Implement audio cross fading

https://github.com/mifi/editly/issues/67

README.md

@@ -110,6 +110,8 @@ Edit specs are JavaScript / JSON objects describing the whole edit operation wit
     transition: {
       duration: 0.5,
       name: 'random',
+      audioOutCurve: 'tri',
+      audioInCurve: 'tri',
     },
     layer: {
       fontPath,
@@ -167,6 +169,8 @@ Edit specs are JavaScript / JSON objects describing the whole edit operation wit
 | `defaults.transition` | | An object `{ name, duration }` describing the default transition. Set to **null** to disable transitions | | |
 | `defaults.transition.duration` | `--transition-duration` | Default transition duration | `0.5` | sec |
 | `defaults.transition.name` | `--transition-name` | Default transition type. See [Transition types](#transition-types) | `random` | |
+| `defaults.transition.audioOutCurve` | | Default [fade out curve](https://trac.ffmpeg.org/wiki/AfadeCurves) in audio cross fades | `tri` | |
+| `defaults.transition.audioInCurve` | | Default [fade in curve](https://trac.ffmpeg.org/wiki/AfadeCurves) in audio cross fades | `tri` | |
 | `clips[]` | | List of clip objects that will be played in sequence. Each clip can have one or more layers. | | |
 | `clips[].duration` | | Clip duration. See `defaults.duration`. If unset, the clip duration will be that of the **first video layer**. | `defaults.duration` | |
 | `clips[].transition` | | Specify transition at the **end** of this clip. See `defaults.transition` | `defaults.transition` | |

audio.js

@@ -4,7 +4,7 @@ const execa = require('execa');
 const flatMap = require('lodash/flatMap');
 const fs = require('fs-extra');
 
-const { getFfmpegCommonArgs, getCutFromArgs, createConcatFile } = require('./ffmpeg');
+const { getFfmpegCommonArgs, getCutFromArgs } = require('./ffmpeg');
 const { readFileStreams } = require('./util');
 
 module.exports = ({ ffmpegPath, ffprobePath, enableFfmpegLog, verbose }) => {
@@ -15,21 +15,23 @@ module.exports = ({ ffmpegPath, ffprobePath, enableFfmpegLog, verbose }) => {
 
     const mergedAudioPath = join(tmpDir, 'audio-merged.flac');
 
-    const segments = await pMap(clips, async (clip, i) => {
+    const clipsOut = await pMap(clips, async (clip, i) => {
       const clipAudioPath = join(tmpDir, `clip${i}-audio.flac`);
 
-      const audioLayers = clip.layers.filter(({ type, visibleFrom, visibleUntil }) => (
+      const { duration, layers, transition } = clip;
+
+      const audioLayers = layers.filter(({ type, visibleFrom, visibleUntil }) => (
         ['audio', 'video'].includes(type)
         // TODO We don't support audio for visibleFrom/visibleUntil layers
         && !visibleFrom && visibleUntil == null));
 
       async function createSilence(outPath) {
-        if (verbose) console.log('create silence', clip.duration);
+        if (verbose) console.log('create silence', duration);
         const args = [
           '-f', 'lavfi', '-i', 'anullsrc=channel_layout=stereo:sample_rate=44100',
           '-sample_fmt', 's32',
           '-ar', '48000',
-          '-t', clip.duration,
+          '-t', duration,
           '-c:a', 'flac',
           '-y',
           outPath,
@@ -39,7 +41,7 @@ module.exports = ({ ffmpegPath, ffprobePath, enableFfmpegLog, verbose }) => {
 
       if (audioLayers.length > 0) {
         const processedAudioLayersRaw = await pMap(audioLayers, async (audioLayer, j) => {
-          const { path, cutFrom, audioCutTo, speedFactor } = audioLayer;
+          const { path, cutFrom, cutTo, speedFactor } = audioLayer;
 
           const streams = await readFileStreams(ffprobePath, path);
           if (!streams.some((s) => s.codec_type === 'audio')) return undefined;
@@ -58,7 +60,7 @@ module.exports = ({ ffmpegPath, ffprobePath, enableFfmpegLog, verbose }) => {
               atempoFilter = `atempo=${atempo}`;
             }
 
