Views: 832

Written by:

Audio Data Compression: Myths and Controversy Part 2 (Lossless)

Uncategorized

This post is part 2 of 2 on audio data compression. The first part focused on lossy compression whereby audio quality is decreased and quite a bit of size is reduced. This post will focus on lossless compression whereby quality is retained and some size is reduced.

As I mentioned in the previous post, there are folks out there who are paranoid about their audio. They feel that anything that touches their audio will degrade the quality. These fears are unjustified in the case of lossless audio compression. Whatever goes into lossless compression is exactly what comes out of lossless compression. In fact, with most modern lossless formats, if you were to compress a raw audio file into a lossless format and then uncompress it back, the two files would be exactly the same, bit-for-bit. As a result of this integrity, lossless compression is perfect for applications where quality is more important than space or bandwidth like archiving or transfering masters. Many folks, myself included, keep lossless backups of their CDs so that they can recreate the audio from the source discs should anything happen to the originals. There also is an added benefit of being able convert these lossless files into any lossy format without having to re-rip the source media. It’s very convenient. But, I digress. For pro audio there are two main applications for lossless compression: archiving and the transfer of masters.

The Formats

Because of the nature of lossless compression, the actual format is not as important as it is for lossy. Since one can always recreate the source file exactly, one can convert any lossless format into any other lossless format with no problems. The areas where the formats differ are features such as checksumming, streaming, compression strength, licenses, and hardware/software support.

Some of the major formats are FLAC, WavPack, Monkey’s Audio, Windows Media Lossless, Apple Lossless, and Shorten. For a long time Shorten was the standard because of its heavy use in online concert-sharing communities (the digital equivalent to tape-swapping). Shorten has since then pretty much been left behind because its compression strength hasn’t held up to its competitors and because it’s features aren’t as robust. The others vary mostly by speed of compression/decompression and compression strength. I won’t even bother listing the individual pros and cons here because Hydrogen Audio has already done so perfectly.

This table is yoinked from there verbatim:

Features WavPack ALAC FLAC Monkey’s OptimFROG Shorten
Encoding speed very fast average fast fast slow very fast
Decoding speed very fast fast very fast average average very fast
Compression* 58.0% 58.50% 58.70% 55.50% 54.70% 63.50%
Flexibility** very good bad very good very good very good bad
 
Error handling yes   yes no yes no
Seeking yes yes yes yes yes yes
Tagging ID3/APE QT tags Vorbis tags ID3/APE ID3/APE no
Hardware support limited good very good limited no limited
Software support good bad very good good average very good
Hybrid/lossy yes no no no yes no
Replay Gain yes sort of yes no yes no
RIFF chunks yes   yes yes yes yes
Streaming yes yes yes no yes no
Pipe support yes no yes yes yes yes
Open source yes Yes (decoding) yes yes no yes
Multichannel yes yes yes no no no
High resolution yes yes yes yes yes no
OS support All Win/Mac All All Win/Mac/Linux All
Features WMA LA TTA LPAC MPEG-4 ALS MPEG-4 SLS Real Lossless
Encoding speed average slow very fast average average slow slow
Decoding speed average slow fast fast fast slow fast
Compression* 56.30% 53.50% 57.10% 57.20% 57.10% ? 57.0%
Flexibility** bad average bad bad very good bad bad
 
Error handling yes no yes no yes yes  
Seeking yes yes yes slow yes yes yes
Tagging Proprietary ID3v1 ID3 no yes yes Proprietary
Hardware support limited no limited no no no no
Software support good bad average average bad bad bad
Hybrid/lossy no no no no no yes no
Replay Gain no no yes no yes yes no
RIFF chunks no yes no yes      
Streaming yes   no no yes yes yes
Pipe support yes yes no       no
Open source no no yes no yes yes no
Multichannel yes no yes no yes yes no
High resolution yes no yes yes yes yes no
OS support Win/Mac Win/Linux All Win/Linux/Sol All All Win/Mac/Linux

Yay. As you can see… it’s pretty much Apple(tm) to Apples. I personally use FLAC because it works on Linux, OSX, and Windows and it supports three things that audio professionals will love: multi-channel support (like 5.1 surround sound), checksumming (if it gets corrupted during a filetransfer it’ll tell you), and high resolution support (24-bit, 192khz, etc is a-ok). There are also a good number of music players that support portable FLAC such as the iAudio I7 series and some neat-o home media center solutions such as the Sonos series or the Squeezebox series (just bought by Logitech).

Applications

Lossless compression is absolutely perfect for archiving your masters or for sending them to and from a mastering engineer over the ‘net. You’ll get almost a 50% savings in file size while retaining all the quality you need. It simply makes no sense to bother sending uncompressed wave or aiff files. To be totally blunt – any mastering engineer who demands uncompressed audio is being unreasonable. Send them to the folks at Hydrogen Audio where they will be mercilessly schooled.

Even though the math and research behind lossless compression is no less complex, the topic of what format to use and why is significantly less so. Do what you feel and what works for your needs.