In all honesty, I started this blog post with the title “Audiophile Myth no 4” pretty sure about the outcome. And then learnt and I changed my mind. Life´s good. The real story is that a few snake oil sales men and women out there, are giving a bad reputation to genuin apothecaries.
A lot has been read and heard recently about high resolution – high definition audio. Neil Young´s somewhat desperate infomercial on Kickstarter last year promoting his Pono player certainly raised some industry eyebrows. Portrayed as life altering, belief shattering, near religious, experience, enrolling everybody with two ears and a name in the industry to burst into tears over the profoundness of what they just witnessed was clearly over the top. Did Pono have a letter r in its name that only I wasn´t seeing? The tale of “The Emperor´s New Clothes” came to mind. I was also bothered by the use of a crowd-funding platform for something that was apparently vastly overfunded from the start, but that is a different story.
I have a rather large area of my brain dedicated to the detection and avoidance of snake oil and these snake oil neurons were firing in epilepsy at the time. Being raised by two doctors I know too well that a placebo sugar pill can very much kill cancer if we only allow ourselves to believe that it might. Now killing cancer is good, and if sugar pills can do that, let´s all sugar up. A lot. By analogy, if we feel better listening to music in a different way, then let´s all listen more in that way. No questions asked. "But what is the science behind it all?" I asked. So I decided to go on another of my quests: How much resolution is enough resolution.
I remember very well when CD players were first introduced back in the 80ies and people were still very much trying to get their heads around the sheer possibility of music being expressed by numbers and not by a wave encoded in some mechanical or magnetic format. At the time the ability of a CD player to “oversample” became a highly valued attribute. As we are accustomed to think in time frames much longer than those of the digital realm, a sudden jump from on step of signal to the other was an appalling thought. This must necessarily sound “bumpy” as opposed to “smooth”, something had to be done. We were told about “over-sampling” a process that basically introduces fictitious steps between bumps to turn a ledge into a staircase. Intuitive it was it was also highly marketable.
The problem at the time was that chipsets were slow and hard to make and oversampling took resources, which in return provided a welcome opportunity to charge more for a device. And many of us paid.
Once the oversampling craze was over and many more inquisitive people realized that bumps can, and are, also smoothed out by the electronic filters found in all amps or at a loudspeaker driver level that simply will not move in small incremental steps, just because that is what the signal demands. It has weight, it has inertia, it has inductivity, it´s a real thing.
We were then told that smoothing the bumps by oversampling was only a cheap way out, and we needed real additional life changing data not fiction, and were thus presented with the SVCD. This was a CD really developed to store Video and a precursor of the DVD, but of course, if we could use all this additional data that can be stored on such a disc and use it for sound alone, what a wonderful world it would be. And hence the story of Hires Music began.
Needless to say, the SVCD has gone the way of the Dodo, and now we are in a similar discussion but this time about different digital formats.
Now to be clear: Hires has nothing to do with Lossless. “Lossy” music formats were a more or less dreaded necessity at a time when digital storage was bulky and expensive but now that memory has become so cheap and abundant and streaming services that take care of storage for us are so readily available, there really is not much reason to loose anything, as little as it may be.
But HRES / HiDEF is a different story. This is about resolution, not compression, and this is where finding a more subtle, honest truth in our current global shouting match for media attention is a more tender task.
When it comes to being honest our industry there are few people I trust as I trust David Chesky with Chesky Records and HDTRACKS. I went back to speak to him to get the low-down on the issue to see what´s the truth behind the hype.
Resolution in music can be compared to resolution in cinema and TV, David explained. The difference is that instead of many little dots up and across on a screen, we now look at time and signal strength as our coordinates. We can probably all agree that watching a full resolution IMAX movie on our smartphone is a vast insult to those concerned about Internet bandwidth. Similarly, as David Chesky put it, Kubrik didn´t film the space Odyssey on 70mm, for us to watch it on those TV sets that was available at the time he filmed it. So transferring this idea to music it is obvious that for HD / HR to make any sense at all, the equipment that is used to express it has to be rather good at expressing fine detail. This is of course the part that interests me most in my quest to design and produce such equipment.
As David explains, listening to Hires music on 10 dollar ear buds makes no sense. But what does?
Without getting into too much technical detail here, it is time to introduce Mr. Harry Nyquist, a man who knew a thing or two about signal analysis in his time. Nyquists’ criterion states that for any signal to be captured unambiguously it has to be sampled at least twice as often as has the highest frequency it contains.
Music of standard resolution is usually sampled 44´000 times a second. This would, in theory, mean it can express frequencies up to 22kHz. 22kHz sounds high enough, to someone middle-aged and male like me as I certainly don´t have much hope at hearing that high, but Nyquist wasn´t much concerned about the quality of a signal, he was only concerned about the limits of detection. A signal sampled only twice per waveform might be recognizable by machines, but we can probably all agree that such See-Saw wave form might not exactly be music in our ears.
And this raggedy waveform is then furthermore filtered and filters tend to respond poorly to sudden changes, and all of a sudden, what seemed comfortable far removed from human hearing at 44kHz makes it well into the realm where it can come to bother us. I personally can certainly hear the additional quality of a higher sampling rate, but it has to be clear, that for this to stand out in any significant way, the audio equipment has to be rather capable of expressing very fine detail. Luckily we believe we make such equipment.
But how much sampling is sampling enough? David´s answer is clear. Anything above 96kHz is clearly overdoing, provides no audible added value and should be considered marketing hype.
Now what about the other axis of resolution, the resolution of the signal itself? Standard music is encoded with 16 bits, that´s a number consisting of 16 ones and zeroes and in human readable form 65535 possibilities of being a “little loud”. At a first glance 65535 seem like a large number, but consider that anything in sound and music is vastly logarithmic meaning that what we perceive as a linear increase in loudness is really exponential in signal strength and energy. So between the softest whisper and the largest roar of a cathedral church organ there are many orders of magnitude of difference in signal strength. 1 trillion times more to be exact and all of a sudden 65535 seems not so large a number anymore.
Most music really can live with this resolution rather well, as it goes through what recording experts call dynamic range compression to where the volume of the loudest portion of a song is reduced, while that of the softest part is enhanced.
What about the purist though, those audiophiles who really want the experience of music as it first happened? For those, 16 bits is pushing it a little and for well recorded music played on very good speakers or headphones with high dynamic range, 24 bits with its 16.7 Million different possibilities does lead to an audible improvement in spaciousness, depth and detail.
So at this point, for anyone who has a knack for audio and likes things perfect, lossless, 24 bit 96kHz is a great format to have, and together with the exceptional binaural recordings available from Chesky Records HDTRacks this really is a great offering that with tangible and enjoyable improvement over what we are used to.
Listen to a Pono at 192kHz and break down in tears of joy? Cry if you must….