The Bootstrap Guide to Music Metadata

This month Marina and I put together a guide for libraries getting their start in DIY music publishing.

We cover the basics of good library metadata practice, including how to tag your tracks, prep your files for delivery, and deliver to a client, distributor or subpublisher.

For more information on selecting a good publishing platform, see our previous interviews with SourceAudio, Harvest, DISCO and CadenzaBox.

 

Metadata is to music search as buttons are to an elevator.

 

In that, maybe the only time you’ve noticed them is when they weren’t working.

 

For a music editor looking for the perfect track to underscore their next commercial, film or prime-time show, searching for one type of track and getting another would be akin to your elevator letting you off on the wrong floor. A music library with no metadata at all is like asking your clients take the stairs.

 

About This Guide

We at TagTeam have spent the last decade tackling the problem of how best to deploy descriptive music metadata in the highly subjective, highly competitive world of production music. Our expertise isn’t authoritative (though we’d like to think we’ve been around the block a few times), but our work with many of the top libraries in the industry, big and small, teaches us more every day.

 

This guide is meant as a starting point for those finding their way in DIY music publishing, and will break down into the following sections:

 

  1. TAGGING: What tags to use, how best to tag, and how to start formulating your own taxonomy.
  2. PREP: How to format your metadata, encode files for delivery, and organize your local library.
  3. DELIVERY: How to QA your completed metadata and work with your distributors to make changes as needed.

 

Laced up? Let’s begin!

 

PART 1: TAGGING

 

At the heart of descriptive metadata is the process of tagging, or assigning search terms to compositions in your library that describe their fundamental musical qualities.

 

If you’re trying to get good search terms associated with your music, you are pretty much limited to the following options:

  1.  Tag your music by hand, using software like Excel, iTunes or Soundminer to record metadata and/or write to audio files directly (more on that in the next section),
  2.  Rely on the automated and/or machine learning options offered by some distributors (ala DISCO’s Auto Tagging, or proTunes SonicSearch), or
  3.  Hire someone with expertise to listen to your music and tag it for you (like one particular company with the word “Tag” in their name)

 

While we at TTA extol the virtues of carefully selected human-listening-based tags, there are cases where machine-based solutions make sense, like when dealing with massive back-catalogs of material where precise keywording is not of the utmost importance.

 

Machine Learning is a fascinating emergent field of research which shows immense promise for the automation of repetitive human tasks. But to our eyes and ears, and especially in the highly subjective and rapidly evolving sphere of music and pop culture, it’s still hard to beat a trained human ear when it comes to optimizing music tags for search.

 

Building a Taxonomy

Another advantage of hand-tagged tracks is the level of control you get over your search terms. It should go without saying, music is not a one-size-fits-all product; in the same way, having control over the level of nuance in your tags may be important to help sell what makes your library unique.

 

For example, if your library is composed primarily of trailer music, you may choose to include a wider variety of words in the “Music For” or “Industry Usage” fields to better distinguish between tracks that might otherwise just be marked as “Trailer”. These could be tags like “Action“, “Epic Battle“, “Superhero“, or “High Fantasy“.

 

Or, if your library features a carefully curated selection of singer-songwriter ballads, taking time to build out a full-fledged set of Mood tags to categorize your tracks effectively may help highlight the depth that would otherwise be glossed over by simple tags like “Happy” or “Sad”. (Instead, consider the expanded palette of “elated”, “excited”, “optimistic”, “feel-good”, or “melancholy”, “depressed”, “gloomy“.)

 

All these careful considerations are part of the process of creating your own taxonomy, the final collection of words that will come to define your music.

 

Important Note: Some distributors and sub-publishers have restrictions on the types of tags allowed in their search engines. Be sure to check the technical specs on any platform you will be delivering to *before* you tag, and reach out to tech support if the information is not readily available.

 

Refining Your Taxonomy

All libraries with descriptive metadata have a taxonomy whether they know it or not, be it something they mindfully constructed and applied while tagging their tracks, or just the summary collection of whatever tags happened to end up on their MP3s. Ultimately though, clarifying and constructing a taxonomy that you understand and have control over gives you the power to shape how your music can be found.

 

For example, one of the immediate benefits of using a standardized taxonomy is to remove clutter. Take for example the following synonymous tags:
  • “guitar”
  • “Guitar”
  • “guitars”**

– or even –

  • “giutar”

 

A well-modeled taxonomy would unify these keywords under the single tag “Guitar”, linking related tracks featuring guitar together in search, as opposed to keeping them isolated and separate.

