The Retiring Mind, Part V: Listening and Ground Truth

Human hearing is limited in the range of frequencies that we can discern. Generally, at the high end, that limit is around 20kHz, which is a very high pitch indeed. But, as we age, our high frequency perception reduces as well, until we may very well have difficulty hearing 8kHz or understanding human utterances in old age. You can test your own approximate limits with a simple YouTube video that raises pitches quickly up through the spectrum. I’m capping out at just north of 13.5kHz using a cheap speaker attached to my monitor, and with normal but quiet ambient background noise.

The original design of the Compact Disc by Phillips and Sony used the 20kHz limit as guidance for the encoding of the digital information on the disks. Specifically, the input analog waveform was sampled at a resolution of 16 bits 44.1kHz, which gives a maximum volume range of 2^16 (96dB) and supports the Nyquist sampling theorem that requires double the maximum frequency of the input stream in order to reconstruct that stream.

And CDs were very good, exceeding the capabilities of vinyl or cassettes, and approaching the best magnetic tape capabilities of the time. They also had some interesting side-effects in terms of mastering by freeing bass frequencies that had to be shifted towards the central channel on vinyl in order to avoid shortening recordings unduly because of the larger groove sizes needed to render low frequencies.

But now, with streaming, we can increase our resolution still further. Qobuz and Tidal offer Hi-Res audio formats that can range up to 24 bit resolution at 192kHz sample rates. Tidal also promotes MQA (Master Quality Authenticated) format that may use lossy compression but preserves aspects of the original master recording.

Are these better than CDs? Objectively yes, in that more of the original signal is available. But can we hear the difference? It’s not altogether clear, and actually part of a whirlwind of opinions and attitudes in the audiophile community. There are several reasons to doubt that we can really discern the extra detail in many Hi-Res streams. First, and foremost, are the limits of our hearing. Second is the rendering chain that goes from the streaming data through a sequence of digital-to-analog conversion (DAC), then through amplification (often with a pre-amp and a power amp), and then finally through transduction into sounds using some kind of speaker system. Each of these components can modify the sound away from what it was originally. Third, there is the room in which one listens to those speakers. Is it treated to reduce sonic reflections? Or are there diffusers to provide more reflections? Bass traps? What is natural at that point? Fourth, what does “originally” even mean? The recording and mastering of a live or studio performance involves its own chain of components, as well as the deliberate enhancement of the sounds to achieve different goals, like soundstage, centering, tonal balance, channel separation, and so forth.

In the audiophile community, people get obsessive about their reproduction chain, including hard-to-fathom ideas about expensive wires, metal disks to vibrationally isolate components from resonant coupling, and even power conditioners to reduce power-supply frequencies from sneaking into the noise floor of their amplifiers. Many will swear that they can hear the tonal colorations of these components, and many will also swear that more expense for new technologies that sport exotic theories results in greater listening pleasure. The alternative is that most of that is wishful thinking driven by a desire to justify the expense of such components.

Still, there is some research on at least the value of Hi-Res versus lower rate, lossy compression schemes. In that meta-analysis, there was a small statistical effect where people could discern the properties of better streams, though it was largely associated with sample rate (44.1kHz versus 192kHz) rather than quantization bit width (16 bits versus 24 bits; a caveat here that most 24 bit schemes only use 21 bits). Even so, one can buy modern DACs that can handle 32 bits at 700+ kHz!

So, given all of that, what sounds best? I approached my audiophile engagement with a kind of neutral skepticism, as is my nature and the fact that I had little experience with high-end audio systems. I was gifted one of the first Sony CD players for my high school graduation, along with a copy of Pink Floyd’s Dark Side of the Moon as well as a set of Beethoven’s symphonies. I remember that well, and how with headphones on it sounded better than anything I had ever heard before. But then decades passed and car and personal audio systems came and went. It was nice to just be able to find music I wanted to listen to and the time to do so.

So I recently decided to do a deep dive, but without spending too much dough (though opinions on this will vary). I assembled two systems. First, I combined Cambridge Audio Minx 22s with a CA subwoofer and a Peachtree Audio Decco125 SKY for a living room. The room is untreated, has a 20’ ceiling, windows all around, and travertine flooring. Not ideal! Spare furniture and a rug help, but just barely. The Minxes are up in the corner due to wiring supplied with the house. The Decco125 SKY has a 24bit/96kHz DAC and can stream Qobuz and iHeartRadio, as well as a other sources like Tidal and Spotify. It can also do Airplay (though not Airplay 2). The software has not evolved very well, but the Qobuz integration serves my needs. The amp is a class D system rated at 125 W into 8 Ohms, which is more than adequate for the little Minxes, and which are notable for their planar tweeters that ostensibly radiate more uniformly than dome tweeters.

