agentleFS
topo-aitopo-ai/ai-video-skills/skills/sound-fx-for-video/SKILL.md

sound-fx-for-video

Skill49 starsChanged 6 months agoMITView on GitHub

用于视频制作中的音效方案:搜索下载免费音效、Web Audio API 合成音效、以及在 HyperFrames/Remotion/HTML5 视频中的音频集成。触发词包括:音效、SFX、配乐、BGM、whoosh、click、transition、impact、typing 等。

  • Reads credentials

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 1

    1. ai-workspace-lab/xworkspace-core-skills★ 8100% same

    Edited · 0

    None found.

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository4 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    CC0: Use freely, no attribution

    line 100
  2. Distinctive:

    CC-BY: Use with attribution (add to video description)

    line 101

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

视频音效制作 Skill

视频需要声音。这份 Skill 覆盖两条主路径:

  1. 搜索并下载 免费音效素材
  2. 合成生成 自定义音效(Web Audio API,无需外部文件)

按需求选择路径,实际项目通常会混合使用。


路径一:搜索并下载免费音效

推荐来源

SourceURLLicenseAPIBest For
Freesoundfreesound.orgCC licenses (check per-file)REST API (key required)Huge library, specific sounds
Mixkitmixkit.co/free-sound-effectsFree, no attributionNoQuick grabs, curated quality
Pixabaypixabay.com/sound-effectsFree, no attributionNoClean UI, good variety
BBC SFXbbcsfx.acropolis.org.ukFree for personal/educationalNoPremium BBC quality
ZapSplatzapsplat.comFree with attributionNoGame/comedy/cartoon sounds
SoundBiblesoundbible.comMixed (check per-file)NoQuick one-off downloads

搜索策略

  1. 关键词要具体。
    差:click sound
    好:mouse click sharp UI feedback
    差:whoosh
    好:fast whoosh air sweep transition
  2. 组合关键词搜索:
    • Object + action: "glass shatter", "paper crumple"
    • Mood + type: "cinematic impact bass", "playful pop notification"
    • Context: "UI button click feedback", "slide transition whoosh"
  3. 常见音效分类与搜索词:

    CategorySearch Terms
    Transitionswhoosh, sweep, swoosh, riser, dive
    UI feedbackclick, tap, pop, blip, notification, ding
    Impactsboom, hit, slam, thud, punch, bass drop
    Typingkeyboard, typing, keystroke, mechanical
    Revealsshimmer, sparkle, magic, chime, glow
    Movementslide, swoop, flutter, bounce, elastic
    Atmosphereambient, drone, hum, tension, pulse

程序化下载

Freesound API (best for automation):

# 1. Get API key from https://freesound.org/apiv2/apply/
# 2. Search
curl "https://freesound.org/apiv2/search/text/?query=whoosh+transition&fields=id,name,previews,duration,license&token=YOUR_API_KEY"

# 3. Download preview (mp3, no auth needed)
curl -o whoosh.mp3 "https://freesound.org/data/previews/ID_ID_preview.mp3"

# 4. Download full quality (OAuth2 needed)

Simple curl download (sources with direct links):

# Pixabay (find the download URL from browser network tab)
curl -L -o click.mp3 "https://cdn.pixabay.com/audio/..."

# Mixkit
curl -L -o transition.wav "https://assets.mixkit.co/active_storage/sfx/..."

Python helper (for batch downloads):

import urllib.request
import json

FREESOUND_TOKEN = "YOUR_TOKEN"

def search_sfx(query, max_results=5):
    url = f"https://freesound.org/apiv2/search/text/?query={query}&fields=id,name,previews,duration,license&token={FREESOUND_TOKEN}"
    with urllib.request.urlopen(url) as r:
        return json.loads(r.read())["results"][:max_results]

def download_preview(sound_id, filename):
    url = f"https://freesound.org/data/previews/{sound_id//1000}/{sound_id}_{sound_id}_preview.mp3"
    urllib.request.urlretrieve(url, filename)

版权检查

发布前必须检查授权协议:

  • CC0: Use freely, no attribution
  • CC-BY: Use with attribution (add to video description)
  • CC-BY-NC: Non-commercial only — do NOT use for monetized videos
  • CC-BY-SA: Derivatives must share same license

