So it turns out that Neil Young in raising money via Kickstarter to develop a high resolution music store and player (‘Pono’), based around 192kHz, 24 bit audio. The Skeptic’s Guide to the Universe had an interesting discussion on in it their most recent episode. Well worth a listen. Here are a few thoughts I posted on the SGU forum:
Interesting about the audio stuff (which is my bread and butter). High sampling rates and high bit depths are valuable in recording and production, but pretty useless in delivery. Having said that, it isn’t quite the slam dunk Steve suggests. Nyquist tells us that a sampling rate of double the maximum frequency which you seek to record is all that is necessary. So why not 40kHz sampling rather than 44.1kHz or 48kHz? A naive analogue to digital conversion of the original leads to artefacts from any ultrasonic content being reflected back down into the audible band. So a low pass filter is applied prior to the conversion. In the early days that was analogue. To use a steep filter to reduce the signal level enormously between, say, 20kHz and the CD-standard Nyquist point of 22.05kHz resulted in tons of phase shift. 48kHz was a bit better. Most of those technical problems, though, are pretty much under good control these days. We’ve come a long way.
In the episode there was some unfortunate conflation between different things. Neil Young apparently wants 192kHz, 24 bits. CD standard is 44.1kHz, 16 bits. I don’t think anyone has yet conducted a sound study capable of demonstrating an improvement in sound from the higher resolution. But what most of us listen to on our iPods isn’t the CD standard. It is a lossily compressed version of the CD standard, usually in MP3 or AAC format. Modern portable players do support lossless (typically ALAC for Apple products, FLAC for others).
Incidentally, Monty overblows possible negatives of 192kHz, 24 bit recording. The problem is pretty much that it’s a waste of space, not that there will be audible intermodulation and similar artefacts. If you play the test audio he provides in the article you may well hear stuff, but the two ultrasonic spikes in his test, for example, peak at -3dBFS. Actual ultrasonic music content at those frequencies — if any, and generally there won’t be any — will be way down somewhere below -80dBFS. They can intermodulate all they like and you’re never going to hear it.
Now Monty has made a good case elsewhere at Xiph that using a good quality MP3 encoder (LAME) set to variable bitrate (VBR) running at about 192kbps (compared to the 1,411.2kbps of the uncompressed audio), the sound cannot be distinguished from the original. I don’t disagree. However the great majority of content on portable players doesn’t meet that standard. Stuff encoded to MP3 back in the 1990s was frequently pretty poor.
One last thought: Neil Young’s quest for 24 bit, 192kHz probably will yield improved quality for catalogue recordings. Not because of the high resolution, but because some of the stuff will be remastered from the original source material using modern technology and engineers concerned with producing a premium product. That’s why DVD Audio and SACD discs typically sounded better than the original CDs.