Techgage logo

Audio Archiving Guide: Part 1 – Music Formats

Date: September 10, 2008
Author(s): Matt Serrano

Have a large music collection just waiting to be archived? We’re here to help! In this three-part series, we’ll be tackling all there is to know about properly building your collection, from choosing the right codec, to ripping, to archiving. In this first article, we’ll take the frustration out of finding the perfect codec – one that fits your style, and needs.



Introduction

The world of digital audio can be a daunting one. While many perplexing challenges have been made more straight-forward over the years with the advent of better software and hardware that’s easier to understand, there are still far too many music lovers that don’t know about the precious digital music they have sitting on their hard drive.

There’s a lot to learn about the inner-workings of digital audio, and although these articles only scratch the surface of their technicality, reading them will hopefully set you on an illuminating path that will help you make better-informed choices in the future. After all, music will follow you for the rest of your life. If you want to keep your collection in pristine form, making the right choices now will save you a headache further down the road. 

This is the first part of a three-part series of in-depth articles about digital audio, and how to rip, convert, or download them and organize in the most effective and convenient means possible. Upcoming articles in the series will go into more detail and technical explanation.

Common Audio Formats

For most non-critical listeners, downloadable MP3s – or for that matter – the default settings many media players carry when ripping CDs may be more than enough, but that’s part of the trick. In my experience with high-quality audio, one never knows how high their standards truly are until they experience high fidelity sound first-hand. There are other options available that may fit your needs and still provide an inconspicuous difference in sound quality when compared to the source.

The MP3 codec, arguably the most recognizable and widely-used audio format, was created by the Fraunhofer research organization in conjunction with other parties (as noted by Wikipedia) in 1991. Contrary to what many may surmise, the format is not free to use and requires a license from Fraunhofer to decode (play back) or encode (convert or create) an MP3 file, which is the primary concern for open source and electronic rights groups.

Although the format is so popular, enthusiastic listeners consider the traditional Fraunhofer encoder (which is used in most media playback software, such as Windows Media Player and iTunes) to be woefully inadequate. A popular alternative MP3 encoder, sardonically named “LAME,” has been coached to replace the traditional codec in many people’s collections.

In the audiophile circle, LAME is considered to boast largely improved audio quality over the vanilla MP3 codec, and it even competes well with other, more non-traditional ones as well. There is almost no technical limitation to the format in today’s world; older MP3 decoders had trouble playing them back or correctly displaying the track length on variable bit-rate files, but the issues have been all but eliminated.

There may be an outstanding issue with work flow, because ripping CDs or converting music files would usually be done with a separate piece of software unless the one you currently use supports it, which, of course, may or may not have an affect on you if you’re willing to switch or go between programs for better audio quality.

More recently, the AAC format (the designated successor to MP3) has begun, albeit slowly, to take hold of the market. Even though the codec may be most-known for its use in the iTunes music store, it is only used by Apple because it is a standards-compliant format that has digital rights management (DRM) control built-in. Like MP3, a license is required to encode or decode the format, but actual content can be distributed without a licensing fee from the standards groups.

iTunes is largely considered to have the best AAC encoder in terms of quality, but many choose not to use it because of the reliance to the program and Apple. Nero’s interpretation comes to a close second, but unlike iTunes, it is not free (update: see below) to use and must be purchased with one of their products. AAC is said to hold comparably better quality than MP3 at lower bitrates, but with competition such as the LAME codec, conclusions are hard to draw in active listening tests. Despite its advantages, it is not as distributed because of the overall lack of hardware and software support.

September 10th Addendum: Nero contacted us to confirm that the Nero AAC codec does in fact perform better at lower bitrates, using the HE-AAC switch with their encoder, because iTunes provides no such option. Their codec is currently a free download from their website.

The Windows Media Audio format has taken the second place in the audio popularity contest, mostly because of the active hardware and software support in the format’s heyday. Most of the press was promoted by Microsoft’s licensees, and much of the format’s success, similar to AAC, can be attributed to its dominance in online music sales and aggressive marketing.

