Audio/Video Conversion

The platform automatically converts uploaded media files into web-friendly formats so they play smoothly across browsers and devices. The pipeline runs FFmpeg under the hood, produces multiple renditions, and stores them alongside the original.

Why Conversion

• Browser compatibility: Not every browser plays every codec. H.264 inside MP4 is the safe lowest-common-denominator.
• Bandwidth: Multiple renditions (480p / 720p / 1080p) so clients adapt to network conditions.
• Mobile: Smaller files for cellular connections.
• Audio extraction: Pull audio-only track for podcast-style playback.
• Poster image: A still-frame thumbnail used as the player's preview.

Format Choices

• Video: H.264 in MP4 (universally supported) plus optionally VP9 in WebM for size savings.
• Audio: AAC in M4A (universal) plus optionally Opus in WebM.
• Streaming: Optional HLS or DASH segmented manifest for adaptive streaming.

Processing Workflow

1. Upload arrives, original stored.
2. A conversion task is queued.
3. Worker picks up, probes the file, decides which renditions to generate.
4. Each rendition produced sequentially or in parallel depending on worker config.
5. On completion the record updates with paths to each rendition.
6. Optionally the uploader is notified by email/push.

Worked Examples

• Tutorial videos: Generate 480p / 720p / 1080p MP4, plus poster JPG. Player picks based on bandwidth.
• Podcast platform: Audio-only — re-encode to AAC 128kbps, generate Opus 96kbps for size-conscious browsers.
• Live event highlights: Original 4K source → 1080p / 720p / 480p / 240p; HLS manifest for live-event playback.
• Internal training: Single 720p rendition; no public access; transcript extracted for indexing.

Some uploaded videos have no audio — screen recordings, time-lapses, security camera clips. The default conversion pipeline expects an audio track and can fail or produce muted output. This article explains how the platform handles these and how to tune the behaviour.

Detection

The probe step uses FFmpeg's stream introspection to detect whether the file contains an audio track. Detection is reliable for all standard container formats.

Behaviour Options

• Auto-detect and mark silent (default): Conversion runs as normal; the resulting metadata marks the file as silent so player UIs can hide the volume control.
• Insert silent audio track: Adds a silent AAC track so the file plays through audio-aware pipelines that require one.
• Reject upload: Refuse silent videos at upload — useful when audio is genuinely required (e.g. an interview platform).
• Auto-mute and proceed: Old behaviour; deprecated.

Why Inserting Silent Tracks Matters

Some downstream systems (broadcast servers, video editors, captioning workflows) expect every video file to have an audio track even if empty. Inserting a silent track keeps those pipelines happy without changing the user's perception.

Worked Examples

• Screen recording library: Default behaviour — silent flag respected, player hides the volume control.
• Broadcast pipeline: Insert silent track so downstream broadcast schedulers don't crash on missing audio.
• Interview platform: Reject silent uploads — audio is the whole point.
• Security camera feed: Default — silent flag set; alerts use motion not audio.