Music Metadata Editor: Naming Stems and Alts

A music metadata editor workflow for stems, alts and edits needs to do more than add “instrumental” to a filename. Musicians, producers and libraries need each file to be identifiable on its own and connected to the right parent track. That makes it easier to deliver the right version, answer a sync request and search a catalog without opening files one by one.

The reliable approach is to separate two jobs: name and record the version relationship yourself, then analyze each audio file for the qualities that are actually present in it. AI music tagging can help describe sound, but it cannot replace a clear version naming convention or catalog structure.

Stems, alts and edits are different assets

A stem is an audio part or grouped submix, such as drums, bass or vocals. An alt is a complete alternate mix or arrangement of the same track, such as an instrumental, TV mix or no-drums mix. An edit is usually a version with a changed structure or duration, such as a 60-second cutdown. These distinctions help people know what they are receiving before they audition it. (soundstripe.com)

Use the category that describes the file, not a vague catch-all like “alt.” For example:

• `Northbound (Instrumental)` is a full-length alternate mix without lead vocals.
• `Northbound (TV Mix)` identifies a mix where the lead vocal is removed or reduced, according to your delivery definition.
• `Northbound (60s)` is a duration-specific edit.
• `Northbound STEM Drums` identifies a part, not a complete alternate mix.

A stem is not necessarily a solo instrument recording. In a production-music catalog it often means a grouped, mixed audio part. Agree on what your organization means by “stem,” then use that definition consistently in filenames and catalog fields.

Set one naming pattern and apply it to every delivery

Choose a stable title as the shared base, followed by a short, controlled version label. A useful pattern is:

`[Track Title] ([Version Type]) ([Duration, if relevant])`

For stems, keep the stem role visible and preserve the common title:

• `Northbound (Full Mix).wav`
• `Northbound (Instrumental).wav`
• `Northbound (TV Mix).wav`
• `Northbound (60s).wav`
• `Northbound STEM Drums.wav`
• `Northbound STEM Bass.wav`
• `Northbound STEM Vocals.wav`

You could instead use underscores or another separator. The separator matters less than using the same terms and order across the catalog. Music delivery recommendations also describe structured filename components for mixes and stems, which supports treating the title and version identifier as distinct information. (aes.org)

Set a few rules before renaming a backlog:

• Keep the base title identical across versions. Avoid one file called `Northbound` and another called `Northbound instrumental final`.
• Pick one spelling for each version type. For example, use `Instrumental` everywhere, not a mix of `Inst`, `Instrumental`, and `No Vox` unless those labels mean different deliverables in your catalog.
• Put duration in a consistent format, such as `60s`, `30s`, and `15s`.
• Give revisions a separate, controlled revision field or suffix if you need to distinguish delivery rounds. Do not let `final`, `final2` and `really-final` become permanent version categories.
• Keep filenames readable and avoid adding descriptive claims that are not useful for identifying the asset.

A naming convention is not a substitute for explicit version relationships in your catalog. A title suffix helps a person scan a folder. A parent-track or family ID lets a database reliably connect the full mix, cutdowns, alts and stems, even if somebody later changes a filename.

Record identity separately from version details

A useful music metadata manager should preserve the difference between track identity and file-specific version information. Your catalog can use a compact set of fields such as:

• Base title: the common work or track name, such as `Northbound`.
• Version type: full mix, instrumental, TV mix, clean, underscore, cutdown or stem.
• Duration label: for example, `60s` or `30s` for timed edits.
• Stem part: drums, bass, vocals, guitars or another agreed grouping, when the file is a stem.
• Parent or family ID: the stable value connecting versions that belong together.
• Revision: a delivery revision, if your workflow needs one.
• Audio filename: the exact name of the file being delivered.

Do not use one field to carry several meanings. `Northbound 60 clean inst final` is hard to filter because it bundles title, duration, lyric status, mix type and revision into a single string. Keep those details in separate fields when your catalog supports them, while retaining a readable filename for people who browse folders.

The musical metadata should also describe the individual audio file, not just copy the full mix's tags. An instrumental may retain the track's tempo and harmony but have no lead vocal. A drums stem may have a different perceived energy or mood from the full mix. A short edit may begin in a different section and communicate a different use from the full version. The parent relationship can be shared; the sound description should be checked per file.

Analyze every version as its own audio file

A useful workflow is to name the family first, then analyze every deliverable that a buyer, editor or platform might search for. TrackTag Studio batch audio analyzer accepts multiple audio files and returns fields including BPM, key, genres, moods, emotions, themes, occasions, instruments, vocals, song structure, production and mixing notes, and a full description.

