Somebody told me that the quality of CD media and the software package used to burn them can affect the sound quality of a recorded CD. He said that the results you get from Adaptec's Easy CD Creator are more compressed (in terms of dynamic range, not data!) than those of CeQuadrat's WinOnCD. At first I just laughed, but then I remembered reading somewhere that people had found that two apparently identical CDs could sound or 'appear' different (if the waveforms from each are analysed in a graphical audio editor). I know that the quality of media, writing speeds and the drive itself can all affect the eventual quality in terms of how well the disk will read, though I call on your knowledge to clear up whether or not these factors can affect the actual audio quality of a CD.
Will B
Editor Paul White replies: Software should only affect the sound of recorded CDs if it does some kind of processing, such as sample‑rate conversion, gain change or a deliberate DSP process. Most editing packages allow you to edit the data without 'changing the numbers', and providing you do this, the software has no mechanism for affecting the sound other than by introducing errors. However, your soundcard can affect the sound if you go in via the analogue inputs (noise or interference can often be added at this point) and even the digital input can produce problems if there's significant clock jitter in the system. The digital link between your source player and the soundcard may also be a source of trouble if the player has a poor clock or if you use substandard digital cable.
Jitter also tends to get blamed for numerically identical CDs sounding different, because if you're storing the right data, one way it can be made to sound wrong is by introducing timing errors (right data but at the wrong time). Jitter leads to a subtle degradation of audio quality and many listeners also find it blurs the stereo imaging of a mix. Media does make a big difference too, as some cheap unbranded CD‑Rs barely work on conventional CD replay machines, and I suspect this means the error correction is working overtime. Once the error correction has to conceal serious errors by 'guessing' missing data, the sound suffers.
The write speed of the media is also important, as CD‑R disc coatings are optimised for certain laser exposure times, and this of course varies with speed, even though some automatic adjustments are done by the recorder. I've had no obvious problems recording at up to 4x speed using quality branded media, but unbranded media is often rejected by my burner as being unsuitable, confirming the notion that it's better to buy branded product.
Another theory concerning identical CDs sounding different is that if the disc isn't absolutely concentric, the laser servo motors have to work overtime to keep the reading optics aligned. In some cases, this can lead to the control electronics modulating the ground currents in the audio system, in effect superimposing noise and distortion on the wanted signal.
The jury is still out on the absolute causes of these perceived differences, but essentially, properly designed software shouldn't affect the sound unless you want it to.