What's Going On Up There? Follow-up From January 2001

A Follow-up appeared in January 2001 (Vol.24 No.1):

After I had finished measuring the review sample of the first DVD-Audio player to reach these shores, Technics' $1200 DVD-A10, for Jonathan Scull's report in the November 2000 Stereophile, I set it up in my listening room.

Basically, I thought the Technics was okay as a CD player—not in the first or even second rank, but its sound was better than I'd expected: excellent clarity, no treble grain. Playing the two sampler DVD-As that were all that I had available to me in October 2000, the Technics' sound quality went up a notch, but not as much as I had expected. Remember, however, that I was listening in two-channel mode, and that the true glory of the DVD-A experience should be revealed when the discs are played back in full surround sound. A Stereophile writer will report on this subject in a future issue.

But one track in particular on the Warner sampler DVD-A that comes with the Technics player puzzled me in two-channel mode. I feel Steely Dan's Two Against Nature CD (Giant/Reprise 7599 24719-2) was a highlight of the 2000 recorded year, both with respect to musicmaking and to sound quality. Not surprisingly, therefore, the first track I played on the Technics DVD-A10 was the surround mix of "Janie Runaway" on the Warner sampler, downmixed on the fly by the Technics to stereo. The experience left me puzzled: While I couldn't swear the mix on the DVD-A was the same as that on the CD, they otherwise sounded pretty much the same! It's a terrible thing for an audiophile to have to admit that he didn't hear a difference, but there it was.

So, as I had just published an article examining the extended spectra of a number of original 24/96 and 24/88.2 recordings, I analyzed the frequency content of the tracks on both DVD-A samplers, as well as on some of the Classic Records and Chesky 24/96 DVD-Video discs. As the Technics redithers and downsamples its digital output to 16/48 for the DVD-A discs (but not for the Classics and Cheskys), I used a professional dCS 904 two-channel A/D converter, running at 24/96, to redigitize the Technics' analog output. The digital data were stored on my PC's hard drive, using CoolEdit 2000 and the TosLink input of the RME Digi96/8 Pro soundcard. Just in case the downmixing was changing things, I fed the dCS converter with the Left and Right surround channels, which represent the raw audio data on the disc after being unpacked. (While MLP is presumably used to losslessly compress or "pack" and decompress the data, there is no mention of MLP on the Warner disc; instead, the Technics' onscreen display cryptically shows "PPCM," for "Packed PCM.")

The first graph shows the spectrum of analog pink noise as encoded by the dCS, something I did to check that the dCS was working properly. The spectral content falls, as expected, at 6dB/octave—which is why, on this linear-scaled graph, the rolloff decreases with increasing frequency—up to 40kHz, above which the dCS's anti-aliasing filter rapidly rolls off the signal to below the -140dBFS floor of the graph.

Fig.1 FFT-derived spectrum of analog pink noise, digitized at 96kHz with a dCS 904 ADC (linear frequency scale, 0Hz-48kHz, 6dB/vertical div., left channel cyan, right channel magenta).

The first 24/96 track to get the treatment was Terry Evans' "Too Many Cooks," from Blues for Thought (Classic DAD 1014). Fig.2 shows a three-dimensional spectrogram of the final 90 seconds of this track, with the vertical axis, frequency, extending up to 48kHz; the horizontal axis is time; and the color of the spectral components corresponds to amplitude, with white, then yellow, being the highest in level, and dark blue, then black, the lowest.

Fig.2 Spectrogram of Terry Evans' "Too Many Cooks," from Blues for Thought (Classic DAD 1014). Original DVD-V data recorded at 24-bit/96kHz.

Although the original master was analog, this recording can be seen to have considerable content between 20kHz and 40kHz, revealed by the vertical red and magenta lines. (As demonstrated in the October article, this is primarily due to cymbals and electric guitars.) The dark blue flecks in the same region are due to analog tape noise, while the black band at the very top of the graph is due to the dCS's intrinsic rolloff above 40kHz, as seen in fig.1.

By contrast, fig.3 shows a spectrogram captured under identical circumstances for the 24/96 Steely Dan track. The vertical lines are chopped off smartly above 24kHz, with just blue flecks filling the space between 24kHz and 44kHz. (Ignore the vertical dotted yellow line, which is a marker.) Fig.4 shows a high-resolution spectrum of one of the hi-hat cymbal crashes in this cut. The musical content extends up to 20kHz, with then a steep low-pass rolloff that plateaus above 24kHz. This plateau extends up to 40kHz, and then is rolled off in the same manner as the pink noise in fig.1.

Fig.3 Spectrogram of Steely Dan's "Janie Runaway," from the Warner DVD-Audio sampler. Original DVD-A data recorded at 24-bit/96kHz.

Fig.4 FFT-derived spectrum of extract from fig.3 data (linear frequency scale, 0Hz-48kHz, 6dB/vertical div., left channel cyan, right channel magenta).