very good / best
  • Requires the least EQ
  • Flatter response, minimal correction needed

RØDE / “Road NT1” (NT1)

  • Described as flatter and easier for EQ
  • Still shows noise/resonance depending on placement

Shure SM7B

  • Common streamer “dark mic”
  • Often needs more high-frequency EQ for clarity
  • Benefits from close placement and sometimes boosters (Fathead/Cloud Lifter style) due to dynamics/high gain needs
  • Notes a possible sharp notch region that may not be audible if it sounds good.

Lavalier mic (DDW Lav Pro on Sennheiser G3)

  • Often described as omnidirectional
    • Picks up more room echo
  • Shows more “carrier”/notch behavior and resonance spikes
    • EQ tends to be more careful:
      • Often subtractive plus low/high-pass
  • Resonance spikes may come from:
    • Room reflections
    • Monitor/table echoes due to omni pickup

Placement insight (mic angle/off-axis)

  • If the mic clips, has noise, or sounds odd:
    • Try changing mic angle/off-axis positioning (even by a few degrees/sideways)
    • Prefer this over relying on aggressive EQ.

Limitations / analysis: what EQ can’t fix and why it fails

EQ limitations

  • Can’t make a voice sound like a completely different vocal profile (e.g., “James Earl Jones” vs. “Morgan Freeman” style transformations).
  • If capture/placement is wrong, EQ won’t fully rescue clarity.

Overusing EQ

  • Heavy band-by-band corrections can create comb filtering/phase issues due to phase delay at EQ crossover points.
  • Guideline:
    • Use EQ as minimum necessary
    • Avoid template thinking like:
      • “Don’t automatically cut every track by 3 dB at 500 Hz.”

Monitoring is critical

  • Flat monitoring is necessary; otherwise room/speaker resonance causes bad decisions.
  • Tools and treatment:
    • Room EQ Wizard
    • Acoustic treatment (bass traps/absorption/diffusion)

Don’t judge while monitoring through effects

  • Advice from experience:
    • Turn EQ off when evaluating so taste doesn’t get biased by processing.

Complementary EQ (dialogue competing with music/SFX)

Core principle

  • Dialogue clarity often lives in the low-mid band (~200–800 Hz) when other audio sources are present.
  • Instead of simply increasing voice volume, the recommendation is to:
    • Gently reduce ~200–800 Hz in the music/SFX so the voice reads clearly.

Automation workflow

  • Demonstrates complementary EQ using a stock DAW EQ example (e.g., Avid/Pro Tools stock EQ):
    • Reduced Q (example: Q ~0.4)
    • Starting cuts around -3 to -4 dB where confusion tends to be greatest
  • Automation can be done on:
    • Submaster / track / bus

Strategy comparison (ducking vs complementary EQ)

  • Volume ducking (e.g., auto-ducker in OBS) is one method.
  • Complementary EQ is another way to reserve space for speech.

Main speakers / sources

Primary speaker

  • Andi (host; referenced directly multiple times)

Software/tools referenced

  • FabFilter Pro-Q3
  • Tokyo Dawn Free EQ
  • OBS (filters + VST2.x plugin workflow)
  • Pro Tools (stock EQ + complementary mix discussion)
  • Room EQ Wizard

External reference mentioned

  • DPA website voice-processing guidance (including boosting 1–5 kHz for clarity)

Original video