 

Search engines on some platforms are more generous than others when it comes to grouping up similar or synonymous terms, but the discrepancies in a poorly applied taxonomy become immediately apparent when your metadata is used to feed any kind of UI, like tag buttons in Source Audio, dropdowns in Harvest, or other software interfaces designed for spotting or DAW integration. The sea of misspellings and minute differences between tags will contribute to a splintering of your taxonomy, with tracks that should belong together winding up separated from one another in search, for no other reason than an inconsistent application of taxonomy.

 

** If you’re wondering about the plural instrument “Guitars” vs. the singular “Guitar”, ask yourself this: will the person searching for this track really want to distinguish between tracks with multiple vs. singular guitars? It’s perfectly fine if so, but can you also then guarantee that this distinction has been made consistently across all tracks in your library? If not, does the extra tag detract from clarity in your search schema?

 

These are all important decisions you get to make as someone constructing your own taxonomy, and this level of thought should be applied to descriptive metadata of all types. A good taxonomy is one that covers the full breadth of your musical catalog with sufficient nuance in terms, but one that is also clear enough in its definitions to be applied to all tracks logically and consistently. A clear conception of how terms in your taxonomy work together will ultimately yield better tags, and thus better search.

 

 

Tagging Best Practices

A second advantage to building your own taxonomy will come when you begin the actual tagging of music.

 

Familiarizing yourself with the words in your library will help you tag consistently and to the full breadth of nuance your taxonomy allows. And if more than one person on your team is tagging your library, a shared understanding of your taxonomy is invaluable in making sure all tracks work together in search. Have frequent discussions over how you use certain tags, and spot check each others’ work to identify potential discrepancies in your choice of tags.

 

Depending on the size of your library, it may also make sense to invest in software that can help make tagging more efficient. TagTeam offers subscription licensing to TuneTagger, the same software we use to tag our tracks, which comes pre-loaded with our customizable TTA Taxonomy; we’ve also heard of clients using FileMaker Pro, Airtable or even SalesForce to help keep tagging consistent. Whatever it is, be sure it is easy to use, allows for customization, and ultimately makes your life easier.

 

 

 

PART 2: PREPARING FOR DELIVERY

 

Filenames, encoding types, and spreadsheets, Oh My.

 

When prepping your files and metadata for delivery, there are a wide variety of formats and expectations to be considered. A helpful starting point is to look at your current distro situation, as well as your end goals:

 

a. Are you on personal terms with music editors whose creative briefs you will be responding to directly?

In this case, you may already be in the habit of assembling files for direct delivery online, for which you may benefit from embedding your metadata directly into your AIFs and MP3s. (WAVs, while sonically superior, hold far less metadata in their wrappers than the formats above.)

 

The advantage of embedded metadata, of course, is that wherever the file goes, the metadata goes, be that into a music editor’s iTunes playlist of “good tracks”, the inbox of another editor in the industry, or the final list of tracks selected for use and in need of licensing information (be sure to embed that too!). The last thing you want is your track to end up perfect for the part but rejected because no one can remember where it came from, or who to contact for licensing.

 

Alternatively, if you are planning on using online platforms to send clients files, be sure to explore the wide variety of playlisting and custom user configurations available via many of the major distributors. Consider whether it makes sense to respond to jobs with your entire library, or with subsets accessible via custom logins or private links. And of course, make sure that all files and metadata you present publicly are presentable, contributing to your overall image and professionalism (more on that later!).

 

 

If you’re not planning on sending files directly, or you are opting for a more general distribution and search solution, you may fall into the next category:

 

b. Are you looking to publish your tracks via an online distributor (such as SourceAudio, Harvest Media, DISCO or CadenzaBox)?

 

While by far the easiest way for newcomers to get their libraries accessible to new clients, these and other distributors like them offer many advantages to seasoned veterans as well, like link-based sharing of playlists and subsections of libraries, automatic one-stop licensing and file delivery (often with automatically embedded metadata), automated sub-publisher distribution, advanced listening analytics, and professional front-end templates that make your library highly presentable. Some sites have even begun rolling out partnership opportunities for select libraries that place your tracks on featured collections targeted at music supes and other players in the industry.

