Computer Audio: A Guide For The Perplexed (HiFi-Plus 70)

Posted by: Alan Sircom at 12:12 am, February 3rd, 2010

There are also systems that take over the whole audio front end. Music client systems like Sonos and Logitech's Squeezebox include the display and D/A conversion in a single box that becomes just another source for your preamplifier or amp, with all the file storage handled by a remote computer in another room. This computer becomes known as the 'music server'. More exotic versions of this arrangement exist specifically for audio (Qsonix, Meridian Sooloos) and for audio and video (Kaledescape).

 

Files of all shapes and sizes

The tracks on your CD are, in fact, individual data files, stored in a format called WAV in PC speak, or AIFF for Apple users (the difference between the two files is down to byte order, but although this might be of great importance to geeks, has no audible bearing to sound quality). These are uncompressed 16bit, 44.1kHz PCM files, exactly the same as those read by every CD player since the early 1980s. As each CD can store up to 650 megabytes of information, storing lots of CDs on a computer without any form of compression will soon eat through a lot of computer hard disk space. Data storage is cheap these days, but the amount of time it would take to archive and access all those gigabytes of information slows things down and kind of defeats the object of the exercise.

This is why compression is used. The word seems to automatically send shivers down the spine of an audiophile, but it need not be so scary in reality, because there's compression and then there's compression. The wholly benign form of compression is lossless. As the name suggests, the file is shrunk to take up less storage space but without any negative effects to the sound file itself. For the digital photographers out there, this is like taking a huge TIFF file and using ZIP compression to store and send it out. When the file is used, it's unpacked to its original size and no one's the wiser. Common lossless compression systems include ALAC (Apple Lossless Audio Codec), APE (Monkey's Audio own codec) and FLAC (Free Lossless Audio Codec) and typically shave the file size by anything between one third and one half.

The next stage is data compression, as opposed to file compression. This uses complex perceptual coding algorithms to determine not what data can be thrown away, but what musical instruments. As we fail to hear sounds that are masked by similar, louder sounds, so these quieter sounds can be removed to save space. There are many different types of this kind of 'lossy' compression (so called because once that data is lost, it cannot be retrieved) but the most well known is MP3. The result of all this space saving is that where FLAC can halve the size of a music file, a 128kbps MP3 file can shrink the file by as much as 11:1.

Different levels of compression can be applied to a piece of music, a trade-off between space saving and audio quality; a 320kbps (or kilobits per second) file is notionally indistinguishable from an uncompressed 16bit, 44.1kHz PCM track but results in a relatively large file, while a 96kbps file of the same piece of music will result in compression artifacts that undermine the performance on almost any kind of playback (these can manifest as swirling 'flanging' sounds). Somewhere between the two lies a comfort zone for most people, but for serious listening on large hi-fi systems, most audiophiles seem to prefer lossless files.

 

War, what war?

With different architectures and file types, this sounds complex and like the start of a format war, but the reality is completely different. Most of these systems are more similar than they are different, and they rely on the same basic protocols throughout. That means in most cases, if you buy one system and decide you want to change, your music comes with you without complaint or without bother on your behalf. In the case of the music player programs, they act independently to the music they play, because data is data. Some formats are proprietary (Apple's Lossless coding, for example) but transcoding to more universally used formats is possible and doesn't affect the music files. The days of the format war are hopefully over in computer audio.

There are decisions to be made over the choice of music player, but these are to do with picking out a system that suits your way of listening to music, not about influencing the sound. For many, music player means iTunes, Apple's almost ubiquitous media player software. It's the default choice of the Apple Mac user and the Apple iPod and iPhone user, which makes it incredibly popular. Other media players are available though, and some prefer the functionality or the simplicity of the alternatives.

For PC users, good choices include MediaMonkey, Foobar2000, WinAmp, and many more. Apple users get Songbird as an alternative to iTunes, with more to follow. Linux users get Rhythmbox and everyone can use VLC. Hardcore users also use a standalone 'ripper' software to transfer your music from CD with absolute clarity. Good rippers include Exact Audio Copy and dBpoweramp for PC, Asunder for Linux and Max for Mac, but unless you have a lot of very scratched discs, you'd struggle to hear the improvement. It's worth exploring the options here, but it's easy to get bogged down in minutiae, especially as each has its own set-up idiosyncrasies.