TrackTag Studio offers Core and Ultra analysis. Both levels use the same engine on the same audio and are equally accurate. Core costs 1 credit per track and returns 9 fields, including keyword tags. Ultra costs 2 credits and returns all 35 fields, including the written description. A Core result can later be re-analyzed at Ultra for the 1-credit difference. Choose the level based on how much metadata you need, not on a belief that one level analyzes the audio better.

For a focused catalog update, use Core where searchable keywords and core filing fields are enough. Use Ultra when the team needs the wider set of fields or a written description for reviewing and pitching a particular version. BPM is measured from the audio signal with TrackTag's Beat This! beat-tracking model, and tracks without a steady beat receive “No BPM” instead of an invented tempo. Key is measured by TrackTag's signal-analysis engine.

Analyze the full mix, each alt and each stem as separate files. Then review whether the returned tags make sense for that version. Do not simply copy the full-mix description onto a vocal stem or assume a short edit has the same structural label as the long version. TrackTag analyzes audio and exports results; use your own catalog or delivery system to preserve parent IDs and version relationships.

You can tag a batch of versions with TrackTag Studio, then export a whole batch, a hand-picked subset or a file per track in CSV, Excel, XML, JSON, Markdown, PDF or TXT. Keep a stable version field in your catalog and map the relevant analysis to the corresponding audio filename. If you need an overview of the wider catalog process, see the guide to batch tagging music files.

A version-family example for a sync library

Suppose a composer delivers a full mix, instrumental, TV mix, 60-second edit, 30-second edit and four stems for `Northbound`. Start with one family ID, such as `NB042`, and a consistent set of filenames:

• `Northbound (Full Mix).wav`
• `Northbound (Instrumental).wav`
• `Northbound (TV Mix).wav`
• `Northbound (60s).wav`
• `Northbound (30s).wav`
• `Northbound STEM Drums.wav`
• `Northbound STEM Bass.wav`
• `Northbound STEM Music.wav`
• `Northbound STEM Vocals.wav`

In your catalog, record `NB042` as the shared family value and set a separate version type for each file. Set duration on the two cutdowns, and stem part on the four stem files. Analyze all nine files if they are individually searchable or licensable. Review the vocal and instrument fields carefully: “instrumental” on the main instrumental is not interchangeable with “vocals” on a vocal stem.

For a sync handoff, this structure makes two questions easy to answer: “What version is this?” and “Which versions belong to this track?” Your delivery notes can point to the family, while the filename and metadata identify each file. For more on organizing a submission-ready collection, use the guide to preparing a catalog for a music library submission.

If the same process needs to run whenever new uploads arrive, TrackTag also offers a public API that returns analysis as JSON and a Zapier app with an “Analysis Finished” trigger, an “Analyze Track” action and a “Get Analysis” lookup. Those can connect analysis to a catalog workflow, but your system still needs to store the version type and parent relationship. See the TrackTag API documentation for the available integration details.

Keep the system useful as the catalog grows

Before applying a convention to thousands of files, test it on a small family with a full mix, two alts, a cutdown and several stems. Ask someone who did not create the files to identify each one using only its name and catalog fields. If they cannot tell an instrumental from a TV mix, revise the labels before a large rename.

Then check three things in a sample export:

• Every file has one clear version type, not a pile of conflicting suffixes.
• Every related file points to the same family or parent value.
• Audio tags describe the file that was analyzed, not an assumed description inherited from the full mix.

The goal is not to make every filename carry the whole catalog record. It is to give people enough information to recognize the file, while using structured metadata for filtering and the parent relationship. That division keeps folders legible and makes your music metadata database more dependable for delivery, discovery and licensing.

FAQ

How should I name stems and alternate mixes?

Use the same base title for the track family, then add a consistent version label such as `Instrumental`, `TV Mix`, `60s` or `STEM Drums`. Choose one separator and naming style, and record the parent relationship separately in your catalog.

Should every stem have its own music metadata?

Yes, if people need to search, deliver or license each stem as a separate file. Analyze and review each audio file on its own, because its instruments, vocals, mood and description may differ from the complete mix.

Can AI music tagging identify which files are related versions?

Audio analysis can describe the sound in each file, but do not rely on it to establish your catalog's parent and version relationships. Use a stable family ID and version-type field for that relationship, then use AI analysis to help describe each recording.

Does TrackTag Studio rename files or connect alts to a parent track?

TrackTag Studio analyzes audio and lets you export the results in formats including CSV, XML and JSON. Keep filenames, parent IDs and version types in your catalog or delivery workflow, and map the analysis to each file there.