HI-RES HI-JInkS

Are hi-res music files on a fast track to hell in a handbasket? Not only are many hi-res music files not hi-res at all, others have insurmountable technical failings. Here, Stephen Dawson looks at one tip of the iceberg…

In the rush to get on the hi-res bandwagon, some tracks with insurmountable technical failings are finding their way onto the market.

How audiophilic are audiophile discs? That’s a silly question because the answer is: ‘It depends on the disc’. So let’s look at just one in which the producers have managed to stuff up a significant part.

The DVD-Audio version of The Beach Boys’ 1966 album ‘Pet Sounds’—which Wikipedia says is ‘ one of the most influential records in the history of popular music’— is actually a hybrid disc. One side is DVD Audio and the flip side is DVD Video. Unusually for a DVD Audio disc, the version I purchased back in December of 2005 has the video—mostly the menus, but also some bonus videos—in PAL (more correctly, 576i50) format. I purchased it locally (from JB HiFi if I recall), but the packaging indicates it was produced in the EU.

The DVD Audio side has the new surround mix, and only the surround mix. This is presented in 5.1-channels in wonderfully high resolution 24-bit and 96kHz sampling. Taking something that was created in the mid sixties for mono and turning it into surround was always going to be a challenge, and the result here is badly flawed. For several reasons.

First, the surround channels carry much the same content as the front channels, and they are slightly louder than the front. Talk about aggressive surround! Second, the centre channel has almost nothing in it. Just a little vocal bleed. And as for the LFE channel, of what little there is there, quite a bit isn’t even bass.

Let’s get technical for a sec to prove the point. Let’s take one representative track to illustrate matters: Sloop John B. The right front and surround channels are at the same average level, while the surround left channel is actually 1.6dB louder than the left front. Meanwhile the average level of the centrechannel is more than 32dB lower than any of the other four.

The LFE channel is even lower: more than 45dB less than the main four. And when I play it alone what I hear is the drum kit, mostly the snare and toms, with a little leakage from the cymbals and possibly a triangle. There’s no discernible bass. That’s all in the other channels. Specifically, the strong bass line around the 40 second mark is strongest in the right front and right surround channels.

Speaking of the right surround channel, it has a problem: a number clipped peaks. These are not quite at, but are close to 0dB digital, and their fairly consistent level suggests digital clipping at some earlier point in the production process.

There aren’t many of those, and they’re almost certainly inaudible.

A spectrum of the main channels suggests that the frequency balance is maintained all the way out to about 30kHz, although what’s actually out there is anyone’s guess. It seems unlikely that a whole lot of actual HF signal would have survived forty years on tape. [One problem with storing magnetic tape for long periods of time is that the high frequencies self-demagnetise. It’s inevitable, unstoppable and irreversible.]

Unlike many DVD Audio discs, there is no two-channel mix on the DVD Audio side. Instead you get a choice from two on the other side, the DVD Video side. Remember the DVD Video specifications support two-channel au- dio in PCM up to 96kHz, 24-bit sampling. And that’s what’s provided here, for two full versions of the album (the DVD Audio side is losslessly compressed using Meridian Lossless Packing).

One is a stereo mix, the other the original mono mix. It is this last that’s closest to the original. Both show the song lyrics as they’re playing.

Now here is where things get really, startlingly odd. Both the stereo and the mono versions are encoded at such a high level that they are quite substantially clipped. Not, once again, by the hard limits of the final PCM encode, but somewhere earlier in the process. The near horizontal lines of samples—up to eight in a row— fall short of the hard PCM encoding limit by around 0.15 decibels.

There’s no good reason for that. None at all. If 24-bits of resolution is good for anything, it is surely for allowing sufficient headroom to avoid clipping.

Can I hear the clipping on this particular disc? Not that I can tell. Perhaps in a levelmatched A–B comparison with the same material without clipping I might be able to reliably detect the difference between the unclipped and clipped versions, but I doubt it. In reality, what is still fairly modest clipping is in fact very hard to hear.

The mastering engineer—Joe Gastwirt—is well-respected and has an enormous list of credits, but still… the capacity of 24-bit/96kHz signal delivery ought to be deployed to provide as pure a signal as possible. Stephen Dawson