Export Composer
You can export content from your Project at any time using the Export Composer. It provides access to all the data Align generates as you use it. This includes script data, transcripts, timecode, the audio itself and more. It’s useful for capturing and communicating issues with other team members, delivering complete and approved projects, and everything in between. You can launch it from the Export icon in the Sidebar or Contextual Toolbar, or via its E hotkey.
The Export Composer is made up of five sections, some of which can be expanded to fine tune your export, including or excluding what you need. Additionally, enabling some options may provide enhanced configurations such as transcoding possibilities or MIDI timecode features (used to create script markers in a DAW). Previous selections are remembered each time you launch the Export Composer.
Make your selection and click "Generate Export". Your export will be processed in the background, added to the Background Tasks queue in the top-right of the Project view. Most exports will be generated very quickly, but audio processing such as transcoding will take longer depending on the amount of audio. A notification will appear when your download is ready, which you can click to download the file to your computer. You can also download and re-download the export at any time by clicking on the item in the Background Tasks queue until it’s removed.
You can dismiss the Export Composer by clicking the dismiss "x" in the top right.
Options
Scripts
Your Scripts are all selected to be included in the export by default. Expanding the Scripts section will reveal each Script where you can click each one to include or exclude it. Click "Include All" to toggle all Scripts or deselect all. At least one Script must be selected to complete and generate an export. A count of the currently selected scripts appears alongside the section title.
Your script selection will inform what files are included for every other selection.
Data
The Data section makes available three different exports. All three are selected by default.
Align Data
Select "Align Data" to include your main data export which includes all text content and matched audio information. This includes script, transcript, match status, timecode, and references for every asset including the originally uploaded audio filenames. A separate file will be created for each Script.
There is a single row of data for every script Phrase, duplicating the project and script data so the dataset is self-contained and compatible with any other export. The information provides everything you need for collaborating off-platform, archiving and navigating the exported assets.
Script Data
Select "Script Data" to include a CSV file which contains the original script, its locale, and any originally imported phrase references. The file is in the same format supported for importing scripts into Align – so can be used to create new projects from the same script by yourself or other users. A separate file will be created for each Script.
Transcript Data
Select "Transcript Data" to export the raw transcript Align created as part of the content alignment process. This is in JSON format and not intended for human consumption. Regardless of the underlying transcription process, Align will normalise the format so any downstream process you might engage will be compatible with all exports.
Beta
The Beta section includes a number of options still under development. They are intended to provide useful utility but may not be suitable for all projects. MIDI Timecode and SRT Files are both selected by default.
MIDI Timecode
Select "MIDI Timecode" to generate a .mid MIDI file containing marker information for matched audio. A separate MIDI file is generated for each audio file, as the marker information is relative to the dialogue location in each originally uploaded audio asset.
Each matched Phrase is included with a marker indicating the dialogue’s start position, and the marker label containing the matched script content (not the transcript). To support the limited character set supported by MIDI files, Align transliterates content to roman letters, including CJK languages (Chinese, Japanese & Korean).
These files can be imported into a typical DAW application, such as Pro Tools, Logic or Reaper, to provide positional markers for the dialogue when aligning the original audio to 00:00 on the timeline. As Align has to convert timecode to bars and beats for the MIDI file, it is important to set the DAW project tempo to 120bpm to match the file.
With MIDI Timecode selected, two sub-options appear:
Include [END] markers in MIDI Export. – This adds additional [END] markers for the out-time for each dialogue instance so both start and end markers are included in the file. Disabled by default.
Use transliterated content as MIDI markers. – This ensures non-roman characters within script content don’t cause any issues generating and using the files. It is enabled by default but can be disabled for more accurate content inclusion. However this may only work reliably for English language content.
SRT Files
Select "SRT Files" to include SRT captions with your export. These files are in plain text, human-readable and support multi-lingual content. A separate file is created for each audio file, as the timings are relative to the dialogue position in each originally uploaded audio asset.