-            const cutToArg = (audioCutTo - cutFrom) * speedFactor;
+            const cutToArg = (cutTo - cutFrom) * speedFactor;
 
             const args = [
               ...getFfmpegCommonArgs({ enableFfmpegLog }),
@@ -109,25 +111,41 @@ module.exports = ({ ffmpegPath, ffprobePath, enableFfmpegLog, verbose }) => {
         await createSilence(clipAudioPath);
       }
 
-      // https://superuser.com/a/853262/658247
-      return resolve(clipAudioPath);
+      return {
+        path: resolve(clipAudioPath), // https://superuser.com/a/853262/658247
+        transition,
+      };
     }, { concurrency: 4 });
 
-    const concatFilePath = join(tmpDir, 'audio-segments.txt');
+    if (clipsOut.length < 2) {
+      await fs.rename(clipsOut[0].path, mergedAudioPath);
+    } else {
+      console.log('Combining audio', clipsOut.map(({ path }) => basename(path)));
 
-    console.log('Combining audio', segments.map((s) => basename(s)), concatFilePath);
+      let inStream = '[0:a]';
+      const filterGraph = clipsOut.slice(0, -1).map(({ transition }, i) => {
+        const outStream = `[concat${i}]`;
 
-    await createConcatFile(segments, concatFilePath);
+        const epsilon = 0.0001; // If duration is 0, ffmpeg seems to default to 1 sec instead, hence epsilon.
+        let ret = `${inStream}[${i + 1}:a]acrossfade=d=${Math.max(epsilon, transition.duration)}:c1=${transition.audioOutCurve || 'tri'}:c2=${transition.audioInCurve || 'tri'}`;
+
+        inStream = outStream;
+
+        if (i < clipsOut.length - 2) ret += outStream;
+        return ret;
+      }).join(',');
 
       const args = [
         ...getFfmpegCommonArgs({ enableFfmpegLog }),
-      '-f', 'concat', '-safe', '0',
-      '-i', concatFilePath,
+        ...(flatMap(clipsOut, ({ path }) => ['-i', path])),
+        '-filter_complex',
+        filterGraph,
         '-c', 'flac',
         '-y',
         mergedAudioPath,
       ];
       await execa(ffmpegPath, args);
+    }
 
     // TODO don't return audio if only silence?
     return mergedAudioPath;

examples/audio-transition.json5

@@ -0,0 +1,31 @@
+{
+  // enableFfmpegLog: true,
+  outPath: './audio-transition.mp4',
+  keepSourceAudio: true,
+  defaults: {
+    duration: 3,
+    transition: { duration: 1, name: 'directional' },
+  },
+  clips: [
+    { layers: [
+      { type: 'title-background', text: 'Clip 1' },
+      { type: 'audio', path: './assets/sample1.m4a' }
+    ] },
+    { transition: { duration: 0.2 }, layers: [
+      { type: 'title-background', text: 'Clip 2' },
+      { type: 'audio', path: './assets/sample2.m4a' }
+    ] },
+    { transition: { duration: 0 }, layers: [
+      { type: 'title-background', text: 'Clip 3' },
+      { type: 'audio', path: './assets/sample1.m4a' }
+    ] },
+    { transition: { audioInCurve: 'exp', audioOutCurve: 'exp' }, layers: [
+      { type: 'title-background', text: 'Clip 4' },
+      { type: 'audio', path: './assets/sample2.m4a' }
+    ] },
+    { layers: [
+      { type: 'title-background', text: 'Clip 5' },
+      { type: 'audio', path: './assets/sample1.m4a' }
+    ] },
+  ],
+}

examples/single.json5

@@ -0,0 +1,8 @@
+{
+  // This is a test of a single clip to make sure that it works
+  outPath: './single.mp4',
+  keepSourceAudio: true,
+  clips: [
+    { layers: [{ type: 'video', path: './assets/lofoten.mp4', cutFrom: 0, cutTo: 2 }] },
+  ],
+}

parseConfig.js

@@ -20,6 +20,8 @@ async function parseConfig({ defaults: defaultsIn = {}, clips, allowRemoteReques
     transition: defaultsIn.transition === null ? null : {
       duration: 0.5,
       name: 'random',
+      audioOutCurve: 'tri',
+      audioInCurve: 'tri',
       ...defaultsIn.transition,
     },
   };
@@ -108,6 +110,7 @@ async function parseConfig({ defaults: defaultsIn = {}, clips, allowRemoteReques
     let layersOut = flatMap(await pMap(layers, async (layerIn) => {
       const globalLayerDefaults = defaults.layer || {};
       const thisLayerDefaults = (defaults.layerType || {})[layerIn.type];
+
       const layer = { ...globalLayerDefaults, ...thisLayerDefaults, ...layerIn };
       const { type, path } = layer;
 
