Thanks Alón, for inviting me to write for the San Francisco Audiophile Society! Hello, SFAS, I’m Dan Gravell from the U.K., and I have a long background in computer audio. I’ve been ripping CDs since 1996 and I’ve grown increasingly interested in how audiophiles have embraced lossless and high quality computer-stored audio.
I tend to think of developing a music library a little like gardening — just as you plant, tend, prune, trim and protect your garden, you curate, purchase, reorganize, and tidy your music library. In this first installment, I’ll share what have I learned and how I now approach music library management after almost twenty years of digital “gardening”.
Imagine a green field – landscaped and ready to plant. This is a little like your music collection when you first start. This early period of building your computer music library is characterized by enthusiasm and a youthful naiveté. Enthusiasm, because all of a sudden you have your entire music collection accessible at your fingertips, without ever having to get out of your chair to change a CD! Naiveté, because these exciting new possibilities can mask hidden dangers…
Acquisition of the music within a music library takes different forms. If you have a large existing collection, then the transfer of music from, say, CD, to computer audio formats gives a “big bang” of a large quantity of music. From there, music can be acquired from different sources on a drip-fed basis; more CDs, downloaded music and so on. So your method of acquiring music is something that needs to be considered throughout your music library’s lifecycle.
Early on, with a generally smaller library, music library management seems easy. More than that, it’s exciting. You realize that it’s possible to classify your music in different ways which allow interesting ways of playing your music collection that hadn’t been apparent before. Sure, you can just play album by album, but then you realize you can shuffle all the tracks from a given year, taking you back to your college years, or maybe queue all the tracks by a given artist in one playlist, allowing you to track the development of that artist.
The underlying classifications that make this possible are called tags. Tags are textual metadata stored inside music files (generally). It’s these tags that say what the name of a given track is, the album name, its artist, genre and year of release, and so on.
Tags go further. You also begin to spot ways of accessorizing your music collection, such as adding more details to the textual metadata (perhaps the lyrics, the mood of the music, occasions for playing the music or maybe your ratings for the music). You notice how it is possible to assign artwork to releases, which make your music collection not only sound beautiful, but look the part too.
But here be dragons. Tags and artwork do not come for free. They require time and attention to detail to get right, not just in terms of correctness, but also in terms of completeness and consistency. These “three Cs” are something I’ll return to.
This realization brings us to where things begin to get more difficult. Every Fall, and every flowering period, your plants need tending, reshaping, pruning and tidying. With a growing music collection and more experience of using your music players you begin to notice gaps, inconsistencies and incorrect metadata. These begin to annoy you more and more, so you set about doing something about them.
There are many reasons that music metadata goes bad. In most cases it starts from acquisition. There are numerous different channels by which you can acquire music. CDs are “ripped” from source and transferred to your computer by software called CD rippers. These rippers apply their own logic to download and insert those useful metadata. Then there are download sites from where you can purchase and download music; these sites also have their own processes and data feeds to insert that metadata.
The trouble is that all these sources of metadata have their own quirks and inconsistencies. Some of them have “style” and “mood” classifications, some don’t. Some include the entire date in the “year” tag, some don’t. Some include high resolution artwork, some very small artwork. Some will tag an album as being by “The Beatles”, some as “Beatles”, and worst of all, some as “Beatles, The” (although there are reasons for that last one; it’s about understanding your music players which I discuss below).
These quirks affect the “three Cs” of music library management. Namely: correctness, completeness and consistency.
Correctness is how correct the metadata is. This can sometimes be pretty obvious, like if the name of a given album is incorrect. It can also be more subtle because many fields of metadata are subjective rather than objective; for example, genre. Finally, sometimes the metadata can be surprisingly correct… For example, if you have a digital remaster of an album it may be tagged as being thirty years younger than the actual release.
Completeness is making sure all the metadata that is required is stored in the music files. The largest barrier to this is data availability. At acquisition time this is less likely to be a problem, although if you later decide you want more metadata and some music is missing this data, you will need to find a way of finding and inserting it.
Consistency is checking that all metadata fields are used and formatted in the same way. For example, all “year” fields should be tagged simply as the year, with no month or date information. This is important so that your music players know how to interpret the metadata and present the music to you in a useful way.
As you begin to learn about these potential problems, you begin to enter the next stages of gardening. You become more confident in planting and pruning, and jobs are completed in short order.
At this point you understand the costs of music library management and take steps to avoid and lower these costs. In many cases, the answer is to minimize your use of metadata where possible and to use automated software to manage that process.
Metadata minimalism works because having fewer items of metadata means fewer opportunities for noticing gaps, typos, incorrect data or formatting errors. Of course, this is a balance — too little metadata and your music collection becomes less useful!
Software to organize your music collection comes in different forms. There are simple “music taggers” which essentially just open up every file and allow you to tinker to your heart’s content. And then there are more automated tools that will change your files to adhere to defined structures and rules. In most cases, using the latter is preferable for an easy life.
It’s also at this point that you understand that the very reason for managing your music library is to make sure all of your music players display your music correctly, making it easy to choose, search and browse your library. This means the metadata you choose depends on a combination of how you want to use your music player(s), and how your music players will interact with the metadata.
Essentially, the music player is king, and kings have a habit of making you work around them.
Not all music players support all metadata, and not all treat it in the same way, so working out what metadata to store is a process of working out the lowest common denominator between the music players you use and then using that to derive the “minimum viable metadata” that you can get away with storing.
And just as you think you understand all this comes… attack! Just as a garden has a number of different attackers: pests and disease, the same exists of music libraries.
Such attacks can come in many forms. It could be accidental or malicious. It could be caused by yourself in a weaker moment or a family member not understanding what the “delete” button does. On the other hand it could be a computer virus that wipes out your music files, or even a burglar who makes off with your music server.
What you need is security. By policing who has access to your music library, and by providing a backstop should the worse happen, you can recover from attack on your music library, be it intentional or otherwise. Your ultimate failsafe when it comes to music library security are backups. If you do nothing else towards security, do backups, because that way, any calamity can be recovered from (apart from, er, destruction of your backups).
There are three things to think about with backups. The first is how they are done. Automatic backups are best, because that way no-one needs to remember to do the backup and so they get done, rain or shine.
The next thing to think about is frequency. The more often the better, but there comes a point where backing up constantly is a waste of time. The frequency should be aligned with how often your music library is changed, either through addition or reorganization. Typically, once a week or month is adequate.
Finally, consider where to store the backup. Off site is best, but for a lossless and high resolution audiophile music collection this will take a long time to transfer. Removable USB discs can store your backups and be moved to another part of your house, or moved off site, to lower the chance of physical damage to your home affecting both your music and its backups.
Stepping back, my observation in twenty years of curating, developing and grooming a computer music library is that it is a long term but rewarding commitment. By dropping back in, fixing inconsistencies, filling gaps, you can keep your music library in order… and even develop your own green thumbs!
Stay tuned to more of my thoughts on computer based music library management, as we roll up our sleeves and begin to get practical! If you have any queries about album artwork, metadata or would like something specific covered – drop me a line in the comments and I’ll look to include that in future articles.
Dan is a music and technology lover from Melton Mowbray, a town in the UK. Dan has been curating and building his computer music library since 1996, and he now puts this knowledge to good use in developing his software product: http://www.blisshq.com