While Align does not attempt to conform the content to any standard, it can be a great way to get the initial timings into another caption authoring tool to fine-tune and master there. Line length will exceed typically accepted limits for longer phrases, with each Phrase mapping to one caption regardless of length.
Transcode & Segment Audio
Enable "Transcode & Segment Audio" to add audio processing to your export. While your original files are preserved, the export will include discrete audio files for each matched phrase in your project. Each audio will be clipped at the exact in and out time of the current alignment, and then transcoded into one of three formats. You can also add optional padding to the start and end of the segment – up to 5s using the provided slider.
Transcode Format
You can choose from three audio formats for the segmented audio. All options produce single channel, mono audio.
- WAV PCM HQ (24/48) – High quality 24-bit WAV audio file with 48KHz sampling rate (lossless).
- WAV PCM LQ (16/48) – Standard quality 16-bit WAV audio file with 48KHz sampling rate (lossless, default).
- AAC-LC (96Kbps) – Low quality 96Kbps AAC audio using a low-complexity profile, effective for speech (compressed).
Segment Padding
By default segments will be clipped at exactly the timecode Align identified via content alignment, which is typically very tight on the start and end of dialogue. If you desire a little safety for the segments, or more "air", you can add up to 5s of audio padding to each side of the segment in 0.1s increments. Importantly the padding is treated as an "up to" value as there are a few cases where the value will be reduced:
- If there is not enough audio in the original file, i.e. it exceeds the start or end of the actual file.
- Intelligent collision detection is used to avoid absorbing dialogue from other known dialogue segments.
Transcode & Segment is not selected by default as it is a billable request, in addition to content alignment, and increases the time required to generate the export dramatically, depending on the amount of processing. Work is done in parallel, so for larger workloads, it can perform in excess of 15x realtime.
Original Audio
Select "Original Audio" to include the originally uploaded audio files with your downloads. This is not enabled by default as it is assumed you may have them available already, but including them will conform their naming with other assets in the self-contained payload. While the files may be renamed, the audio within is delivered exactly as uploaded with no additional processing. Including them will also not significantly delay the export generation, but the download itself will of course be bigger.
Understanding the Export
All exports are delivered as a single compressed archive (.zip). Any modern operating system should have no issues expanding the contents and presenting the resulting export as a single folder.
All files in your export are conformed to a pattern, with consistent labelling and references, which are also used in the Align Data file. This should make drilling down to required elements straightforward, and easily comparable with subsequent exports.
Folder Hierarchy
All file and folder names follow a consistent pattern based on the Project Key, Script number, locale, and asset or phrase number. Each export will only contain the files relevant to both the Project and your Export Composer selection.
- 📁{project-key}/Export root, named after Project Key.
- 📁{project-key}_{script}_{locale}/One folder per Script, indexed with locale.
- 📄…_Align.csvAlign Data export.
- 📄…_Script.csvScript Data export.
- 📁…_{asset}/One folder per audio asset, labelled with Asset reference.
- 📄…_{asset}.wavThe Original Audio file.
- 📄…_{asset}_Transcript.jsonRaw transcript data (JSON).
- 📄…_{asset}_Subtitles.srtSRT caption file.
- 📄…_{asset}_Markers_120bpm.midMIDI marker file.
- 📁…_Segments/Transcoded phrase audio.
- 📄…_{phrase}.wavOne audio file per Phrase.
For example, in a project with key 00123-XYZ, the first script in British English with one audio asset would produce:
| Token | Example | Meaning |
|---|---|---|
{project-key} | 00123-XYZ | Project key as shown in the dashboard |
{script} | S001 | Script reference, indexed in order of creation |
{locale} | en-GB | Language tag for the script content |
{asset} | A001 | Audio asset number, indexed in order of upload |
{phrase} | P001 | Phrase number, matching the script line |
The naming is consistent across the folder structure, and referenced within the Align Data CSV — making it straightforward to cross-reference any asset with its corresponding data row.