|
|
Audio Compare
NSTL Audio Comparison Test: Fact Sheet
NSTL, Inc. conducted a consumer study comparing the audio quality of Microsoft Windows
Media 4.0 to RealNetworks RealSystem G2 and MP3. Seventy-seven
consumers participated in audio testing using both Rock and Classical music.
Download NSTL's Full Report 
Consumers were played a clip from an original CD, followed by a Windows Media
clip and a G2 clip (in random order), and asked which sounded closer to the
original. This test was performed at 20 kilobits per second (Kbps) and 32 Kbps data rates. The same
consumers were played a clip from an original CD, followed by Windows Media at
64 Kbps and MP3 at 128 Kbps (in random order) and asked which more closely
resembled the original.
Results
For tests at 20 Kbps and 32 Kbps
2 genre of music x 2 comparisons x 77 consumers = 308 comparisons
- Windows Media sounds more like the original: 249 or 80.84%
- RealSystem G2 sounds more like the original: 59 or 19.15%
For tests at 64 Kbps Windows Media and 128 Kbps MP3
2 genre of music x 77 consumers = 154 comparisons
- Windows Media at 64 Kbps sounds more like the original: 79 comparisons or
51.29%
- MP3 at 128 Kbps sounds more like the original: 45 comparisons or 29.22%
- Could not distinguish between Windows Media and MP3: 30 comparisons or 19.48%
Methodology
Four multimedia systems
were set up in an office-type environment. Each system was equipped with a
Sound Blaster! Live audio card, and a Cambridge Soundworks, Inc. Microworks
CSW350 speaker system, including a sub-woofer. Audio players were installed
with the default configuration and the volume was fixed at a reasonable level
for the office environment. The test subjects were seated away from the test
system and were not able to see the computer screen. The two speakers were
positioned approximately two to three feet away from the test subjects head.
The speakers were pointed toward the test subject. The subwoofer was placed on
the floor approximately five feet from the test subject, with the bass tube
facing toward the test subject.
The original audio clips were PCM WAV files at 44.1 kHz, 16-bit stereo,
resulting in approximately 1,379 kilobits per second. The MP3 audio clips were
encoded using 128 Kbps constant bit rate (CBR) joint stereo
encoder from MusicMatch Jukebox. The RealSystem G2 audio clips were encoded
using RealNetworks 20 and 32 Kbps stereo music encoder. The
Window Media clips were encoded using Microsoft's Windows Media Encoder 4.0
using the MSAudio codec at 20 and 32 Kbps stereo music settings
for the G2 comparison and 64 Kbps for the MP3 comparison.
The sample audio clips were installed on the hard drive of the test system to
avoid any issues associated with Internet traffic and to normalize the test
process. The Windows Media Player v6.02.05.0331 was used to play the audio
clips encoded using the new Microsoft codec. Winamp v2.10 was used to play
audio clips encoded using the MP3 codec. RealSystem G2 encoded audio clips were
played on the G2 Player v6.0.5.27.
Test subjects listened to the original audio file and then listened to two
comparison audio clips. They were then asked which of the two clips sounded
more like the original. The original and the sample clips were played again if
the test subject requested. The length of the audio clips was fixed at
approximately ten seconds. The order in which the clips were played was varied
to minimize bias in the listening order.
Download NSTL's Full Report
 Back to the top
This report was prepared
by NSTL, Inc. under contract for Microsoft Corporation. NSTL does not guarantee
the accuracy, adequacy or completeness of the services provided to Microsoft or
the data included herein. NSTL MAKES NO WARRANTIES, EXPRESSED OR IMPLIED, AS TO
RESULTS TO BE OBTAINED BY ANY PERSON OR ENTITY FROM USE OF THE CONTENTS OF THIS
REPORT. NSTL MAKES NO EXPRESS OR IMPLIED WARRANTIES OF MERCHANTABILITY OR
FITNESS FOR A PARTICULAR PURPOSE OF ANY PRODUCT MENTIONED IN THIS REPORT.
|
|
