# Output Presets | AudioDN Docs

> Configure AudioDN output presets for transcoding, previews, waveforms, transcriptions, and analysis.

Source: https://audiodeliverynetwork.com/de/docs/output-presets/

---

# Output Presets

An **output preset** is a recipe for what to make from an upload—a transcode, preview, waveform, analysis, and more. ADN runs the presets you’ve configured against every upload and stores each result as an **output** on that track. You can also add a track-scoped preset after the fact when a single track needs an extra output. (Output presets were previously called variants.)

[Transcoded Version](#transcoded-version)[Preview Clip](#preview-clip)[Audio Analysis](#audio-analysis)[Waveform Video](#waveform-video)[Waveform Image](#waveform-image)[Transcription](#transcription)

## Choosing output types

Output presets transform a single upload into multiple deliverables without extra work on your side. Common patterns:

| Use case | What you get | Output type |
| --- | --- | --- |
| **Consistent playback formats** | Users upload WAV, FLAC, MP3, etc. — presets normalize everything to a single codec and bitrate so your player doesn’t have to care. | Transcoded Version (`transcode`) |
| **Low-bandwidth streaming** | Serve a 64 kbps Opus output for mobile listeners and a 320 kbps MP3 for desktop — the player switches automatically. | Transcoded Version (`transcode`) |
| **Try-before-you-buy previews** | Auto-generate a 30-second clip so free users can sample tracks while full-length stays behind a paywall. | Preview Clip (`preview`) |
| **Player waveforms** | A default 320-sample waveform ships with every track for free. Need more detail? Add an Audio Analysis preset with higher sample counts. | Audio Analysis (`levels`) |
| **Social media clips** | Generate animated waveform videos with embedded audio — ready to share on Instagram, TikTok, or X without any video editing. | Waveform Video (`waveform_video`) |
| **Custom player artwork** | Produce static waveform PNGs with transparent backgrounds for use as scrub-bar overlays or decorative artwork. | Waveform Image (`waveform_image`) |
| **Multiple download formats** | Offer MP3, FLAC, and WAV downloads from a single upload — no server-side encoding needed at request time. | Transcoded Version (`transcode`) |
| **Voice-optimized encoding** | Use Opus at low bitrates (24–48 kbps) for podcasts, audiobooks, or voice content — Opus delivers near-transparent speech quality at a fraction of the file size compared to MP3. | Transcoded Version (`transcode`) |
| **Cost-effective storage** | Keep the original for re-processing later but serve compressed outputs to save bandwidth and storage costs. | [Original Upload](#original-upload) (retention) |
| **Searchable speech and captions** | Turn spoken audio into text on the track and a timed caption file, so voice can be searched or shown alongside playback. | [Transcription](#transcription) (`transcribe`) |

### Transcoded Version

Convert your audio file into another format with customizable quality and settings.

Output type `transcode`. Converts uploads into playback-ready formats. Create multiple presets for different quality tiers, codecs, or download options.

| Setting | Description |
| --- | --- |
| `Codec Type` | Output format — MP3, AAC, Opus, FLAC, WAV, etc. |
| `Bitrate` | Target quality and file size (e.g. 128 kbps, 320 kbps) |
| `Stereo / Mono` | Downmix to mono for smaller files or keep stereo for full fidelity |
| `Strip Metadata` | Remove embedded tags (artist, album, etc.) if not needed |

### Preview Clip

Generate short, clipped previews automatically based on your rules — perfect for 'try before you buy' scenarios.

Output type `preview`. Automatically clips a portion of the track for previews. Control the duration, start position, and minimum/maximum percentages.

| Setting | Description |
| --- | --- |
| `Preview Length Type` | Fixed duration in seconds or a percentage of the full track |
| `Preview Length` | Exact duration for the preview (e.g. 30 seconds) |
| `Minimum Percentage` | Ensure a minimum portion of the track is always included |
| `Maximum Percentage` | Cap how much of the track can be previewed |
| `Start Position` | Where the preview begins within the original track |
| `Transcode Options` | All settings from Transcoded Version (codec, bitrate, channels, metadata) apply to the preview clip too |

### Audio Analysis

Generates detailed track insights for powering waveforms, visualizations, and audio metrics.