While Microsoft previously stated that WMA had comparable quality to MP3 at half the bit-rate, most professionals agree that the statement is largely false. Though a WMA file encoded at 64 to 96 kbit/s, in its standard form, may be able to compete with the traditional Fraunhofer MP3 encoder, the real world benefits are considered negligible when paired against MP3 files at equally high bitrates.

As one might expect, there are other options available, but for the most part, the alternative formats have failed to capture the public’s eye. In the next section, we’ll be taking a look at these lesser known formats.

Less Common Formats, Lossy vs. Lossless

Vorbis (sometimes commonly referred to as OGG Vorbis) is a fully-open source, royalty-free format managed by the Xiph.org foundation. Even though the audio quality supplied by the encoder is viewed as superior to MP3 and many other codecs, hardware makers have yet to embrace the format with the exception of some iRiver, Samsung, Cowon and other devices, and neither Microsoft or Apple have shown interest in including support for the codec.

There are some additional advantages and disadvantages to using Vorbis. Because the technology is open, anyone is free to use it under the GNU Public license. Listening tests have also come to the conclusion that Vorbis offers better quality than WMA and AAC at lower bitrates of 48 to 64 kbit/s.

Though some listeners believe they can hear an apparent difference at bitrates of 128 kbit/s and higher, many cannot. With the format’s open nature, many forked projects have appeared, including aoTuV, Tremor, and FFmpeg, which may intimidate potential users with the sheer amount of options and potential compatibility issues.

Like Vorbis, FLAC is another free, open source format. It was brought under the Xiph.org umbrella of codecs in 2003 and is given away under the same GNU Public License. However, unlike the format’s distant cousin, FLAC is a lossless codec and works differently than the formats listed above.

FLAC does have more support with hardware than Vorbis, with noticeable mentions being iRiver, Cowon, Squeezebox, and Sonos, but because of the increased file size (we will go into more detail later) the practicality of using such a format on a portable device isn’t likely for most people with sizable collections.

Monkey’s Audio is another free lossless codec created primarily for the Windows platform. It offers far less compatibility (which is considered its main downside) than its main competitor, FLAC, but it does have a higher compression rate, and subsequently, smaller file sizes. Unfortunately, some questions have been raised about the format’s licensing agreement, which has prevented it from being included in GPL projects.

Finally, Apple Lossless (ALAC), as the name suggests, is a codec created by Apple. Even though it is used in the MP4 container (with a .m4a extension), it is more akin to FLAC than the company’s widely-used AAC codec. Originally released in 2004 as a new feature of iTunes 4.5, the codec is undocumented by Apple, so it cannot be officially decoded or encoded by a third party. However, an open source decoder was created by two software engineers in 2005, and dBpoweramp has released a reverse-engineered encoder.

Hardware support is officially limited to Apple’s iPod and iPhone, but Cowon does offer a ALAC support on the popular A3 PMP. Because of the closed nature of the format, it isn’t as widely used as other alternatives, but it does have a 40% to 60% compression rate, which means it can perform better than more popular codecs in some cases.

Finally, Microsoft’s WMA Lossless derivative provides another option that integrates directly with Microsoft’s Windows Media Player and Zune software. It doesn’t have much hardware support, which is mainly limited to Toshiba V and S line of MP3 players and the company’s own Zune player (4, 8, 16, 80 and 120 models), but the compatibility with Microsoft’s media players, including Media Center, make it viable for some people.

Lossy vs. Lossless

The “lossy” codecs differ from formats like FLAC, ALAC, Monkey’s Audio and WMA Lossless in a few key ways. For example, all lossless formats contain the same exact quality of the originating files, compared to codecs such as MP3, AAC and Vorbis which remove quality to make the file size smaller. This means that lossless codecs can be converted to other lossless formats without a hit on quality, but an MP3 file will lose quality with every subsequent conversion.

Left: WAV file ripped from CD, MP3 encoded at 128 kbit/s using LAME
Right: WAV file ripped from CD, FLAC file encoded at compression level 8

Lossless formats work in a similar fashion as .zip files. The audio can be compressed to a smaller file size, but all of the information is still intact. In spite of this, lossless files, even highly compressed ones, take up a lot of space compared to a lossy file, even at a higher data rate. For some people with limited storage, using lossless codecs would be a practical impossibility because they are able to store much less than they normally would.