Comments

Steven Stone -- Wed, 02/03/2010 - 19:24

Nice piece.

Even I understood it.

:)

Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications

superfalcon (not verified) -- Thu, 02/11/2010 - 12:37

I understood the article and it was a great discussion of the technologies involved, but it was hardly a beginners guide. I was disappointed because I thought this was going to be a step by step how-to on getting the best fidelity from a computer as a music source. How about a detailed guide on the best way to to rip, store and play music from internet sources? I know this is basic knowledge for some, but I am just now considering using sites like lastfm to download music. I want to do this with the greatest fidelity possible from these sites and in any of the playback modes I choose. If this is too basic for most of your readers, where can I go, and what is the best source to learn these methods?

Mr Plus -- Thu, 02/11/2010 - 15:44

This is merely the first of a series of computer audio features. The purpose of this feature (which, I admit, loses some of its impact online) is to introduce people to the topic. Later features will focus on more specific aspects.

There is still a lot of hesitancy surrounding computer audio among audiophiles, and many UK audiophiles are something like a year or so behind our American counterparts, in part because we still have HMV Shops selling CDs in every town.

Alan Sircom
Editor, Hi-Fi Plus Magazine
London, England
editor [at] hifiplus [dot] com

superfalcon (not verified) -- Thu, 02/11/2010 - 23:40

Thanks for your reply! I have been a fan of high quality music reproduction for many years, but I am a newbie when trying to use my computer as a music source. I have had friends try to help me but I ended up with mixed results. Also, until I read your article, I never knew that the MP3 format used compression that lost some musical detail. I'm glad I waited before I dove into that technology. I also want to make sure my next vehicle has a quality music system. Some standard six speaker systems sound great, while some hyped 10 and 11 speaker systems are less than stellar. Anyway, thanks for the info and I'll look forward to your next contribution.

Mr Plus -- Fri, 02/12/2010 - 06:23

Perceptual coding works on the principle that if you play two otherwise identical sounds at different volume levels, the quieter sound is masked by the louder and therefore the quieter sound can be eliminated. This is a very well documented observation. Audible problems surrounding compression systems are based on either the compression system failing to do its job properly or doing its job too well. In reality, that usually means the latter as codecs 'failing to do the job properly' are weeded out long before they make it to the general public.

If a codec (enCOder DECoder) is applied too heavily for the replay system its intended to be used with, the data reduction can be clearly audible and at its worst ends with the 'phasey' sound I described in the feature (for 'phasey'... think of the whooshing guitar sound of Jimi Hendrix playing the opening bars of 'Castles made of Sand' only slightly milder and faster sounding). At its most benign though, data compression can be at least notionally impossible to distinguish from the original sound.

This needs careful listening before deciding to go MP3 though. Many of those who develop the code test it on music that gives it an awfully easy ride - it's worth remembering that the bulk of the original development work on MP3 used "Tom's Diner" by Suzanne Vega, because it's easy to hear imperfections on such a relatively uncomplicated track. Not all music is so straightforward - see what happens when you throw half an hour of Bruckner or an Albert Ayler or Eric Dolphy improvisation at the same compression rate before adopting a compression rate.

Alan Sircom
Editor, Hi-Fi Plus Magazine
London, England
editor [at] hifiplus [dot] com

1likeh1f1 (not verified) -- Thu, 02/11/2010 - 15:27

Alan,

This is a very good article and I enjoyed reading it. This is a very exciting area of our hobby and one that, I believe, can bring in more enthusiasts than any other form of hi-fi playback existing today. I would like to suggest that you add some content regarding how the "bugaboo" of digital, jitter, is being addressed with today's technologies and approaches. Folks, especially those who have had exposure to high quality anologue playback systems, can hopefully appreciate how "pre-echo" and other phenomena of jitter are now being effectively tackled. That, coupled with high bit/sample rate music content and high quality digital playback components provides the contemporary approach to the highest quality music playback available and avoids alot of the "bugaboos" associated with top shelf vinyl playback (high cost, high set-up and maintenance requirements, highly specialized knowledge requirements, etc.) Also, we all know that the availability of digital music content sources are going to increase, while analogue/vinyl music sources are going to remain at the margins - presumably this will help with relative price/selection optimization for digital music content over time.