Output type `levels`. A default 320-sample waveform is generated for every track at no extra cost. Add an Audio Analysis preset only if you need higher sample counts or different metrics.

| Setting | Description |
| --- | --- |
| `Number of Samples` | Resolution of the waveform — more samples means more visual detail |
| `Analysis Metric` | What is measured per sample (e.g. RMS, peak, loudness) |

### Waveform Video

Creates an animated waveform video with customizable styling and optional embedded audio.

Output type `waveform_video`. Produces an MP4 with an animated waveform — ready to share on social media platforms without any video editing.

| Setting | Description |
| --- | --- |
| `Video Width` | Output width in pixels |
| `Video Height` | Output height in pixels |
| `Video Frame Rate` | Frames per second — controls smoothness of the animation |
| `Background Color` | Fill color behind the waveform |
| `Waveform Line Color` | Color of the animated waveform line |
| `Include Audio` | Whether to embed the track’s audio in the video file |
| `Transcode Options` | When audio is included, all Transcoded Version settings (codec, bitrate, etc.) apply to the embedded audio |

### Waveform Image

Creates a static waveform graphic — great for scrubbing bars or artwork overlays.

Output type `waveform_image`. Produces a PNG with a transparent background. Use it as a scrub-bar overlay, decorative artwork, or embed in your own UI.

| Setting | Description |
| --- | --- |
| `Image Width` | Pixel width of the output image |
| `Image Height` | Pixel height of the output image |
| `Waveform Line Color` | Color of the waveform trace |

### Original Upload

Preserves the exact file that was uploaded, before any processing.

Original Upload is a retention setting (type id `original`), not a normal transcode preset. Enable or disable it in the dashboard. Keeping the original lets you generate new outputs later. Without it, new outputs cannot be generated from that track. Delete originals only after your preset set is finalized if you need to reclaim storage.

### Cover Image

Extracts an embedded cover image from the audio file and generates image colors automatically.

All uploaded files are scanned for embedded cover art. When found (or uploaded separately), ADN extracts a color palette and auto-selects a `player_color` for theming, plus contrast-adjusted `player_color_light` and `player_color_dark` colors for light and dark backgrounds.

The `theme` array in the track response contains one or more extracted colors with hex values, area, lightness, and saturation data — use these for custom UI theming.

You can override cover images and image colors at the organization or collection level.

### Transcription

An output preset that writes spoken text onto the track and stores a timed caption file as its output.

Add a transcription output preset when spoken audio should be searchable or captioned. It is optional and uses AI Credits. Standard audio presets do not.

When that preset runs, processing does both of the following:

-   **Track record.** Writes `transcription` (plain-text paragraphs of what was said) and `transcription_language` (a BCP-47 language code such as `en`) onto the track. Both stay `null` until this preset has been applied. Language is detected automatically unless the preset sets one. These fields are read-only — see the [Tracks API](/docs/api/tracks).
    
-   **Output.** Stores a WebVTT (`.vtt`) caption file as the preset’s output, with cue timing for captions or a custom transcript UI. That file is a normal track output. The [Output webhook](/docs/webhooks/outputs) reports it when the file is ready. The track fields above are on the track, not on that per-file event.
    

Language can stay on auto-detect or be pinned. Speaker labels are off unless you turn them on. With speaker labels, the track text is grouped by speaker and each caption cue is tagged. Transcription is billed from AI Credits by audio duration after it succeeds. Playback does not wait for it unless you mark the preset required.

Pricing for this preset is in [AI Credits](/pricing). For recording, retrieval, and WebVTT access, see the [Voice & transcription](/docs/guides/voice) guide. A worked example is also on [Handling Voice](/examples/handling-voice#transcribe).