@@ -129,10 +132,7 @@ async function parseConfig({ defaults: defaultsIn = {}, clips, allowRemoteReques
         const inputWidth = isRotated ? heightIn : widthIn;
         const inputHeight = isRotated ? widthIn : heightIn;
 
-        // Compensate for transition duration
-        const audioCutTo = Math.max(cutFrom, cutTo - transition.duration);
-
-        return { ...layer, cutFrom, cutTo, audioCutTo, inputDuration, framerateStr, inputWidth, inputHeight };
+        return { ...layer, cutFrom, cutTo, inputDuration, framerateStr, inputWidth, inputHeight };
       }
 
       // Audio is handled later
@@ -178,10 +178,7 @@ async function parseConfig({ defaults: defaultsIn = {}, clips, allowRemoteReques
 
         const speedFactor = clipDuration / inputDuration;
 
-        // Compensate for transition duration
-        const audioCutTo = Math.max(cutFrom, cutTo - transition.duration);
-
-        return { ...layer, cutFrom, cutTo, audioCutTo, speedFactor };
+        return { ...layer, cutFrom, cutTo, speedFactor };
       }
 
       if (layer.type === 'video') {

transitions.js

@@ -27,37 +27,37 @@ function getTransitionEasingFunction(easing, transitionName) {
 function calcTransition(defaults, transition, isLastClip) {
   if (transition === null || isLastClip) return { duration: 0 };
 
+  const getTransitionDefault = (key) => (defaults.transition && defaults.transition[key]);
+
   let transitionOrDefault = {
-    name: (transition && transition.name) || (defaults.transition && defaults.transition.name),
-    duration: (transition && transition.duration != null) ? transition.duration : (defaults.transition && defaults.transition.duration),
-    params: (transition && transition.params) || (defaults.transition && defaults.transition.params),
-    easing: (transition && transition.easing !== undefined) ? transition.easing : (defaults.transition && defaults.transition.easing),
+    name: (transition && transition.name) || getTransitionDefault('name'),
+    duration: (transition && transition.duration != null) ? transition.duration : getTransitionDefault('duration'),
+    params: (transition && transition.params) || getTransitionDefault('params'),
+    easing: (transition && transition.easing !== undefined) ? transition.easing : getTransitionDefault('easing'),
+    audioOutCurve: (transition && transition.audioOutCurve) || getTransitionDefault('audioOutCurve'),
+    audioInCurve: (transition && transition.audioInCurve) || getTransitionDefault('audioInCurve'),
   };
 
   assert(!transitionOrDefault.duration || transitionOrDefault.name, 'Please specify transition name or set duration to 0');
 
   if (transitionOrDefault.name === 'random' && transitionOrDefault.duration) {
-    transitionOrDefault = { easing: transitionOrDefault.easing, name: getRandomTransition(), duration: transitionOrDefault.duration };
+    transitionOrDefault = { ...transitionOrDefault, name: getRandomTransition() };
   }
 
-  const getTransitionByAlias = () => {
   const aliasedTransition = {
     'directional-left': { name: 'directional', params: { direction: [1, 0] } },
     'directional-right': { name: 'directional', params: { direction: [-1, 0] } },
     'directional-down': { name: 'directional', params: { direction: [0, 1] } },
     'directional-up': { name: 'directional', params: { direction: [0, -1] } },
   }[transitionOrDefault.name];
-    if (aliasedTransition) return { ...transitionOrDefault, ...aliasedTransition };
-    return transitionOrDefault;
-  };
-
-  const outTransition = getTransitionByAlias();
+  if (aliasedTransition) {
+    transitionOrDefault = { ...transitionOrDefault, ...aliasedTransition };
+  }
 
   return {
-    name: outTransition.name,
-    duration: outTransition.duration || 0,
-    params: outTransition.params,
-    easingFunction: getTransitionEasingFunction(outTransition.easing, outTransition.name),
+    ...transitionOrDefault,
+    duration: transitionOrDefault.duration || 0,
+    easingFunction: getTransitionEasingFunction(transitionOrDefault.easing, transitionOrDefault.name),
   };
 }