路径二:使用 Web Audio API 合成音效

无需音频文件,直接在浏览器中实时合成,适合代码驱动视频流程(HyperFrames、Remotion、HTML5)。

快速参考:常见视频音效模式

UI Click / Tap

function playClick(audioCtx) {
  const osc = audioCtx.createOscillator();
  const gain = audioCtx.createGain();
  osc.connect(gain).connect(audioCtx.destination);
  osc.frequency.setValueAtTime(1200, audioCtx.currentTime);
  osc.frequency.exponentialRampToValueAtTime(600, audioCtx.currentTime + 0.05);
  gain.gain.setValueAtTime(0.3, audioCtx.currentTime);
  gain.gain.exponentialRampToValueAtTime(0.001, audioCtx.currentTime + 0.08);
  osc.start(); osc.stop(audioCtx.currentTime + 0.08);
}

Notification Pop / Ding

function playPop(audioCtx) {
  const osc = audioCtx.createOscillator();
  const gain = audioCtx.createGain();
  osc.type = 'sine';
  osc.connect(gain).connect(audioCtx.destination);
  osc.frequency.setValueAtTime(880, audioCtx.currentTime);
  osc.frequency.exponentialRampToValueAtTime(1760, audioCtx.currentTime + 0.05);
  gain.gain.setValueAtTime(0.4, audioCtx.currentTime);
  gain.gain.exponentialRampToValueAtTime(0.001, audioCtx.currentTime + 0.2);
  osc.start(); osc.stop(audioCtx.currentTime + 0.2);
}

Whoosh / Transition Sweep

function playWhoosh(audioCtx) {
  const bufferSize = audioCtx.sampleRate * 0.4;
  const buffer = audioCtx.createBuffer(1, bufferSize, audioCtx.sampleRate);
  const data = buffer.getChannelData(0);
  for (let i = 0; i < bufferSize; i++) {
    data[i] = (Math.random() * 2 - 1) * Math.pow(1 - i / bufferSize, 2);
  }
  const source = audioCtx.createBufferSource();
  source.buffer = buffer;
  const filter = audioCtx.createBiquadFilter();
  filter.type = 'bandpass'; filter.Q.value = 5;
  filter.frequency.setValueAtTime(200, audioCtx.currentTime);
  filter.frequency.exponentialRampToValueAtTime(4000, audioCtx.currentTime + 0.2);
  filter.frequency.exponentialRampToValueAtTime(200, audioCtx.currentTime + 0.4);
  const gain = audioCtx.createGain();
  gain.gain.setValueAtTime(0, audioCtx.currentTime);
  gain.gain.linearRampToValueAtTime(0.5, audioCtx.currentTime + 0.1);
  gain.gain.linearRampToValueAtTime(0, audioCtx.currentTime + 0.4);
  source.connect(filter).connect(gain).connect(audioCtx.destination);
  source.start();
}

Impact / Bass Hit

function playImpact(audioCtx) {
  const osc = audioCtx.createOscillator();
  const gain = audioCtx.createGain();
  osc.type = 'sine';
  osc.connect(gain).connect(audioCtx.destination);
  osc.frequency.setValueAtTime(150, audioCtx.currentTime);
  osc.frequency.exponentialRampToValueAtTime(30, audioCtx.currentTime + 0.3);
  gain.gain.setValueAtTime(1, audioCtx.currentTime);
  gain.gain.exponentialRampToValueAtTime(0.001, audioCtx.currentTime + 0.4);
  osc.start(); osc.stop(audioCtx.currentTime + 0.4);
  // Add noise burst layer
  const bufferSize = audioCtx.sampleRate * 0.1;
  const buffer = audioCtx.createBuffer(1, bufferSize, audioCtx.sampleRate);
  const data = buffer.getChannelData(0);
  for (let i = 0; i < bufferSize; i++) data[i] = (Math.random() * 2 - 1);
  const noise = audioCtx.createBufferSource();
  noise.buffer = buffer;
  const noiseGain = audioCtx.createGain();
  noiseGain.gain.setValueAtTime(0.5, audioCtx.currentTime);
  noiseGain.gain.exponentialRampToValueAtTime(0.001, audioCtx.currentTime + 0.15);
  noise.connect(noiseGain).connect(audioCtx.destination);
  noise.start();
}

Typing / Keystroke

function playKey(audioCtx) {
  const bufferSize = audioCtx.sampleRate * 0.03;
  const buffer = audioCtx.createBuffer(1, bufferSize, audioCtx.sampleRate);
  const data = buffer.getChannelData(0);
  for (let i = 0; i < bufferSize; i++) {
    data[i] = (Math.random() * 2 - 1) * Math.pow(1 - i / bufferSize, 4);
  }
  const source = audioCtx.createBufferSource();
  source.buffer = buffer;
  const filter = audioCtx.createBiquadFilter();
  filter.type = 'highpass'; filter.frequency.value = 2000;
  const gain = audioCtx.createGain();
  gain.gain.value = 0.15;
  source.connect(filter).connect(gain).connect(audioCtx.destination);
  source.start();
}

