This morning, the EMI Group announced that they would begin offering their music library for sale online without DRM. Not only that, but the DRM-free files would be encoded at a higher bitrate than their "protected" counterparts. Of course, this comes with a cost. Rather than the now-standard $0.99 per song, these files (which certainly do offer more to the consumer) would be priced at $1.29 per song.
As I was reading this, I began to think about what the real value of a digital music purchase is and whether the current pricing model makes sense, so I looked up the numbers. What follows are my results.
To properly consider the value of a downloaded track, you must first know the value of that same track on a physical CD. According to the RIAA's own numbers, the average price of a CD in the year 2005 was $14.91, which I will be rounding off to $15 for the sake of simplicity. Dave Taylor tackled this question using his personal collection and found an average of 12.5 songs per disc. Again rounding in the favor of the RIAA, I'll just say 12 for the purpose of this article. This gives us an on-disc value of $1.25 per song, on average.
Now we should define what the CD is. It is an uncompressed representation of audio, using two channels sampled 44,100 times per second at a 16 bit resolution. It is stored as standard PCM audio with absolutely no encryption or DRM, and can be converted to other formats by a wide variety of free and commercial tools.
Based on the above, one CD quality, DRM free audio track should be sold online for around $1.25 . Some argue that since you are not getting the physical media, liner notes, etc. that a download is worth less. I disagree, because of the convenience. I can have my downloaded song NOW, any time day or night, and I don't have to go through a potentially complicated process to import the audio to my computer and/or put it on my portable audio player of choice. Convenience is worth money, and I argue that it balances out the lack of physical content and drastically reduced distribution costs.
Obviously distributing the raw CD audio online is wasteful at 1.4mbit/sec, but this is where lossless audio compression comes in to play. Using an encoder such as FLAC, the file size can be cut in half on average with absolutely no decrease in sound quality. If you combine lossless audio with an appropriate cue sheet, you can burn a CD that is bit-for-bit identical to the retail packaged version but took half the space on your hard drive and half the time to download.
Some may be wondering "where do bandwidth costs fit in all this?" After researching the costs for big bandwidth at a colocation center, I've determined that bandwidth is basically negligible in all this. At a Level(3) datacenter near my location, for large consumers (as an online music store certainly would be) bandwidth runs about $240/TB. That's under 24 cents per gigabyte, meaning that a full CD worth of losslessly compressed audio would incur roughly eight cents of bandwidth cost and for single tracks the cost goes below one cent. Tracks sold with lossy compression would cost even less, easily going below one cent per album if compressed to a level more suited for flash memory based portable players.
Back to the main topic, since we've now determined that DRM-free perfect copies of the CD audio are worth $1.25 on average, what about lossy compression or DRMed tracks? Both lose value to the consumer, arguably more is lost from DRM than compression because DRM limits the number of devices the content could be played on, where lossy compression only makes it sound a bit worse (which some people wouldn't even notice).
I'm going to start with lossy compression, since it's the most straightforward in my opinion. Let's say a hypothetical music store offers FLAC content for the already determined $1.25 per track, but also offers the same music in three grades of MP3. The top grade would be 320kbps constant bitrate, generally regarded as "archival quality" for MP3 and nearly indistinguishable from the source or lossless copies. A full album at this bitrate takes up around 175MB. The middle grade would be a variable bitrate copy averaging around 190kbps. A full album would be around 100MB, though it can fluctuate depending on how complex the source material is. This quality is still considered to be good enough and again would be hard to tell from the source on mid-to-low end audio equipment. The lowest grade would be around 115kbps, again variable bitrate, and while the quality would be noticably lower than the source material it would be good enough for a portable player or low-grade audio equipment.
I'd price the highest grade of MP3 somewhere around $1.15 per track because it is hard to tell the difference in quality. The middle grade would land right at the current sweet spot of $1, and the lowest grade would be between 75 and 80 cents, if even that. It may be true what the record industry claims that prices can't be lowered much further than $1 for most tracks, but at least popular ones where the initial costs have already been covered should offer this option.
DRM is another beast entirely, because there are so many variables. Some, like iTunes, allow effectively infinite CD burning but are locked in to either one or a small number of portable players. Others, like Microsoft PlaysForSure, have a wide variety of restrictions which can be set limiting burning, transfer to portable players, and even number of plays. The one thing that can be guaranteed is that any DRM locks out a portion of the potential customers. Apple's Fairplay only works with iTunes and the iPod, for example, while Microsoft's PlaysForSure only works on Windows PCs with Windows Media Player and locks out the most popular portable player on the market. With either of those two choices, you've automatically halved your market. With that in mind, it is my opinion that DRM should be limited to all-you-can-eat subscription services like Napster where the trade is made giving up flexibility for unlimited access to a large collection of music. For those who insist on using it on "purchased" media, I'd put the value at below 50% of the unencumbered media, since at least half of the transferability/compatibility is lost.
When looking at this information, it's clear that EMI is making a step in the right direction. Their higher quality DRM-free offerings are still overpriced compared to CDs, but will be the best bang-for-the-buck in the big-label music download business for the near future. If the entire iTunes catalog went DRM-free at the current prices and quality, it would be right where I'd expect it to be given these numbers.
In conclusion, the problems of DRM seem to be finally catching up to the music industry and things are beginning to change. It is my opinion that we'll see prices competitive with CDs reaching online music stores very soon after the industry finally abandons their attempts to encrypt everything.
Comments and corrections are welcomed and encouraged.