Who needs 192kHz audio anyway?

This post was first published in April 2014 at projectstudiohandbook.com/PSHforum

This week we were reminded of how lucky we are to be working in the audio, rather than video, sector. 

Although Project Studio Handbook is all about video tutorials, we are more than happy using AVCHD camcorders and off-the shelf backup solutions for our growing library of media clips and edits. We're not working at 4 or 5K, so although our video may not be as future proof as those that are, we feel the quality is more than acceptable for the type of content we produce. We have a number of 4TB eSATA connected drives, which we are nowhere near to filling. We can duplicate and keep backups of our finished edits, and before the drives have reached the end of their life we will have cloned the data to new bigger and better ones.

But spare a thought for those who are working with the highest quality cutting edge video formats. For them, a maxed out new MacPro is not just an option, it's essential. Add the cost of backup RAID and they are looking at an investment of $15,000 plus, and that's without the camera.

Working at increasingly higher qualities, presents some added challenges ..

- You need edit-ready hard drives and interconnections capable of high read/write times
- You need raw CPU power capable of real-time or fast as possible offline rendering
- You need a backup solution for all the raw data, which may have cost a fortune to film, and which probably cannot be re-filmed if lost, and the truth is, there is no adequate long-term solution for the kind of file sizes 4 and 5 k produce

Returning to the world of audio, we have almost none of these problems. And why? Because, for most of us 44.1 or 48kHz at 24-bit is all we will ever need, and even the most cost-effective of todays computer systems is capable of running scores of tracks and plug-ins without complaint.

If you understand digital theory (our videos on the subject are planned for later this year) you will know that frequencies towards the top end of the audio spectrum (20kHz) are poorly represented at 44.1 and 48kHz sample rates. At 44.1kHz a single cycle at 20kHz is represent by just over 2 samples, barley enough to represent the change between compression and rarefaction. It looks kinda worrying!

But can you honestly say you could differentiate 44.1kHz, 96kHz, 192kHz, and a reasonable quality MP3 (say 224kbps) in a blind test? I know we would all like to think we can, and certainly MP3 can handle modulated sounds such as phasing and ambience poorly, but for the most part, the differences are a lot smaller than you'd think.

Are you willing to take the test? Make a recording at 192kHz, duplicate it and convert it. Then ask a collage to re-name the files A, B, C etc (keeping a reference of which is which). Now, listen to the files and try to identify them.

Even if you can hear a difference, it will be small. We couldn't consistently identify the files.

And that's why there is no general out-cry about the quality of audio. No overwhelming demand for higher sample rates and word lengths, and why the attempts to introduce commercial playback systems failed. We think there is a case for moving on from MP3. Networks (the web, mobile networks etc) and storage are more than capable of handling raw PCM (uncompressed) audio data, but do we NEED better quality uncompressed audio? We don't think so.

So sit back, enjoy being able to process a snare drum with a chain of plug-ins as long as your arm, forget about a new MacPro, and spare a (smug) thought for our video brethren. They're going through hell right now. At least they don't have to worry about that 3D nonsense anymore!

Thanks for reading.
FairFax