Rise / Tension Builder

function playRise(audioCtx, duration = 1.5) {
  const osc = audioCtx.createOscillator();
  osc.type = 'sawtooth';
  const filter = audioCtx.createBiquadFilter();
  filter.type = 'lowpass'; filter.Q.value = 8;
  const gain = audioCtx.createGain();
  osc.connect(filter).connect(gain).connect(audioCtx.destination);
  filter.frequency.setValueAtTime(100, audioCtx.currentTime);
  filter.frequency.exponentialRampToValueAtTime(3000, audioCtx.currentTime + duration);
  gain.gain.setValueAtTime(0.01, audioCtx.currentTime);
  gain.gain.linearRampToValueAtTime(0.3, audioCtx.currentTime + duration);
  osc.start(); osc.stop(audioCtx.currentTime + duration);
}

Sparkle / Shimmer

function playSparkle(audioCtx) {
  const notes = [261.63, 329.63, 392.00, 523.25, 659.25]; // C5 E5 G5 C6 E6
  notes.forEach((freq, i) => {
    const osc = audioCtx.createOscillator();
    osc.type = 'sine';
    const gain = audioCtx.createGain();
    osc.connect(gain).connect(audioCtx.destination);
    const start = audioCtx.currentTime + i * 0.06;
    osc.frequency.value = freq;
    gain.gain.setValueAtTime(0, start);
    gain.gain.linearRampToValueAtTime(0.2, start + 0.02);
    gain.gain.exponentialRampToValueAtTime(0.001, start + 0.4);
    osc.start(start); osc.stop(start + 0.4);
  });
}

模式速查表

SoundCore TechniqueKey Parameters
Click/TapShort sine + fast decayfreq 800-1500Hz, dur < 100ms
Pop/BubbleSine + freq ramp upfreq sweep up, short
WhooshFiltered noise + bandpass sweepbandpass sweep 200→4k→200
ImpactLow sine + noise burstfreq 150→30Hz, noise < 150ms
TypingHighpass filtered noiseHPF 2kHz+, dur < 50ms
Rise/TensionSawtooth + filter sweepLPF 100→3kHz over duration
SparkleArpeggiated sine clusterC-E-G-C-E, staggered 60ms
Boom/RumbleVery low sine + slow decayfreq 40-80Hz, dur 0.5-2s
Slide/MoveSine with pitch bendfreq ramp up or down
Error/BuzzSquare wave buzzlow freq square, short burst

路径三:HyperFrames / Remotion 集成

HyperFrames 音频集成

在 HyperFrames 中,音频常见有两种方式:

方式 A:Web Audio API(实时合成)

// In your HyperFrames component
const audioCtx = new AudioContext();

function MyAnimation() {
  const handleClick = () => playClick(audioCtx);

  return (
    <div onClick={handleClick}>
      <h1>Click Me</h1>
    </div>
  );
}

方式 B:预加载音频文件

// 1. Import audio file
import whooshSfx from "./sfx/whoosh.mp3";

// 2. Play on animation event
const audio = new Audio(whooshSfx);
audio.play();

方式 C:Remotion <Audio> 组件(使用 Remotion 时)

import { Audio, useCurrentFrame } from "remotion";

export const MyComp = () => {
  const frame = useCurrentFrame();
  return (
    <>
      <Audio src={staticFile("whoosh.mp3")} startFrom={30} volume={0.5} />
    </>
  );
};

视听同步建议

  1. Trigger sounds at animation keyframes, not at random times
  2. Keep sounds short (50-300ms for UI, 300-800ms for transitions)
  3. Layer sounds — a whoosh + impact combo feels better than either alone
  4. Volume hierarchy: BGM (0.1-0.2) < Transition SFX (0.2-0.3) < Key SFX (0.3-0.5) < Voiceover (1.0)
  5. Always test with audio on — silent playback hides timing issues

选择流程

Need a sound effect?
├── Is it a simple UI/click/pop? → Synthesize with Web Audio API (instant, no files)
├── Is it a complex/natural sound (footsteps, crowd, rain)? → Download from Freesound/Pixabay
├── Need precise control over timing? → Web Audio API synthesis
├── Need realism? → Download real recordings
└── Both? → Mix: download base layer + synthesize accents on top

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.