 

Each of the distributors listed above has their own guidelines and formats for metadata ingestion, but most will require at minimum a spreadsheet (in either Excel or CSV format) with some or all of the following information:

 

  • Filename
  • Track Name
  • Artist Name
  • Full Composer Info (including First and Last name, PRO name, IPI, Percentage Share/Split, Region of Collection)
  • Full Publishing Info (Full name of Publishing Entity/Entities, Publishing PRO Name, Publisher IPI, Percentage Share/Split, Region of Collection)
  • Moods (Descriptive Metadata)
  • Keywords/Styles/Music For (Descriptive Metadata)
  • Track Description (Descriptive Metadata)
  • Instrumentation (Descriptive Metadata)
  • Genre/Subgenre (Descriptive Metadata)
  • Release Information:
        – Album Name
        – Label Name
        – Album Descriptive Metadata
        – Catalog Number
        – UPC
  • ISRC
… and more!

 

If you’re just starting out, you may not have all the information above assembled for every track in your library. That’s fine! But this larger profile of information about each of your tracks helps get your music found, licensed, and paid out, so working to assemble a definitive database of full track metadata for your own records is highly recommended as your collection grows and opportunities arise.

 

On Databases: Many composers and self-published libraries just starting out will rely on iTunes and ID3 tags to keep all associated metadata linked to tracks; we highly recommend looking into software like Excel or SoundMiner to manage your data, which in the long run are much more reliable ways of storing, viewing, editing and delivering metadata. It is also much easier to assure quality and detect errors in metadata that is all stored in one place, linked via filename to the final masters.

 

Many libraries using distributors like Harvest or SourceAudio can also take advantage of the online metadata editing tools available via the admin panels; this is a great way to keep your online metadata definitive and up to date, but be sure to download periodic backups in the event of technical errors, server outages, or misplaced edits.

 

c. Are you delivering your tracks to sub-publishers or aggregators who will be hosting and/or licensing your music for you?

 

Every sub-publisher relationship is unique, though in general everything mentioned above relating to online distributor delivery will still apply. Many sub-publishers are accustomed to receiving metadata in Harvest Media format, which includes all the fields mentioned above and more, so if you are just starting your database you can download a Harvest template here to make future deliveries easier.

 

Working with sub-publishers also makes it all the more important to have accurate composer information and splits on file for all of your tracks. Be sure to reach out to anyone in your roster with missing IPI, percentage share, or PRO information to ensure that whoever down the line handling royalty reporting has everything they need to get your artists paid.

 

 

PART 3: DELIVERY

 

Provided you’ve done the due diligence of tagging your tracks, assembling a solid taxonomy, preparing your metadata and collecting your files, the final step of delivering your music should be smooth sailing. All the same, it still helps to follow through to the last mile to make sure that all of your metadata ends up exactly where it needs to be:

 

  • If delivering files by hand, take the time before uploading to thoroughly review embedded track metadata in iTunes, SoundMiner or other ID3 reading software to make sure everything looks right. If embedding to files using SoundMiner’s “Import From Text” feature, listen to the tracks manually while reviewing the metadata to make sure no spreadsheet errors have written your data to the wrong place, or assigned the wrong names to tracks. Double- and triple-check your composer info and splits to make sure every composer is listed for their fair share, before uploading and sending to potential clients. Once a file is out there, it’s out there.

     

  • If uploading to an online distributor, be sure not only to review all visible and hidden metadata (re-downloading your library CSV if needed), but to make sure you’ve fully configured the site to its full potential as well:
      • Are your alts and stems nested properly under their main tracks (if the platform supports this functionality)?
      • Have you set up a landing page, and customized the aesthetics of your template to present a polished, professional face for your brand?
      • Have you confirmed that the dropdowns and other UI elements built into your site are showing up properly, and working well with your descriptive metadata to provide an elegant search experience for clients?

    Pro Tip: Many modern music search distributors pride themselves on providing the best music for a particular search in the fewest clicks possible. How many clicks does it take in your library to arrive at a unique, stand-out search result that makes users say “THIS IS IT!” ?

     

  • If delivering to a sub-publisher, confirm the formatting conventions required for delivery, and double check that your data is complete and organized accordingly. Secondly, make sure all files being delivered match up with the metadata in your accompanying CSVs, with full writer and publisher info filled out accurately. Lastly, consider embedding metadata to the files themselves as a last line of defense in the event of a lost spreadsheet or upload error down the road.

 

While far from comprehensive, we hope this guide serves as a stepping stone on the journey towards the actualization of your music through metadata. It’s easy to get caught up in the fine details, but just remember, what matters most in the end is the quality of the music you make, your ability to present it well, and just doing your best to set the stage for serendipity to strike.

 

You’ve got this!

 

Tagteam Analysis offers music tagging and metadata services to companies and libraries in need of advanced music search tools and optimization. Read more about us on our website: tagteamanalysis.com

 

You may also like