Skip to content

Editor & Export

Jump to bottom
Tarmo Saranen edited this page May 6, 2026 · 2 revisions

Editor & Export

The Editor tab provides a chunk-by-chunk view of your audiobook with tools for editing, regeneration, preview, and export.

Chunk Editor

Each chunk shows:

  • Status — Pending (no audio), Done (generated), or Error
  • Speaker — Character name (editable)
  • Text — Spoken dialogue or narration (editable)
  • Instruct — TTS voice direction (editable)

Editing Workflow

  1. Click any field to edit it inline
  2. Changes are saved automatically
  3. After editing, click Generate on that chunk to regenerate its audio
  4. Or use Batch Render Pending to regenerate all chunks that need audio

Render Modes

Render Pending (Standard)

Sends individual TTS calls in parallel using the configured worker count.

  • Per-speaker seeds for reproducible output
  • Works with all voice types

Batch (Fast)

High-speed batched rendering. See Batch Generation for details.

  • 3-6x real-time throughput
  • Chunks grouped by voice type and sub-batched by text length
  • Single batch seed (configure in Setup)

Single Chunk Generation

Click the generate button on any individual chunk to regenerate just that line. Useful for:

  • Fixing a single line after editing text or instruct
  • Testing different instruct directions
  • A/B comparing voice options

Audio Preview

  • Play individual chunks by clicking the play button
  • Play Sequence — Plays all chunks in order with natural pauses:
    • Configurable pause between different speakers (default 500ms)
    • Configurable pause between consecutive same-speaker segments (default 250ms)
    • Both values can be adjusted in the Setup tab under Speaker Change Pause and Same Speaker Pause

Merging

Click Merge All to combine all generated chunks into a single audiobook file with natural pauses. The result appears in the Result tab.

Requirements:

  • All chunks must have generated audio (status: Done)
  • Chunks are merged in order with automatic pauses

Export Options

Combined Audiobook

A single MP3 file with all voices and natural pauses.

  • Downloaded from the Result tab
  • File: cloned_audiobook.mp3

Individual Voicelines

Separate MP3 files per line, stored in the voicelines/ directory:

voicelines/
├── voiceline_0001_narrator.mp3
├── voiceline_0002_elena.mp3
├── voiceline_0003_marcus.mp3
└── ...

Files are numbered in timeline order with speaker names for easy import into DAWs.

Audacity Export

One-click export of per-speaker WAV tracks for multi-track editing.

Click Export to Audacity in the Result tab to generate:

audacity_export.zip
├── project.lof       # Open in Audacity to auto-import all tracks
├── labels.txt        # Import via File > Import > Labels for annotations
├── narrator.wav      # Full-length track with only NARRATOR audio
├── elena.wav         # Full-length track with only ELENA audio
├── marcus.wav        # Full-length track with only MARCUS audio
└── ...

How to use:

  1. Download and unzip audacity_export.zip
  2. Open project.lof in Audacity — all speaker tracks import automatically
  3. Import labels.txt via File > Import > Labels for chunk annotations
  4. Each track is padded to the same total duration with silence — playing all tracks simultaneously sounds identical to the merged MP3

This format is ideal for:

  • Per-character volume adjustment
  • Adding effects to specific speakers
  • Fine-tuning timing between dialogue
  • Professional post-production

MP3 Encoding Note

If you get broken or tiny (428 byte) MP3 files, conda's bundled ffmpeg may lack the MP3 encoder. Fix:

conda install -c conda-forge ffmpeg
# or remove conda's ffmpeg to use system one:
conda remove ffmpeg

Alexandria auto-detects this and falls back to WAV output.

Clone this wiki locally