0.1.15

fix: Исправлена проблема с звуком и телепортами по частям видео

README.md

@@ -28,20 +28,29 @@ brew install ffmpeg gpac
 Before running, make sure `FFmpeg` and `MP4Box` are installed (see Install).
 
 ```bash
-# Run via npx (no install)
 npx @gromlab/create-vod video.mp4
-
-# Or install globally
-npm install -g @gromlab/create-vod
-create-vod video.mp4
 ```
 
-**Output:** A folder `video/` in the current directory with segments under `{profile}-{codec}/`, DASH/HLS manifests in the root, poster, and thumbnail sprite/VTT (both DASH and HLS are always generated).
+**Output:** In the current directory you'll get:
+```
+video/
+├── manifest.mpd     # DASH manifest
+├── master.m3u8      # HLS master playlist
+├── poster.jpg       # Poster frame
+├── thumbnails.{jpg,vtt}  # Sprite + VTT cues
+├── audio/           # Audio init + segments (AAC)
+├── 1080p/           # H.264 1080p init + segments
+├── 720p/            # H.264 720p
+├── 480p/            # H.264 480p
+├── 360p/            # H.264 360p
+├── 1080p-av1/       # AV1 1080p (if av1 selected)
+└── ...              # Other profiles/codecs as {profile}-{codec}
+```
 
 ## CLI Usage
 
 ```bash
-create-vod <input-video> [output-dir] [-r resolutions] [-c codec] [-p poster-timecode] [-e encoder] [-d decoder]
+create-vod <input-video> [output-dir] [-r resolutions] [-c codec] [-p poster-timecode] [-e encoder] [-d decoder] [-m]
 ```
 
 ### Main arguments
@@ -60,33 +69,37 @@ create-vod <input-video> [output-dir] [-r resolutions] [-c codec] [-p poster-tim
 | `-p, --poster`      | Poster timecode         | `HH:MM:SS` or seconds      | `00:00:00` | `-p 00:00:05` or `-p 10`      |
 | `-e, --encoder`     | Video encoder           | `auto`, `nvenc`, `qsv`, `amf`, `vaapi`, `videotoolbox`, `v4l2`, `cpu` | `auto`   | `-e nvenc` |
 | `-d, --decoder`     | Video decoder (hwaccel) | `auto`, `nvenc`, `qsv`, `vaapi`, `videotoolbox`, `v4l2`, `cpu` | `auto`   | `-d cpu`   |
+| `-m, --muted`       | Disable audio track     | flag                      | off      | `-m`                            |
 
 ### Examples
 
 ```bash
 # Default (DASH + HLS, auto profiles)
-create-vod video.mp4
+npx @gromlab/create-vod video.mp4
 
 # Custom output directory
-create-vod video.mp4 ./output
+npx @gromlab/create-vod video.mp4 ./output
 
 # Selected resolutions
-create-vod video.mp4 -r 720,1080,1440
+npx @gromlab/create-vod video.mp4 -r 720,1080,1440
 
 # High FPS
-create-vod video.mp4 -r 720@60,1080@60
+npx @gromlab/create-vod video.mp4 -r 720@60,1080@60
 
 # Poster from 5th second
-create-vod video.mp4 -p 5
+npx @gromlab/create-vod video.mp4 -p 5
 
 # Force CPU encode/decode
-create-vod video.mp4 -c h264 -e cpu -d cpu
+npx @gromlab/create-vod video.mp4 -c h264 -e cpu -d cpu
 
 # Force GPU encode + CPU decode
-create-vod video.mp4 -c h264 -e nvenc -d cpu
+npx @gromlab/create-vod video.mp4 -c h264 -e nvenc -d cpu
 
 # Combined parameters
-create-vod video.mp4 ./output -r 720,1080@60,1440@60 -p 00:00:10
+npx @gromlab/create-vod video.mp4 ./output -r 720,1080@60,1440@60 -p 00:00:10
+
+# No audio
+npx @gromlab/create-vod video.mp4 -m
 ```
 
 ### Supported resolutions

README_RU.md

@@ -27,20 +27,30 @@ brew install ffmpeg gpac
 ## Быстрый старт
 Перед запуском убедитесь, что в системе установлены `FFmpeg` и `MP4Box` (см. Install).
 ```bash
-# Использование через npx (без установки)
 npx @gromlab/create-vod video.mp4
-
-# Или глобальная установка
-npm install -g @gromlab/create-vod
-create-vod video.mp4
 ```
 
