We took advantage of a newly-commissioned test system on which we ran all tests. Timings obtained are of course a function of the underlying processor, memory, motherboard, DVD drive and other system components. The numbers we will shortly provide serve best as relative measures to compare to one another, rather than as predictors of how other systems might behave under the same test load. We also deliberately chose what you might call a "middle edge" system - that is, one that is neither on the leading nor the trailing edge of PC technology - as more likely to be representative of the kind of system a typical user (whatever that really means) might employ. The following table shows the key characteristics of the system we used for this test.

As modern systems go, this one is certainly adequate but by no means a screamer. We'd expect those who own Core Duo or Core 2 Duo systems, or Socket AM2 based systems with 4200+ processors or better, to experience at least 20% better performance than what we report here.

Test Design

To get a sense of how the various packages do their jobs when burning DVDs we used identical Memorex DVD-R media for all three burn tests, and compiled a collection of audio, DVD video, and TV capture files so as to force each package to exercise its ability to deal with different formats. To that end, we collected a set of standard DVD files (about 168 MB of video material with embedded Dolby Digital soundtracks), a 345 MB segment of TV footage from the 2000 Olympics in Windows Media Center dvr-ms format, and a 4 MB audio file in Windows Media Audio (wma) format and put all of them in a directory tree to make it as easy as possible to burn these files to a DVD. The next table provides a complete listing of all the materials used in this test

This collection of content is large enough to exercise the burning software and to invoke transcoding behaviour in both of the commercial products. Nero transcoded the dvr-ms file into Nero Digital video format, while Roxio transcoded it into plain-vanilla MPEG-2 format, and Infrarecorder left all source materials alone.

Test Results

Even though the two commercial products transcoded the TV capture video from one format to another, both did so as a separate step prior to beginning their data burns. We were pleasantly surprised that both packages also outperformed Infrarecorder in overall burn time. The following table describes our test results.

The difference between Nero and Roxio is slightly more than 6%, where Nero outpaces Infrarecorder by just over 24% and Roxio by just under 20%. Though Nero appears to enjoy a slight performance edge, it's not significant enough to make prospective buyers lean one way or the other, and is certainly not compelling enough to induce a satisfied Roxio customer to switch to Nero (or vice-versa, for that matter).

Although we didn't time the transcode intervals for the two commercial packages, our observation during repeated runs was that if you added this time to the actual burn time, which we did keep track of, all three packages would finish in a dead heat because Infrarecorder did not re-encode the Windows Media Center video file, but the other two packages did.