Episode 26: High Fidelity

HiFi over WiFi. This episode of Conversely we rant about Neil Young’s Pono music player, a string of recent acquisitions by Beats and Spotify, and we wax rhapsodical about SXSW of yesteryears.

Subscribe (iTunes)
Subscribe (Stitcher)
Subscribe (RSS)

[beginrant]

And since Mike kind of forgot to say this during the Pono section, here’s one really important reason why Pono doesn’t matter: 192kHz audio is bunk. Well, not technically, since it obviously has some uses, but when it comes to music, there really is no need.

See, there’s this thing in digital signal processing called the Nyquist rate. The Nyquist frequency of 192 kHz audio is half, or 96kHz. In other words, if the sampling rate of a FLAC file played through the Pono is 192 kHz, the audio output will include all frequencies up to 96 kHz.

Well that all sounds well and good, except that you can’t hear it. Human ears typically max out at around 20 kHz – anything above that is essentially a dog whistle, a sound that might be reproducible but you really can’t hear it. And as you get older, your hearing range shrinks, especially for men.

CDs were designed with this in mind. They have a standard 44.1 kHz sample rate at 16 bits per sample. This faithfully reproduces frequencies up to 22.05 kHz, which covers the entire spectrum of human hearing.

In other words, there’s really not much need to sample audio any higher than 44.1 kHz. So why does Pono go up to 192 kHz?

Well some people say that while we can’t hear those higher frequencies, we feel them. I say this is bunk, and most experimental evidence says the same. Others will say that those higher frequencies are inherent in all sound – for example, a violin actually creates some very high overtones that get lost when we record/playback at lower sample rates. I say this is bunk too, as the vast majority of frequencies we associate with music fall within our normal hearing range. Moreover, our brains likely can’t process those sounds, and certainly don’t feel them as part of music.

But here’s an even bigger reason 192 kHz audio is bunk: almost no music whatsoever is recorded at these frequencies. Most digital audio workstations – the ones used in every digital studio in the world – use standard sample rates of 44.1 kHz or 48 kHz. They are capable of higher, but few musicians chose to use those higher sample rates, as it bogs down processing, taxes disk storage, and is generally just unnecessary, especially given that CDs are ubiquitous and they are fixed at 44.1 kHz. Even if Neil manages to license the original bit-for-bit audio masters of your favorite records of all time, they’ll still be 48 kHz.

There’s a case to be made for converting analog to digital at higher sampling rates, but at the end of the day, you’re probably playing back at much lower rates.

Here’s another thing: most audio systems can’t reproduce frequencies higher than standard hearing ranges. They just aren’t designed to do that. The Sennheiser HD 800 headphones, their top of the line cans priced at 1500 bucks, are rated for frequencies between 14 Hz and 44.1 kHz, with topline frequency response at 51 kHz. And while they can impressively reproduce audio from 96 kHz source, this means that even the best headphones in the world can’t reproduce the sound of 192 kHz sample rate audio.

Don’t get me wrong, I love the idea of increasing audio fidelity – MP3s and AACs really don’t sound very good compared to lossless sources – but I can’t make any real case for the level of audio in the Pono. Give me lossless versions at 48 kHz sampling, 24 bit, and throw a decent DAC into the iPhone, and I’ll be a happy man.

[/endrant]

Show Notes:

Episode 26: High Fidelity

Episode 25: Summon The Heroes