Again, I thank you for this good, basic article and hope to see more, regularly scheduled hi-def digital music system/content information for those of us that believe that this is the future of our hobby.

Kind regards and happy listening!

Mr Plus -- Thu, 02/11/2010 - 15:52

Thanks for you feedback. I agree, jitter is something that needs addressing (especially as many think - incorrectly, IMO - the whole topic has gone away, thanks to sample-rate conversion), but I'd say is a feature in its own right.

Alan Sircom
Editor, Hi-Fi Plus Magazine
London, England
editor [at] hifiplus [dot] com

Perry (not verified) -- Thu, 02/11/2010 - 21:58

I waited two years for articles such as this. Finally Olive came out with the HD4 and I have a great music server! Thanks.

Blaine (not verified) -- Fri, 02/12/2010 - 21:52

thanks - this is very helpful. I would like advice on the optimal solution for 1) a multi-room complex consisting of three lodges next door to each other with 15 listening rooms (sitting rooms, bedrooms etc) and 2) three sites in different parts of the world. Ideally for the first situation, one would have one server and for to second situation one would be able copy and synchronise music (and video?) files to the different sites.

Mr Plus -- Sat, 02/13/2010 - 07:06

There's a useful solution to both called NetStreams. The company specializes in AV over IP. It's not the sort of thing normal humans can (or would want to) do, so it's best to see if there's a custom installer in your region. The audio quality is good, but it integrates well with NaimNet products if you want the sound to move up a gear.

I'd search for the NetStreams site and CEDIA (Custom Electronic Design & Installation Association), for a good nearby installer.

Alan Sircom
Editor, Hi-Fi Plus Magazine
London, England
editor [at] hifiplus [dot] com

Mr Plus -- Sun, 02/21/2010 - 19:03

@Bob Parish:

Unfortunately, iTunes does that to AIFF files. No idea why... the secret is buried deep within Apple.

The 'easy' solution to this is to convert the files in another program first - Max if you are a Mac user or dB Poweramp if you use PC/Linux (with Wine). You can transcode WAV to AIFF and retain the 24/96 or 24/192 files, then get all the metadata advantages to AIFF.

Hope that helps!

Alan Sircom
Editor, Hi-Fi Plus Magazine
London, England
editor [at] hifiplus [dot] com

Mr Plus -- Sun, 02/21/2010 - 19:09

@ Malcolm Parks:

Thanks, that's the plan too. Not forgetting the rise of the memory player (I just got a Bladelius Embla – now that's a really fascinating cross between computer audio and traditional hi-fi).

Keep watching the site, as I expect to put up a First Look at the new HRT Streamer Pro USB DAC very soon...

Alan Sircom
Editor, Hi-Fi Plus Magazine
London, England
editor [at] hifiplus [dot] com

John of old (not verified) -- Fri, 05/21/2010 - 16:26

Hi Nobby
I'm several weeks behind on this, just doing a browse on some subjects and stumbled upon this topic and alas I'm going to be picky.
CD is not composed of wav files, cd is is composed of 16 binary digit, 44khz CD-DA. (22khz bandwidth)
WAV is an audio waveform at 16 binary digits and 48khz (24khz bandwidth)
It is not possible to extract 16/44 CD-DA, but all grabber software can extract as 16/48 (or variable bitrate/sampling) wav/wave (It's all about not stepping on toes).  And though in a technical sense a wave could be termed a form of upsampling - 44khz to 48khz, there is no interpolation of extended data so as long as levelling is not applied then there should be no discernable difference between a 16 bit CD-DA and a 16 bit wav, but they are not the same.
As for Apple being the choice for computer audio, oh sigh. Check out me as "fed up lurker" for my views on that..
I'll stop there, I just wanted to say Hi and wish you well.
 

jeremytaylor -- Thu, 07/22/2010 - 23:09

As with many applications a, software and hardware associated with computing, it just does not stop! For the novice user it is intact the most daunting and often confusing time =. There is no solution to this as the product range is hug he and some are from different operating platforms. The simplest approach is to develop a trusting relationship with an audio supplier and accept their advice.

All content, design, and layout are Copyright © 1999 - 2011 NextScreen. All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.