FILE FORMAT STEW
The early days of file-based music were a bit like the Wild West: whatever worked is what was used. Before a legitimate retail infrastructure for online music sales was created, file types and sizes varied (though the mp3 format was by far the most popular), and though the sources were mainly ripped CDs, rampant file sharing made the provenance of many music tracks murky, at best.
The arrival of Apple’s iTunes, in April 2003, not only provided stability to the online music industry in terms of a retail sales and distribution model, but also established some effective standards for music file technology. For iTunes, Apple opted for the AAC (Advanced Audio Coding) codec, which like the mp3 codec, is still lossy, meaning it uses an algorithm to eliminate redundant information to make the file smaller, but includes technical improvements over mp3 such as a variable bit rate, the ability to sample up to 96 kHz (mp3 was limited initially to 48 kHz), and support for multichannel audio, which means it can support surround-sound music. In 2007, Apple expanded its AAC files to 256 kbps (kilobits per second, aka “bit rate”), doubling the format’s transfer rate, which increases fidelity significantly. Meanwhile, the mp3 format has undergone its own improvements, including the mp3PRO version that offers higher bit rates.
The Next Generation
These improvements to the basic music file formats helped move the sonic quality of digital music forward. However, newer digital formats offer even more capacity for improved sonic performance. AIFF (Audio Interchange File Format) uses uncompressed PCM digital data, the same 44.1 kHz sampling rate and 16-bit bit rate as a Compact Disc, and is popular with professional audio users, though the format cannot be played on Apple devices. FLAC (Free Lossless Audio Codec) is described as “losslessly compressed,” meaning that its file size can be reduced by about half without eliminating any information. The HD-AAC codec, released in 2011 by applied research lab Fraunhofer Institute -- which originally developed the mp3 audio codec -- can keep the bandwidth and bit structure of a 96-kHz/24-bit track intact and deliver it at a bit rate of about 700 kbps, approximately half the bit rate of the original recording, and considerably higher than the 256 kbps of the higher-resolution files available on the Apple iTunes web store.
While mainstream music distributors have continued to use mp3 and AAC file formats and their variants, a new generation of record labels is focusing on these and other high-resolution codecs that offer vastly increased sonic quality. Warner Music Group and Universal Music Group now both supply some artists’ music tracks in the FLAC and AIFF formats to HDtracks, a high-resolution online sales venture by audiophile-record label Chesky Records, as well as 320-kbps mp3 files that offer universal compatibility at a higher resolution than the more typical 128k files. HD Tracks also offers 96-kHz/24-bit and 88.2-kHz/24-bit FLAC files, which take music to the DVD level of resolution.
Another high-definition music service, iTrax, takes the notion a step further, releasing only music it has recorded at its own studio, thus assuring that the original recordings were created at as high-definition a level as possible.
Like any new frontier, high-def music has its philosophical differences. iTrax founder Mark Waldrep asserts that distributing music in high-resolution formats is meaningless if the source material is taken from so-called “standard-definition” sources, such as CDs and analog master tapes.
“To me, high-resolution music is defined by the resolution of the source material when it was recorded,” he says. “You can download a ‘high-resolution’ version of [Miles Davis’] ‘Kind Of Blue’, but there were no HD recording decks sixty years ago. So to take an old [Rolling] Stones record from 1964 and put it into what, in reality, is just a bigger bit bucket… might sound a little better but you’re not getting additional resolution, because there’s only what was originally recorded to work with.” However, Waldrep does agree that non-HD content in high-resolution file formats offers a vast improvement over standard MP3 or AAC files.
HDtracks founder David Chesky acknowledges that there is no single high-definition standard for music, but stresses that the files HDtracks sells use the highest native formats -- 48-kHz/24-bit and higher -- available from the labels and the artists, which then undergo quality-control testing in their studios before being converted to the various release codecs.
Mastering engineers, the finals arbiters of sonic quality before recordings leave the professional domain, find some benefit in both points of view. “There isn't much benefit when the original multitrack session is recorded at 44.1 kHz,” says Bob Ludwig, owner of Gateway Mastering in Portland, Maine, where he has mastered recordings for artists including Dire Straits, Paul McCartney and Bruce Springsteen. “Mixing it through an analog desk and recording your mixes at 96-kHz/24-bit, for instance, will not gain you that much except [access to the] the distortion overtones of the analog gear, which now will go higher than [44.1 kHz]. Still, if I was in that situation, I would [remaster] it at a high resolution just to keep it as good as it can be.”
There are more interesting and exciting developments in high-resolution audio codecs coming in the near future, all aimed at enhancing the listening experience for music aficionados in pursuit of better sound.
 “Redundant” is a subjective term. In a passage of music, instruments that play in approximately the same frequency range can be said to be supplying identical information. Clinically speaking, compression algorithms (i.e., codecs) analyze this and decide we only need to hear one instrument playing in those frequencies at that time and eliminate some of the other instrument information to make the file smaller. The reality is that this kind of “lossy’ compression also eliminates the subtle harmonic overtones that the lost instrumentation brings to the track. This is the sonic-quality-versus-convenience tradeoff that has characterized digital music from the start.