audio-production（音频制作）

Audio Production

Comprehensive guide for professional audio production across music, podcasts, video, and interactive media.

Audio Fundamentals

The Audio Signal Chain

SOURCE → CAPTURE → PROCESS → OUTPUT

Detailed:
┌─────────┐   ┌─────────┐   ┌─────────┐   ┌─────────┐
│ Source  │→→→│   Mic   │→→→│ Preamp  │→→→│Interface│
│(voice,  │   │         │   │         │   │  (A/D)  │
│ instr.) │   │         │   │         │   │         │
└─────────┘   └─────────┘   └─────────┘   └────┬────┘
                                               │
                                               ▼
┌─────────┐   ┌─────────┐   ┌─────────┐   ┌─────────┐
│Speakers │←←←│  (D/A)  │←←←│   Mix   │←←←│   DAW   │
│Headphone│   │Interface│   │  Master │   │Recording│
└─────────┘   └─────────┘   └─────────┘   └─────────┘

Key Audio Concepts

Frequency Spectrum

20 Hz                                              20,000 Hz
├───────────────────────────────────────────────────────┤

SUB-BASS │  BASS  │  LOW-MID │  MID  │ HIGH-MID │ HIGH
20-60    │ 60-250 │ 250-500  │500-2k │  2k-6k   │ 6k-20k
         │        │          │       │          │
Rumble   │ Warmth │  Body    │ Pres- │ Presence │ Air
Feel     │ Punch  │  Muddiness│ ence  │ Harshness│ Sibilance

Decibel Reference Points

dBFS (Full Scale):
 0 dBFS  ████████████  Digital maximum (clipping)
-3 dBFS  ██████████    Peaks (leave headroom)
-6 dBFS  █████████     Recommended peak ceiling
-12 dBFS ███████       Average loud signal
-18 dBFS ██████        Recommended recording level
-24 dBFS ████          Conservative level
-60 dBFS █             Noise floor target

Recording

Microphone Types

Polar Patterns

CARDIOID          FIGURE-8           OMNI
   ╭─────╮          ╭───╮           ╭─────╮
 ╭─┤     ├─╮      ╭─┤   ├─╮       ╭─┤     ├─╮
│  │     │  │    │  │   │  │     │  │     │  │
│  │  ●  │  │    │  │ ● │  │     │  │  ●  │  │
│  │     │  │    │  │   │  │     │  │     │  │
 ╰─┤     ├─╯      ╰─┤   ├─╯       ╰─┤     ├─╯
   ╰─────╯          ╰───╯           ╰─────╯
 Front pickup    Front + Back      All around

Recording Levels

Target Peaks: -12 to -6 dBFS
Average Level: -18 to -12 dBFS

Why?
- Leaves headroom for processing
- Matches analog equipment sweet spot
- Prevents digital clipping
- 24-bit has plenty of resolution

Recording Best Practices

• [ ] Set gain before recording (never during)
• [ ] Record at 24-bit minimum
• [ ] Use 48kHz for video, 44.1kHz for music
• [ ] Monitor with headphones
• [ ] Record a few seconds of room tone
• [ ] Check phase when using multiple mics
• [ ] Label takes immediately

Essential Processing

EQ (Equalization)

Types:
- High-pass filter (HPF): Removes low frequencies
- Low-pass filter (LPF): Removes high frequencies
- Bell/Peak: Boost or cut specific frequency
- Shelf: Boost or cut above/below frequency

Common Applications:
┌─────────────────────────────────────────┐
│ Vocal:                                  │
│ HPF at 80-100Hz (remove rumble)         │
│ Cut 200-300Hz (reduce mud)              │
│ Boost 3-5kHz (presence)                 │
│ Shelf boost 10kHz+ (air)                │
├─────────────────────────────────────────┤
│ Kick Drum:                              │
│ Boost 50-80Hz (sub)                     │
│ Cut 300-400Hz (boxiness)                │
│ Boost 3-5kHz (attack/click)             │
├─────────────────────────────────────────┤
│ Electric Guitar:                        │
│ HPF at 80Hz                             │
│ Cut 400-600Hz if muddy                  │
│ Boost 2-4kHz (bite)                     │
└─────────────────────────────────────────┘

Compression

Parameters:
┌────────────────────────────────────────────────┐
│ Threshold: Level where compression starts      │
│ Ratio: Amount of compression (4:1, 8:1, etc.) │
│ Attack: Time to engage (fast=punch, slow=body)│
│ Release: Time to disengage                     │
│ Knee: Hard (sudden) or soft (gradual)         │
│ Makeup Gain: Restore lost volume               │
└────────────────────────────────────────────────┘

Common Settings:
┌─────────────────────────────────────────┐
│ Vocals: 3:1-4:1, medium attack/release  │
│ Drums: 4:1-8:1, fast attack             │
│ Bass: 4:1, medium attack                │
│ Mix Bus: 2:1-3:1, slow attack, 1-3dB GR│
└─────────────────────────────────────────┘

Reverb

Key Parameters:
- Pre-delay: Time before reverb (50-100ms for vocals)
- Decay/RT60: How long it lasts
- Damping: High-frequency absorption
- Size: Room dimensions
- Mix/Wet-Dry: Balance with original

Delay

Mixing

Gain Staging

Signal Flow Levels:
┌─────────────────────────────────────────┐
│ Input:     -18 dBFS (sweet spot)        │
│            ↓                            │
│ Plugin 1:  Output matches input level   │
│            ↓                            │
│ Plugin 2:  Output matches input level   │
│            ↓                            │
│ Fader:     Unity (0 dB) ideally         │
│            ↓                            │
│ Bus:       -12 to -6 dBFS peaks         │
│            ↓                            │
│ Master:    -6 dBFS peaks (before limit) │
└─────────────────────────────────────────┘

