FLAC on the Mac â€” A brief summary of digital audio/music file formats/types
This article covers the most common digital audio file types. I discuss why they were developed, how they’re used, and their advantages and disadvantages. Then I explain how best to preserve and protect a music collection from physical loss, or loss of compatibility with current hardware and music players.
High resolution, or hi-res, music files, also called tracks, are big news these days. Since most Mac users play their music tracks on the default iTunes music player included with the Mac’s operating system, those users will be curious about the advantages of hi-res music. Is it worth the effort to download a new music player that will play those hi-res tracks?
The most common music file types are the ones I’ve dealt with over the past 17 years since the MP3 file type revolutionized personal audio, particularly portable audio. My first MP3 player was the Diamond Rio, which came with a ripper: a software program that converted music tracks on Compact Discs (CDs) to 128 kbps MP3 tracks. CD tracks used approximately eleven million bytes of data per minute to play the music, and the compressed MP3 file used only one million bytes per minute, or eleven times less.
Using MP3 files circa 1998, my 32 MB (not GB) Diamond Rio could store about 30 minutes of music, but the CD file format (WAV) would play for less than three minutes on that device. Most of today’s computers store hundreds of gigabytes of data, and where a typical 4-minute pop-music track in the 96k hi-res format consumes nearly 100 mb of data space, a lossless track in the FLAC format compresses that 100 MB to 60-70 MB on average.
The term kbps (kilobits per second) is used with most music players to describe the rate at which data is consumed by playing the music file. The term 96k (96 khz) is used to describe the sampling rate of the recording process, which is 44k for CD tracks. If the music track is completely uncompressed, the 96k tracks will use about 2.2 times as much data as the 44k tracks.
The term lossless indicates that even if a music track is compressed (i.e. FLAC format), the music player will uncompress the data automatically, hence nothing in the music data is lost. The compression used to make MP3 format files is called lossy, since only the musically relevant data is retained in the MP3 and the remainder is discarded.
The iTunes music store currently provides music tracks in the M4a file format, where I presume the a relates to audio and the M probably designates media. The current M4a files that I purchase are compressed to approximately 256 kbps, or 2/11 of a CD track’s data rate. Given that the compressed (but lossless) FLAC format is usually about 1/3 smaller than a corresponding CD track, I assume that the 256 kbps data isn’t lossy to the extent of retaining just 2/11 of the data. Using the FLAC factor, i.e. that the CD WAV file can be shrunk by 1/3 without losing any real data, then the 256 kbps track probably retains about 3/11 of the original music data.
Looking at both ends of the issue, most people are satisfied with the sound of 256 kbps music files, yet a growing number of users are demanding resolution much higher than the 44k CD format. Many listeners who rely on digital music want to know exactly what is being lost with the lossy file formats, and what if anything is being gained by purchasing music in the hi-res formats. My experience is that most people won’t hear a difference between a CD-ripped MP3 or M4a file in the 256 to 320 kbps range when taking a quick listen, or even when sitting through a comparison test where another person controls the test equipment and procedures.
With good hi-fi gear on the other hand, including a late-model Mac or Macbook computer with a USB DAC (digital to analog converter) such as the Dragonfly, Microstreamer, or Audioengine D3, and an unhurried solo listening session, most users with good hearing will detect a loss of air or a less-than-lifelike sound with the MP3 or M4a. Moving up the other direction to hi-res files, it gets more complicated. When CDs were developed, the so-called Red Book data rate of 44k was considered to encompass all of human hearing, since humans can hear tones up to 20 khz at best. But there are other audio properties besides frequency response, and much of that is controversial in relation to hi-res music tracks.
Regarding the sonic advantages of these hi-res music files, any number of impressions are possible due to the myriad recordings available and the equipment and methods used to record that music. My first response is similar to the comparison I made above. The CD track might lack some of the air or realism of the hi-res track, although the differences in most cases are very subtle when both tracks are made from the same master, with meticulous care.
My other consideration is for situationsÂ where the 44k CD master incurs distortion due to the 44k resolution limit. In those cases the hi-res music file may sound as though it has less air or less fidelity than the CD track. It actually is lacking the extra effects added by the down-sampling process, which converts the original high resolution recording to the 44k limited file format. Skeptics will contend that nothing audible should be lost in that down-res process, since the Red Book standard says so. Since some audio information is actually lost, because 44k is a lot less than 96k. Therein lies the controversy: whether the lost information is audible or not.
Most of the file types discussed here are open or free standards. Converting from one format to another is as easy as downloading a converter, and possibly a plugin for a specific file type not already included in the converter. There are a number of proprietary audio file types in use today that offer advantages over the open types for certain music playback applications. Be aware that converting away from those types to an open type may be costly or extremely time-consuming, if the user decides afterward to quit using the proprietary system.
Business people have been accustomed to backing up their computer files onto various backup media for decades, especially the high-capacity external hard drives that have capacities of a few hundred gigabytes to several terabytes (one terabyte is one trillion bytes). Digital music users who have accumulated a few hundred to even a few thousand digital music tracks often have not backed that data up to an external drive, which can restore their files to a new computer when the old computer fails. It’s best to use backup drives with software that can automatically back up new files, such as Appleâ€™s Time Machine. Older files then get updated, so those files aren’t lost when the original computer drive fails.
This article compresses a lot of concepts into a small space, in the hope that more people will take the time to read it. But this presentation focuses on ‘linear’ differences in sound, i.e. how much loss is audible in compressing CD tracks to the M4a or MP3 format, and on the other end of the spectrum, how much loss (if any) is audible when a high resolution master is downsampled to make a CD.
But that’s all about linear losses. Another, even more controversial topic involves extra-digital data that intrudes into playback systems via low-quality cables and connectors, and imperfect filtering by the DACs that turn digital data into analog sound.