Behind The Numbers Page 4
Maximum output level is also not something to base a purchase on. We measure it and publish the results on the off chance that you'll do a head-to-head listening comparison between players that deviate enough from the typical output level of 2.0 volts to make one seem better than the other for that reason alone. If Player A has a maximum output level, say, of 1.88 volts and Player B's is 2.12 volts - a 1-dB difference - Player B will almost always be preferred in a listening comparison regardless of any other measurable differences. Since you can compensate for below-par maximum output by turning up the volume setting slightly, it's not a significant "defect" unless the output level is very low, because raising the volume will also increase the background noise.
Frequency response was of supreme importance in the heyday of the LP record. It still is important, but since so few digital disc players we've encountered deviate from audible - if not measurable - perfection in this respect, it is almost a moot point now. Deviations of 0.1 dB or less from 20 Hz to 20 kHz in any audio frequency-response measurement can be considered audibly perfect. Larger deviations may occur at both ends of the frequency range, especially with cheap or portable players, but unless they are very large (greater than 1.5 dB), it's unlikely they'll be audible either.
For all sampling rates above 44.1 kHz, the ideal upper-frequency limits for DVD-Audio performance are precisely half the sampling rate, as shown in our sample lab box. None of the DVD-Audio players we've tested so far get up this high - in fact, all of them cut off at virtually the same frequency and with the same decibel variation. That's because many of them are using the same digital-to-analog (D/A) converter chips! Not to worry - if you're old enough to read and understand this article, you can't hear anything above 24 kHz anyway, and probably not even above 18 or 19 kHz.
Measurements of audio distortion (THD+N, or total harmonic distortion plus noise) are always stated in percent mainly out of tradition. I have yet to see a DVD player that has an audibly significant distortion percentage - greater than 0.1% if you're really critical, or 0.3% if you're not - certainly not at full output level (0 dBFS) and not even at our reference level (-20 dBFS). If you're going to use a player's analog outputs, anything more than 0.3% distortion should disqualify it from consideration, regardless of any other virtues it may have. But if you're using its digital outputs, the only relevant distortion measurement is of the digital receiver or preamp you connect it to.
To be more useful in distinguishing one player from another, distortion should be measured like noise level, in decibels below some reference, and then you could compare the two figures (ideally, they'd be within 3 dB of each other). But that is a really radical concept to some, who take (false) comfort in the seemingly more friendly percentage concept. The problem is, perfect distortion performance is not zero but is limited by the resolution of the test signal.
For CD players, theoretically ideal distortion performance should be 0.00153% at 0 dBFS and ten times worse at -20 dBFS (0.0153%). For DVD-Audio players, the test disc we're now using, which is the only one I've been able to track down so far, doesn't have suitable signals for distortion measurements (but I hope this will be remedied within the year). In any case, lower distortion is better.
Defect tracking is the most variable measurement we publish since merely reloading the Pierre Verany test disc (which has data-layer errors in ascending calibrated sizes engraved on it) can produce slightly different results each time. Nonetheless, no better "objective" test for this characteristic is available. The minimum defect-tracking required of a player by the CD standard is 200 micrometers (µm). Ideally, a player should track defects up to the 3,000-µm maximum on the Pierre Verany test disc - in any case, the higher the figure in our lab results, the better.