-**Результат:** В текущей директории будет создана папка `video/` с сегментами в папках `{profile}-{codec}/`, манифестами DASH и HLS в корне, постером и превью спрайтами.
+**Результат:** В текущей директории появится структура выходных файлов:
+```
+video/
+├── manifest.mpd       # DASH манифест
+├── master.m3u8        # HLS мастер-плейлист
+├── poster.jpg         # Постер с указанного таймкода
+├── thumbnails.jpg     # Спрайт превью
+├── thumbnails.vtt     # Таймкоды превью
+├── audio/             # Аудиосегменты (init + m4s)
+├── 1080p/             # Сегменты H.264 1080p
+├── 720p/              # Сегменты H.264 720p
+├── 480p/              # Сегменты H.264 480p
+├── 360p/              # Сегменты H.264 360p
+├── 1080p-av1/         # Сегменты AV1 1080p (если выбран av1)
+└── ...                # Остальные профили/кодеки по схеме {profile}-{codec}
+```
 
 ## Параметры CLI
 
 ```bash
-create-vod <input-video> [output-dir] [-r resolutions] [-c codec] [-p poster-timecode] [-e encoder] [-d decoder]
+create-vod <input-video> [output-dir] [-r resolutions] [-c codec] [-p poster-timecode] [-e encoder] [-d decoder] [-m]
 ```
 
 ### Основные параметры
@@ -59,27 +69,31 @@ create-vod <input-video> [output-dir] [-r resolutions] [-c codec] [-p poster-tim
 | `-p, --poster` | Таймкод для постера | `HH:MM:SS` или секунды | `00:00:00` | `-p 00:00:05` или `-p 10` |
 | `-e, --encoder` | Видео энкодер | `auto`, `nvenc`, `qsv`, `amf`, `vaapi`, `videotoolbox`, `v4l2`, `cpu` | `auto` | `-e nvenc` |
 | `-d, --decoder` | Видео декодер (hwaccel) | `auto`, `nvenc`, `qsv`, `vaapi`, `videotoolbox`, `v4l2`, `cpu` | `auto` | `-d cpu` |
+| `-m, --muted` | Отключить аудио дорожку в выходных файлах | `flag` | `off` | `-m` |
 
 ### Примеры использования
 
 ```bash
 # Базовая конвертация (DASH + HLS, авто кодек, автопрофили)
-create-vod video.mp4
+npx @gromlab/create-vod video.mp4
 
 # Указать выходную директорию
-create-vod video.mp4 ./output
+npx @gromlab/create-vod video.mp4 ./output
 
 # Только выбранные разрешения
-create-vod video.mp4 -r 720,1080,1440
+npx @gromlab/create-vod video.mp4 -r 720,1080,1440
 
 # Высокий FPS для игровых стримов
-create-vod video.mp4 -r 720@60,1080@60
+npx @gromlab/create-vod video.mp4 -r 720@60,1080@60
 
 # Постер с 5-й секунды
-create-vod video.mp4 -p 5
+npx @gromlab/create-vod video.mp4 -p 5
 
 # Комбинация параметров
-create-vod video.mp4 ./output -r 720,1080@60,1440@60 -p 00:00:10
+npx @gromlab/create-vod video.mp4 ./output -r 720,1080@60,1440@60 -p 00:00:10
+
+# Без звука
+npx @gromlab/create-vod video.mp4 -m
 ```
 
 ### Поддерживаемые разрешения

package.json

@@ -1,7 +1,7 @@
 {
   "name": "@gromlab/create-vod",
   "author": "Gromov Sergei",
-  "version": "0.1.9",
+  "version": "0.1.15",
   "description": "DASH/HLS video converter with hardware acceleration (NVENC/QSV/AMF/VAAPI), thumbnails and poster generation",
   "type": "module",
   "main": "./dist/index.js",
@@ -19,6 +19,7 @@
     "dist",
     "bin",
     "README.md",
+    "README_RU.md",
     "LICENSE"
   ],
   "scripts": {

src/cli.ts

@@ -31,6 +31,7 @@ let av1CQ: number | undefined;
 let av1CRF: number | undefined;
 let accelerator: HardwareAccelerationOption | undefined;
 let decoder: HardwareAccelerationOption | undefined;
+let muted = false;
 
