With the release of Pono, quite a few journalists have decided to destroy it. I think this bad, Pono is quite a unique product in the market and I would like that it has every chance of success.
Personally, I have no intention of buying a portable high definition audio player. To be honest, I source most of my music from iTunes. Sure it’s only 44.1kHz, 16bit with 256kbps MPEG2 compression added. Sometimes I get music from other channels which use 44.1/16/320kbps MPEG1. I love it that sites like bandcamp give me the option of downloading FLAC or uncompressed at 44.1/16. This is good enough for most cases, plus I get the advantages of portability. I’m speaking as an audio engineer and a pragmatist.
The purpose of this article is to point out that there are some checks, balances and common-sense in this equation. My perspective on this issue is that of an audio engineer, (repressed) hi-fi dude, musician and DJ.
After working with various pieces of studio microphones, pre-amps, EQs and compressors; most are of a lesser grade than a real hi-fi system. It is common practice to drive some devices (i.e. compressors or pre-amps) hard so that they add a bit of extra character (i.e. distortion) in a musical way. There is no exact representation of a sound source. Every device in the audio chain adds an amount of colour (AKA distortion) ranging from subtle to overt. This is why recording is as much of an art as it is engineering. The goal is always to bring out the mood/emotion/spirit/message of the music. High fidelity matters take second place. I’m not arguing that this is wrong, I’m trying to compare the hi-fi puritan world to that of the typical studio. A hi-fi system is a minimal system of components that only do one simple thing but do it extremely well. A studio is a complex maze of signal paths feeding into multiple other pieces of equipment. It then gets converted to something else, unconverted and then pushed back into the maze again.
The point I’m trying to make here is that going too far down the hi-fi road will only end in frustration. The better your equipment, the better you can hear the faults in the recording. I can still recall the day when I realised that there was some music I never wanted to hear on a hi-fi system but loved to listen to it on a PA. This decision to go out and buy some PA gear fundamentally changed my enjoyment of music for the better. A PA, with it’s dense sound, will hide ugly artefacts introduced during the recording/mixing/mastering process and let the music come through.
Digital Audio Resolution
I’ve experimented with multiple people to see if they can hear the difference between 16bit and 20bit audio recordings. The difference is obvious but subtle. Most people cannot hear the difference between 20bit and 24bit audio. For most people 16bit is good enough. I find that a good 16bit DAC is just as musical as a 24bit recording played on a mediocre 24bit DAC. So bit depth is no big deal if your equipment is up to it.
Sample rates are a different can of worms. Yes it’s true that the human ear tops out at 20kHz. I can hear just past 18kHz. I’ve heard of audiologists who have measured people to 22kHz (but that’s rare). The point about the 44.1kHz sample rate that is usually missed is that it requires a brick-wall type low pass filter at 22kHz so that artefacts are filtered out. Depending on the filtering technique used, this will have either phasing or combing effects from 10kHz and up. This is audible. This is also why people love vinyl; it does not require this kind of extreme filtering. These higher frequencies are what add the feeling of space or depth in a listening experience. Oversampling is a way of dealing with this, but oversampling algorithms add subtly audible ringing to the signal. Hence there is always some kind of trade-off.
Ideally if the standard sample rate was around 60kHz then low-pass filters could be designed to not interfere with the audible spectrum. We do have higher sample rates available (i.e. 88.2, 96 and 192kHz). But everything over 60kHz is just a waste of bandwidth.
Accessibility and Price Often Trumps Quality
How many times have you bought fruit and vegetables from the supermarket rather than the organics shop? How many times have you bought music from iTunes rather than going to the CD store? This is the same for every human; we make a choice between availability, ease of use and price every time we purchase something. Pono makes hi-fi music more accessible and I love it for that. But I already carry around an iPhone and I don’t want to carry another device. For music on the go, my phone is my device. When I want to seriously listen, I’ll use my system at home.
Get to the Point!
I’m going to buy the bulk of my music form iTunes because it is useful for listening on the go. The quality is good enough, especially when listening on a PA system and it doesn’t chomp space on my phone. In the rare case where I know I will want focused listening to a recording on my home hi-fi system, I’ll get the high-definition version and play that on high-definition equipment. Jazz at the Pawnshop is a great example of this.
As an aside, other high definition albums I love are Mark Knopfler’s Privateering and Daft Punk’s Random Access Memories. I’ve taken a song from each, used various techniques (i.e. dithering types) to downgrade it and then listened to the differences on various DACs. My conclusions are:
- A DAC designed to do 44.1kHz at 16bit well comes really close.
- 16bit sources played on a 24bit DAC are less musical
- Anything at or above 88.2kHz at 24bit does have noticeably more air, subtle detail and space.
- Differences in dithering types or no dithering are very subtle. But I can hear why it would be important when going lower than 16bit.
- 88.2kHz at 20bit is a (slightly) higher resolution than vinyl, without added character.
- It’s easy to make a 24bit high sample rate DAC, but very hard to make a great 16bit 44.1/48kHz DAC.
- When you really enjoy the music and you are pulled in by it, this extra detail doesn’t matter.