Foster's
Test Bench #3

iPod Touch
Audio Measurements
 
Foster's
Test Bench #2
Selecting CD
Storage Format
 
Foster's
Test Bench #1

Centrace MicPort;
MXL Mic Mate
 
 
 
Foster's Test Bench #2:
Selecting a Computer Storage Format for CD's

By Alvin Foster & Stephen Owades

Deciding on the right compression format to store and play back CDs led to a format investigation. The search was limited to the most popular lossless formats - FLAC and WMA lossless. The goal: to adopt a 'transparent' (no coloration) application to store and play back CDs on the hard drive of my new media computer.

To insure complete transparency, Steve Owades and I decided to conduct the ultimate test, a bit-by-bit file comparison.

TEST PROCEDURE

The CD source file chosen for the ultimate test was a three minute and fifty-two second rap song. The music was selected because it has an extremely limited dynamic range and, unfortunately, represents today's most dominant, popular recording technique. MediaMonkey was used for all the hard drive recording tests.

The CD source file was first compressed and stored in FLAC and WMA lossless, and also stored as an uncompressed WAV file. The second step uncompressed the FLAC and WMA files and converted them into the WAV format. After the conversion, the three WAV files were carefully trimmed to the same starting and ending points, and Microsoft's File Compare was used to determine if they were the same. This program evaluates two files bit-by-bit to determine if they are identical; i.e., a perfect match. Happily, both FLAC and WMA lossless passed. The compressed and then uncompressed files were no different from the original CD music file.

TABLE OF RESULTS

FLAC offers variable compression rates. We selected three: Hi, Medium, and Low. The File Type describes the type of file tested. The File Size in kilobytes is the resulting size after compression. The Storage Speed in Seconds is the amount of time required for the music to be stored onto the computer's hard drive (including reading from the original CD, compression, and writing to disc). The Playback CPU Usage column gives an indication of how much the central processing unit was utilized to return the compressed files to a playable format. The numbers were obtained from the Windows Task Manager. The CPU usage differences on playback were too small to cause a modern computer to 'slow' down, or to make one prefer either compression scheme. The original, uncompressed WAV file results show the track's original size, and the copying speed without any compression.

FILE TYPE FILE SIZE
IN kb
STORAGE SPEED
IN SECONDS
PLAYBACK
CPU USAGE
FLAC Hi Compression 28,200 37 5%
FLAC Medium Compression 28,320 34 5%
FLAC Low Compression 31,273 28 5%
WMA (Lossless) 27,905 27 6%
WAVE (PCM Audio) 40,026 26 4%

Both FLAC and WMA lossless use variable rate compression. The compression speed and stored file size differ with the amount of complexity and 'silence' in the music. WMA is a proprietary file format and FLAC is 'open source'; i.e., the source code is available free to users. All the compression results noted above will fluctuate with the speed of the computer and the CD reader. For example, classical music is typically more complex than a hard-hitting rap song, and an ultra-fast CPU will compress files more rapidly.

CONCLUSION

The differences between the FLAC and WMA formats are too small to be of much significance. Both will save around 30% compared with the size of uncompressed files. The results of the bit-by-bit comparison to the original file indicate that both formats offer an inaudible difference and vary insignificantly in playback CPU usage. FLAC's open source availability is important to me, so it is my choice.

January 2008
 

updated 10/9/09

 

The Boston Audio Society
PO BOX 260211
Boston MA 02126

problems? email Barry: webmaster@bostonaudiosociety.org

updated 10/9/09