 // First pass: extract flags and their values
 for (let i = 0; i < args.length; i++) {
@@ -113,6 +114,8 @@ for (let i = 0; i < args.length; i++) {
     }
     decoder = acc as HardwareAccelerationOption;
     i++;
+  } else if (args[i] === '-m' || args[i] === '--muted') {
+    muted = true;
   } else if (!args[i].startsWith('-')) {
     // Positional argument
     positionalArgs.push(args[i]);
@@ -132,6 +135,7 @@ if (!input) {
   console.error('  -p, --poster         Poster timecode (e.g., 00:00:05 or 10)');
   console.error('  -e, --encoder <type> Hardware encoder: auto|nvenc|qsv|amf|vaapi|videotoolbox|v4l2|cpu (default: auto)');
   console.error('  -d, --decoder <type> Hardware decoder: auto|nvenc|qsv|amf|vaapi|videotoolbox|v4l2|cpu (default: auto)');
+  console.error('  -m, --muted          Disable audio track (no audio in output)');
   console.error('\nQuality Options (override defaults):');
   console.error('  --h264-cq <value>    H.264 GPU CQ value (0-51, lower = better, default: auto)');
   console.error('  --h264-crf <value>   H.264 CPU CRF value (0-51, lower = better, default: auto)');
@@ -278,7 +282,7 @@ console.log(`   File: ${input}`);
 console.log(`   Size: ${fileSizeMB} MB`);
 console.log(`   Resolution: ${metadata.width}x${metadata.height}`);
 console.log(`   FPS: ${metadata.fps.toFixed(2)}`);
-console.log(`   Duration: ${Math.floor(metadata.duration / 60)}m ${Math.floor(metadata.duration % 60)}s`);
+console.log(`   Duration: ${metadata.duration.toFixed(2)}s`);
 console.log(`   Codec: ${metadata.codec}`);
 if (metadata.videoBitrate) {
   console.log(`   Video Bitrate: ${(metadata.videoBitrate / 1000).toFixed(2)} Mbps`);
@@ -309,7 +313,7 @@ if (customProfiles && customProfiles.length > 0) {
     profileResult.warnings.forEach(warn => console.warn(`   - ${warn}`));
   }
 
-  displayProfiles = profileResult.profiles.map(p => p.name);
+  displayProfiles = profileResult.profiles.map(p => p.fps ? `${p.name}@${p.fps}` : p.name);
 } else {
   const autoProfiles = selectProfiles(
     metadata.width,
@@ -317,7 +321,7 @@ if (customProfiles && customProfiles.length > 0) {
     metadata.fps,
     metadata.videoBitrate
   );
-  displayProfiles = autoProfiles.map(p => p.name);
+  displayProfiles = autoProfiles.map(p => p.fps ? `${p.name}@${p.fps}` : p.name);
 }
 
 const manifestDesc = [
@@ -351,6 +355,7 @@ console.log(`   Poster: ${posterPlanned} (will be generated)`);
 console.log(`   Thumbnails: ${thumbnailsPlanned ? 'yes (with VTT)' : 'no'}`);
 console.log(`   Encoder: ${acceleratorDisplay} (available: ${encoderListDisplay})`);
 console.log(`   Decoder: ${decoderDisplay} (available: ${decoderListDisplay})`);
+console.log(`   Audio: ${muted ? 'disabled (muted)' : 'enabled'}`);
 
 // Build quality settings if any are specified
 let quality: QualitySettings | undefined;
@@ -405,6 +410,7 @@ try {
     quality,
     generateThumbnails: true,
     generatePoster: true,
+    muted,
     parallel: true,
     onProgress: (progress) => {
       const stageName = progress.stage === 'encoding' ? 'Encoding' :

src/core/converter.ts

@@ -51,6 +51,7 @@ export async function convertToDash(
     generatePoster: shouldGeneratePoster = true,
     posterTimecode = '00:00:00',
     parallel = true,
+    muted = false,
     onProgress
   } = options;
 
@@ -97,6 +98,7 @@ Formats: ${formats?.join(',') || 'dash,hls'}
       shouldGeneratePoster,
       posterTimecode,
       parallel,
+      muted,
       onProgress
     );
   } finally {
@@ -137,6 +139,7 @@ async function convertToDashInternal(
   generatePosterFlag: boolean,
   posterTimecode: string,
   parallel: boolean,
+  muted: boolean,
   onProgress?: (progress: ConversionProgress) => void
 ): Promise<DashConvertResult> {
 
