Fixing #626 with ffmpeg's low default audio bitrate

[oweals/peertube.git] / server / helpers / ffmpeg-utils.ts

diff --git a/server/helpers/ffmpeg-utils.ts b/server/helpers/ffmpeg-utils.ts
index f0623c88bd25cb82e9ed3b98adb4e79161e82c5b..eb1c86ab96a11a8a613357ba11cebaac51dd148c 100644 (file)
--- a/server/helpers/ffmpeg-utils.ts
+++ b/server/helpers/ffmpeg-utils.ts
@@ -5,6 +5,7 @@ import { CONFIG, VIDEO_TRANSCODING_FPS } from '../initializers'
 import { unlinkPromise } from './core-utils'
 import { processImage } from './image-utils'
 import { logger } from './logger'
+import { checkFFmpegEncoders } from '../initializers/checker'
 
 async function getVideoFileResolution (path: string) {
   const videoStream = await getVideoFileStream(path)
@@ -85,12 +86,8 @@ function transcode (options: TranscodeOptions) {
   return new Promise<void>(async (res, rej) => {
     let command = ffmpeg(options.inputPath)
                     .output(options.outputPath)
-                    .videoCodec('libx264')
                     .outputOption('-threads ' + CONFIG.TRANSCODING.THREADS)
-                    .outputOption('-movflags faststart')
-                    .outputOption('-b_strategy 1') // NOTE: b-strategy 1 - heuristic algorythm, 16 is optimal B-frames for it
-                    .outputOption('-bf 16') // NOTE: Why 16: https://github.com/Chocobozzz/PeerTube/pull/774. b-strategy 2 -> B-frames<16
-                    // .outputOption('-crf 18')
+                    .preset(standard)
 
     let fps = await getVideoFileFPS(options.inputPath)
     if (options.resolution !== undefined) {
@@ -149,3 +146,117 @@ function getVideoFileStream (path: string) {
     })
   })
 }
+
+/**
+ * A slightly customised version of the 'veryfast' x264 preset
+ *
+ * The veryfast preset is right in the sweet spot of performance
+ * and quality. Superfast and ultrafast will give you better
+ * performance, but then quality is noticeably worse.
+ */
+function veryfast (ffmpeg) {
+  ffmpeg
+    .preset(standard)
+    .outputOption('-preset:v veryfast')
+    .outputOption(['--aq-mode=2', '--aq-strength=1.3'])
+  /*
+  MAIN reference: https://slhck.info/video/2017/03/01/rate-control.html
+  Our target situation is closer to a livestream than a stream,
+  since we want to reduce as much a possible the encoding burden,
+  altough not to the point of a livestream where there is a hard
+  constraint on the frames per second to be encoded.
+
+  why '--aq-mode=2 --aq-strength=1.3' instead of '-profile:v main'?
+    Make up for most of the loss of grain and macroblocking
+    with less computing power.
+  */
+}
+
+/**
+ * A preset optimised for a stillimage audio video
+ */
+function audio (ffmpeg) {
+  ffmpeg
+    .preset(veryfast)
+    .outputOption('-tune stillimage')
+}
+
+/**
+ * A toolbox to play with audio
+ */
+namespace audio {
+  export const get = (ffmpeg, pos = 0) => {
+    // without position, ffprobe considers the last input only
+    // we make it consider the first input only
+    ffmpeg
+      .ffprobe(pos, (_,data) => {
+        return data['streams'].find(stream => {
+          return stream['codec_type'] === 'audio'
+        })
+      })
+  }
+
+  export namespace bitrate {
+    export const baseKbitrate = 384
+
+    const toBits = (kbits: number): number => { return kbits * 8000 }
+
+    export const aac = (bitrate: number): number => {
+      switch (true) {
+      case bitrate > toBits(384):
+        return baseKbitrate
+      default:
+        return -1 // we interpret it as a signal to copy the audio stream as is
+      }
+    }
+
+    export const mp3 = (bitrate: number): number => {
+      switch (true) {
+      case bitrate <= toBits(192):
+        return 128
+      case bitrate <= toBits(384):
+        return 256
+      default:
+        return baseKbitrate
+      }
+    }
+  }
+}
+
+/**
+ * Standard profile, with variable bitrate audio and faststart.
+ *
+ * As for the audio, quality '5' is the highest and ensures 96-112kbps/channel
+ * See https://trac.ffmpeg.org/wiki/Encode/AAC#fdk_vbr
+ */
+async function standard (ffmpeg) {
+  let _bitrate = audio.bitrate.baseKbitrate
+  let _ffmpeg = ffmpeg
+    .format('mp4')
+    .videoCodec('libx264')
+    .outputOption('-level 3.1') // 3.1 is the minimal ressource allocation for our highest supported resolution
+    .outputOption('-b_strategy 1') // NOTE: b-strategy 1 - heuristic algorythm, 16 is optimal B-frames for it
+    .outputOption('-bf 16') // NOTE: Why 16: https://github.com/Chocobozzz/PeerTube/pull/774. b-strategy 2 -> B-frames<16
+    .outputOption('-movflags faststart')
+  let _audio = audio.get(_ffmpeg)
+
+  if (!_audio) return _ffmpeg.noAudio()
+
+  // we try to reduce the ceiling bitrate by making rough correspondances of bitrates
+  // of course this is far from perfect, but it might save some space in the end
+  if (audio.bitrate[_audio['codec_name']]) {
+    _bitrate = audio.bitrate[_audio['codec_name']](_audio['bit_rate'])
+    if (_bitrate === -1) {
+      return _ffmpeg.audioCodec('copy')
+    }
+  }
+
+  // we favor VBR, if a good AAC encoder is available
+  if ((await checkFFmpegEncoders()).get('libfdk_aac')) {
+    return _ffmpeg
+      .audioCodec('libfdk_aac')
+      .audioQuality(5)
+  }
+
+  return _ffmpeg.audioBitrate(_bitrate)
+}