And how does it sound? The room is huge and highly resonant with echoes that seem to have their own echoes. Moreover, the room extends in modern, open floor plan ways, past a fireplace, through a dining room, and into the kitchen and family room, giving a tunnel for echoes to work down and back. Still, for certain types of music, I like the listening experience quite a lot, spending part of every evening in there. Evening listening is typically classical from KUSC Los Angeles, WGUC Cincinnati, or WQXR New York. And in that space, most recordings with limited bandwidth (128 kbps, KUSC and WQXR; 192 kpbs WGUC) sound fairly good. Noise floors appear, especially with WGUC, but the cavernous nature of the room resembles a live performance space. Piano takes on a liveliness while voices become hauntingly expansive (Steve Reich’s Proverb is worth a listen in that room). Qobuz and Hi-Res are mostly irrelevant.

But it is too alive for much listening. The details disappear in clashing echoes. The space itself is an instrument. So I had to try something different. In my office, then, I decided this year to build out a radically different set-up. The space is smaller, 12’ x 12’ or so, carpeted, and with fewer windows. The walls are standard dry-wall but with a heavier finish. I first acquired a set of natural cork panels (40” x 20”, 60 dB SPL) via Amazon and mounted them on a hardboard backing surface to work towards a very neutral environment. There is also a soft daybed in the space and a central listening chair, both of which help deaden the space.

The speakers are Tekton Double Impacts, which are admittedly huge at 100 lbs each, but are consistently reviewed as spectacularly neutral in their presentation. These are, in turn, powered by Schiit Audio class A monoblock Aegirs, fed from a Schiit Freya S via XLR cables. The DAC is a Schiit Yggdrasil capable of 24 (eh, 21 really) bits at 192kHz, which is, in turn, fed by a Bluesound Node 2i. I tried a Sonos Port, which had better Amazon Alexa integration, but it wouldn’t even pass along Hi-Res streams, instead truncating them to 16/44.1. The streaming sources are the same, Qobuz and internet radio, with Apple Music as a backup. The Node 2i is Airplay 2 and the Sonos Port is Apple Music-embedded.

And the sound is, well, very different from the living room experience. There is little to no echo and the speakers provide very detailed stereo imaging. I’ve collected together demo lists that piece together Hi-Res and notable CD-quality audio recordings. Some are spectacular, others are interesting for other reasons, from historical to innovative instrumentation or recording techniques. The class A monoblocks are well regarded, though perhaps a bit overspec’d in terms of their power delivery (20 W/channel into 8 Ohms, 40 into 4; I’m monoblocking them into 4 Ohms, which is not spec’d by Schiit, but should deliver 80 W and I am never short of sound), and the Tektons may have the same issues of having slightly ridiculous claims of 100+ dB sensitivity, though I haven’t found any reliable tests of the Double Impacts yet.

With very Hi-Res streams recorded and mastered in recent times, noise is non-existent and imaging is perfect in a vast soundstage. I have experimented with several different speaker placement strategies, most recently setting them to point to an apex five feet behind my listening chair. I also switched the monoblocks left and right, which corrected a slight left skew of the image center. Why that should work is not clear. It could be component mismatches between the two monoblocks, or between the two speakers, or some kind of exaggeration of the properties of one by the other. But the switch fixed that in any case.

Below are some of my recent demo streams from Qobuz. They are mostly 24/96, with a few at 24/192 like the Norah Jones, and a few at CD quality.

So what’s the conclusion? Can I hear any differences with Hi-Res streams? Well, I only tried very mild A/B testing (switching back and forth between versions), and certainly did not do the more sophisticated ABX testing that requires instrumentation! I’m not sure, quite honestly. I think the room treatment and speakers are the critical parts of the coupling of my hearing to the music. I’m sure a less capable system would be a lesser listening experience, but whether the bit rates matter I’m less sure about. Headphones are one way to isolate out room dynamics even more, but the rendering chain still requires a DAC and a headphone amplifier, so I’m not sure we can ever really get to some kind of ground truth in the matter of what sounds best.