@@ -160,7 +163,12 @@ async function convertToDashInternal(
 
   // Get video metadata
   const metadata = await getVideoMetadata(input);
-  const hasAudio = metadata.hasAudio;
+  const hasAudio = !muted && metadata.hasAudio;
+  const durationSeconds = metadata.duration;
+
+  // Подгоняем длительность сегмента под общий хронометраж, чтобы не оставался короткий хвост
+  const segmentCount = Math.max(1, Math.ceil(durationSeconds / segmentDuration));
+  const effectiveSegmentDuration = durationSeconds / segmentCount;
 
   // Determine hardware accelerator (auto by default)
   const preferredAccelerator: HardwareAccelerationOption =
@@ -311,13 +319,14 @@ async function convertToDashInternal(
       videoCodec,
       codecPreset,
       metadata.duration,
-      segmentDuration,
+      effectiveSegmentDuration,
       metadata.audioBitrate,
       parallel,
       maxConcurrent,
       type, // Pass codec type to differentiate output files
       codecQuality, // Pass quality settings (CQ/CRF)
       undefined, // optimizations - for future use
+      muted,
       selectedDecoder === 'cpu' ? undefined : selectedDecoder,
       (profileName, percent) => {
         const profileIndex = profiles.findIndex(p => p.name === profileName);
@@ -350,7 +359,7 @@ async function convertToDashInternal(
     codecMP4Paths,
     videoOutputDir,
     profiles,
-    segmentDuration,
+    effectiveSegmentDuration,
     codecsSelected,
     formatsSelected,
     hasAudio

src/core/encoding.ts

@@ -53,6 +53,7 @@ export async function encodeProfileToMP4(
   codecType: 'h264' | 'av1',
   qualitySettings?: CodecQualitySettings,
   optimizations?: VideoOptimizations,
+  muted: boolean = false,
   decoderAccel?: HardwareAccelerator,
   onProgress?: (percent: number) => void
 ): Promise<string> {
@@ -157,12 +158,15 @@ export async function encodeProfileToMP4(
 
   // Build video filter chain
   const filters: string[] = [];
+  const targetWidth = profile.width;
+  const targetHeight = profile.height;
 
-  if (decoderAccel === 'nvenc') {
+  const useCudaScale = decoderAccel === 'nvenc';
-    // CUDA path: keep frames on GPU
+  if (useCudaScale) {
-    filters.push(`scale_cuda=${profile.width}:${profile.height}`);
+    // CUDA path: вписываем в профиль с сохранением исходного AR
+    filters.push(`scale_cuda=${targetWidth}:${targetHeight}:force_original_aspect_ratio=decrease:force_divisible_by=2`);
   } else {
-    filters.push(`scale=${profile.width}:${profile.height}`);
+    filters.push(`scale=${targetWidth}:${targetHeight}:force_original_aspect_ratio=decrease:force_divisible_by=2`);
   }
   
   // Apply optimizations (for future use)
@@ -178,17 +182,34 @@ export async function encodeProfileToMP4(
     }
   }
 
+  // Если использовали GPU-скейл, возвращаем кадры в системную память перед CPU-фильтрами
+  if (useCudaScale) {
+    filters.push('hwdownload', 'format=nv12');
+  }
+
+  // Центрируем кадр, чтобы браузеры (Firefox/videotoolbox) не игнорировали PAR
+  filters.push(`pad=${targetWidth}:${targetHeight}:(ow-iw)/2:(oh-ih)/2`, 'setsar=1');
+
   args.push('-vf', filters.join(','));
 
-  // Audio encoding
+  if (!muted) {
-  // Select optimal bitrate based on source (don't upscale)
+    // Audio encoding с нормализацией таймингов и автоподпаддингом тишиной
     const targetAudioBitrate = parseInt(profile.audioBitrate) || 256;
     const optimalAudioBitrate = selectAudioBitrate(sourceAudioBitrate, targetAudioBitrate);
     args.push('-c:a', 'aac', '-b:a', optimalAudioBitrate);
 