Panning Guidelines

Stereo Field:
Hard L    L    L-C    CENTER    R-C    R    Hard R
  │       │     │        │        │     │       │
  │       │     │        │        │     │       │
Rhythm   Gtr1  Keys1   Vocal    Keys2  Gtr2   Rhythm
 Gtr            BVox1   Bass    BVox2          Gtr
                        Kick
                        Snare

Common Positions:
- Center (0): Lead vocal, bass, kick, snare
- Slight L/R: Main instruments, harmonies
- Wide L/R: Rhythm guitars, stereo keys
- Hard L/R: Doubled parts, effects

Frequency Balancing

Making space (subtractive EQ approach):
┌────────────────────────────────────────────┐
│ Kick:  Boost 60Hz   │ Cut from Bass here   │
│ Bass:  Boost 80-100 │ Cut from Kick here   │
│ Gtr:   Owns 1-3kHz  │ Cut vocal slightly   │
│ Vocal: Boost 3-5kHz │ Cut guitar slightly  │
└────────────────────────────────────────────┘

Each element should have its own frequency "home"

Mix Bus Processing

Typical Chain:
1. Subtle EQ (tonal shaping)
2. Light compression (2:1, 1-3dB GR)
3. Stereo enhancement (if needed)
4. Limiter (only for reference bounce)

Keep mix bus processing minimal!
Leave headroom for mastering.

Mastering

Mastering Goals

1. Tonal Balance: Even frequency response
2. Dynamics: Appropriate loudness/dynamics
3. Stereo Image: Width and mono compatibility
4. Translation: Sounds good everywhere
5. Format: Prepare for distribution

Loudness Standards

Mastering Chain (Typical)

1. Reference Import
   └─→ Compare to commercial releases

2. EQ (Linear Phase)
   └─→ Broad adjustments only

3. Multiband Compression
   └─→ Control specific bands

4. Stereo Width
   └─→ Enhance or narrow

5. Limiting
   └─→ Final loudness, prevent clipping

6. Dither (if needed)
   └─→ 16-bit conversion only

Podcast Production

Recording Setup

Recommended Signal Chain:
Mic → Audio Interface → DAW

Equipment Tiers:
┌─────────────────────────────────────────┐
│ Budget:    USB Mic (Blue Yeti, AT2020)  │
│ Mid:       SM7B + Cloudlifter + Interface│
│ Pro:       Large diaphragm condenser    │
│            + high-end preamp            │
└─────────────────────────────────────────┘

Podcast Processing Chain

1. Noise Reduction
   └─→ Remove room noise, hum

2. De-essing
   └─→ Tame harsh "s" sounds

3. Compression
   └─→ Even out volume (3:1-4:1)

4. EQ
   └─→ HPF at 80Hz, presence boost

5. Limiter
   └─→ Catch peaks, -1 dBFS ceiling

Target: -16 LUFS (mono), -19 LUFS (stereo)
True Peak: -1 dBTP

Podcast Export Settings

Standard:
- Format: MP3 or AAC
- Bitrate: 128 kbps (mono), 192 kbps (stereo)
- Sample rate: 44.1 kHz

High Quality:
- Format: AAC
- Bitrate: 256 kbps
- Sample rate: 48 kHz

Chapters: Use ID3 tags for chapter markers
Artwork: 3000x3000 minimum, JPG or PNG

Sound Design

Creating Sounds

Approaches:
1. Recording: Capture real sounds (foley)
2. Synthesis: Create from oscillators
3. Sampling: Manipulate existing sounds
4. Processing: Transform through effects
5. Layering: Combine multiple sources

Synthesis Types

Layering Formula

Sound Design Stack:
┌─────────────────────────────────────┐
│ TOP: High frequencies (air, presence)│
│ MID: Character, tone                │
│ LOW: Foundation, weight             │
│ SUB: Felt, not heard (20-60Hz)      │
│ TRANSIENT: Attack, punch            │
└─────────────────────────────────────┘

Process each layer separately, combine at end.

DAWs (Digital Audio Workstations)

File Formats

Export Specifications

For Mastering:
- Format: WAV
- Bit depth: 24-bit (or 32-bit float)
- Sample rate: Match session (typically 44.1/48kHz)
- Headroom: -6 dBFS peaks

For Distribution:
- Format: WAV or FLAC (lossless)
- Bit depth: 16-bit (with dither)
- Sample rate: 44.1 kHz
- Loudness: Platform target LUFS

Best Practices

DO:

• Record at proper levels (-18 dBFS average)
• Use reference tracks
• Take breaks (ear fatigue is real)
• Check mono compatibility
• Use subtractive EQ first
• Process in series, not parallel (usually)
• Save project versions regularly
• Export stems for backup

DON'T:

• Record with processing (usually)
• Mix at high volumes
• Over-compress everything
• Solo instruments too long
• Trust only one playback system
• Rush the mixing process
• Forget to dither (16-bit only)
• Master your own mixes (ideally)

Quality Checklist

Pre-Export

• [ ] Gain staging correct throughout
• [ ] No clipping or distortion
• [ ] Phase issues resolved
• [ ] Mono compatibility checked
• [ ] Low-end translation verified
• [ ] High frequencies not harsh

Final Check

• [ ] Listen on multiple systems
• [ ] Check in car, phone, earbuds
• [ ] Verify format specifications
• [ ] Confirm LUFS target met
• [ ] True peak below -1 dBTP
• [ ] Metadata correctly embedded