To better show the differences between both FLAC and high-quality MP3, we calculated the total sum of our music collection, which we keep in both FLAC and 320 kbit/s MP3 formats. Included are 6,516 tracks from 412 albums:

FLAC (Compression Level 6): 172 GB
MP3 (LAME 320 kbit/s): 58 GB

Aside from that, lossless files typically have a higher requirement for CPU processing. Although most older machines should have no problem playing most formats, a machine that’s multitasking resource-intensive applications might have trouble handling the load and preserving the playback quality (i.e., there could be increased noise or stuttering playback).

Using lossless formats on portable devices may be a concern for two reasons. First, the increased size could prevent users with more exhaustive libraries from storing larger portions of their collections on the device. Second, the files could drain the battery faster because the music has to be cached in memory more often, forcing the device go back and forth between collecting and caching the data. Some lossless decoders could be less efficient as well, but this limitation is on a per device basis, and may not apply to you.

ABX Testing, What’s Best For You

With the amount of codecs available, naturally, a user is left with many options. The question of which format offers the best quality is best answered by your ears. A simple solution, which is widely preferred by the audiophile community, is a double-blind ABX test, which will test your senses and show you a clear cut answer.

The concept of a basic ABX test is to compare two samples. A user is given three total options: sample A, sample B, and an unknown. After listening through the three choices, one will have to decide whether the unknown is sample A or B. With each right answer, the probability that the person is guessing decreases.

Another adaptation of the test is the ABXY mode, where four choices are given in total. The person being tested is expected to listen to all four choices, where two of them are the same, and find the correct match. Like the standard ABX test, the probability of the person guessing goes down when a right answer is selected.

One method that can be used to determine which format and bitrate to use is to compare two rips from a CD, or conversions from a single source at different bitrates, one at the lowest you believe you are willing to go, and then one slightly higher. If you pass the tests with a low probability of guessing, you should move both bitrates, until you have trouble.

As an example, say I want to know the minimum bitrate that I should rip my CD collection to. I rip a song to MP3 at 128 and 160 kbit/s. After ten trials, I get eight tests right and receive a 95% confidence level, so I move on to 160 and 192 kbit/s. If I do relatively well, I would continue on, but if I do poorly, I would stop at 160 kbit/s.

ABX testing could also be used for figuring out if one is able to hear the difference between CD-quality music and compressed music. A test between a WAV file and an MP3 would show you if you could hear the difference between compressed music and the highest possible alternative, leaving you to adjust your settings accordingly.

Programs designed for ABX testing include WinABX and foobar2000 with the ABX component installed for Windows, ABC/HR (Java) for OS X, and ABX Comparator for Linux (Gnome).

What’s Best For You

It’s important to remember that music quality is subjective. Even though the source material, such as a shrink-wrapped CD, would offer the best possible quality, the sacrifices made will be the most important factor to weight. With that said, when choosing a format, the first choice should be compatibility.

If you’re running Windows, your immediate options will be MP3 and WMA. If you’re on the Mac, your best choices will be MP3, AAC, or Apple Lossless. Finally, if you’re on Linux, or another open source operating system, the most widely used options will be MP3, Vorbis, and FLAC. Of course, if you want to have the ability to play other formats, downloading other software will be mandatory.

The next choice to consider is the transferability between platforms. If you have an MP3 player or PMP, the first thing to do is take a look at what the device supports. Zunes, for instance, play MP3 and WMA files fine, while iPods are more suited for AAC and MP3. In addition, if you use multiple operating systems, you’re best off choosing a format that is supported by the software you use.

While lossless codecs like FLAC are an option because they can be used to preserve tracks and albums, the space needed may be too much for some users. In this case, it’s worth evaluating whether or not the space you can allocate to your music will how encompass the storage requirement, and the rate of growth for your collection.

Join us in the coming weeks for our audio ripping guide, where we’ll take a look at different applications and go in-depth about the technology involved.

Discuss in our forums!

If you have a comment you wish to make on this review, feel free to head on into our forums! There is no need to register in order to reply to such threads.

Copyright © 2005-2019 Techgage Networks Inc. - All Rights Reserved.