-  // Audio optimizations
+    const targetDur = duration.toFixed(3);
+    const audioFilters: string[] = [
+      'aresample=async=1:min_hard_comp=0.1:first_pts=0',
+      `apad=whole_dur=${targetDur}`,
+      `atrim=0:${targetDur}`
+    ];
     if (optimizations?.audioNormalize) {
-    args.push('-af', 'loudnorm');
+      audioFilters.push('loudnorm');
+    }
+    args.push('-af', audioFilters.join(','));
+  } else {
+    args.push('-an'); // без аудио дорожки
   }
 
   // Output
@@ -217,6 +238,7 @@ export async function encodeProfilesToMP4(
   codecType: 'h264' | 'av1',
   qualitySettings?: CodecQualitySettings,
   optimizations?: VideoOptimizations,
+  muted: boolean = false,
   decoderAccel?: HardwareAccelerator,
   onProgress?: (profileName: string, percent: number) => void
 ): Promise<Map<string, string>> {
@@ -239,6 +261,7 @@ export async function encodeProfilesToMP4(
           codecType,
           qualitySettings,
           optimizations,
+          muted,
           decoderAccel,
           (percent) => {
             if (onProgress) {
@@ -269,6 +292,7 @@ export async function encodeProfilesToMP4(
         codecType,
         qualitySettings,
         optimizations,
+        muted,
         decoderAccel,
         (percent) => {
           if (onProgress) {

src/types/index.ts

@@ -86,6 +86,9 @@ export interface DashConvertOptions {
   /** Предпочитаемый аппаратный ускоритель для декодера (auto по умолчанию) */
   hardwareDecoder?: HardwareAccelerationOption;
 
+  /** Отключить аудиодорожку (muted). По умолчанию false. */
+  muted?: boolean;
+  
   /** Quality settings for video encoding (CQ/CRF values) */
   quality?: QualitySettings;
   

src/utils/video.ts

@@ -90,13 +90,16 @@ export function selectAudioBitrate(
   sourceAudioBitrate: number | undefined,
   targetBitrate: number = 256
 ): string {
+  const MIN_AUDIO_KBPS = 64; // не опускаться ниже базового качества
+
   if (!sourceAudioBitrate) {
     // If we can't detect source bitrate, use target
     return `${targetBitrate}k`;
   }
 
-  // Use minimum of source and target (no upscaling)
+  // Не занижаем слишком низко: clamp к минималке, но не выше целевого
-  const optimalBitrate = Math.min(sourceAudioBitrate, targetBitrate);
+  const clampedSource = Math.max(sourceAudioBitrate, MIN_AUDIO_KBPS);
+  const optimalBitrate = Math.min(clampedSource, targetBitrate);
   
   // Round to common bitrate values for consistency
   if (optimalBitrate <= 64) return '64k';

web-test/index.html

@@ -5,8 +5,8 @@
         <meta name="viewport" content="width=device-width, initial-scale=1.0">
         <title>Document</title>
 
-        <script src="https://cdn.dashjs.org/latest/dash.all.min.js"></script>
+        <script src="dash.all.min.js"></script>
-        <!-- <script src="https://cdnjs.cloudflare.com/polyfill/v2/polyfill.min.js?features=es6,Array.prototype.includes,CustomEvent,Object.entries,Object.values,URL"></script> -->
+        <script src="https://cdnjs.cloudflare.com/polyfill/v2/polyfill.min.js?features=es6,Array.prototype.includes,CustomEvent,Object.entries,Object.values,URL"></script>
         <script src="https://unpkg.com/plyr@3"></script>
         <link rel="stylesheet" href="https://cdn.plyr.io/3.7.8/plyr.css" />
 
@@ -15,6 +15,9 @@
                 margin: 50px auto;
                 max-width: 1500px;
             }
+            video {
+                max-height: 800px;
+            }
         </style>
     </head>
     <body>
@@ -24,7 +27,8 @@
 
         <script>
             document.addEventListener('DOMContentLoaded', () => {
-                const source = 'https://bitmovin-a.akamaihd.net/content/sintel/sintel.mpd';
+                const source = 'http://localhost:3000/test-videotoolbox/manifest.mpd';
+                // const source = 'http://localhost:3000/test-nvenc/manifest.mpd';
                 const dash = dashjs.MediaPlayer().create();
                 const video = document.querySelector('video');
                 dash.initialize(video, source, true);