Digital Can Come Back to Bite
Tdawn shared the following AES paper titled “Stop Counting Samples” on twitter, and it confirmed a whole multitude of things I’ve run into over the past few years. In a nutshell, digital audio is a vastly different animal than analog, especially post mastering, and if such is not considered, it will come back to bite, often in a huge way. Whats perhaps most distressing is that multitudes of young engineers grew up on digital, and yet it seems, and is shown in the analysis of recordings in table 2, they are blowing it.
Historical Hot Recordings
Back in my day, if you cranked up the levels too high, waveforms would get their peaks clipped and as a result, multiple harmonics of said waveform would result. In addition, the B-H curve / saturation characteristics of mag tape can make for some really interesting types of distortion, (and it approached with moderation, might even be considered pleasing to some folks ears).
Digital Music is often Subject to Post Production Processing
Digital post production processing is very common. On the one hand, it could be tricks used within a mp3 player to save silicon die space, on the other it could be tweaks used to optimize streaming audio, or even OTA broadcasting. Post production processing assumes that the input signal is good to go… When the input signal has digital artifacts, even if normally not audible, when you post process such, they can come back in spades. Fig 7 shows how this may occur in a huge way.
Note the 5.5Khz tone… and the big spikes in spectral content mirrored above AND below 5.5Khz.
When it such shows up, outside of a sine wave in the lab, is that something may have sounded perfect in the mastering room, and even on a few different pieces of equipment… but a slightly different CODEC / re-sampling methodology in the customers home, and boom, all bets may be off.
Whose problem is it, artist or listener?
Granted, there likely are some folks who take the stance of, hey its good to go here, and if a customer plays it back on something else and the quality takes a dive, its their problem. I can sort of understand such… but then if a customer buys music, and has a bad experience, are they likely to blame the musician, or blame the signal processing going on in their device. More than likely, the musician is going to be the one taking the heat…
What about live recording?
Now, consider a recording of a live performance, where the musicians are using DSP amplifiers on stage for psycho acoustical effect. Micing such amps and then bringing them into the desk in addition to a direct feed for effect would seem to be all analog… but is it really? What about leakage into the vocal or crowd mics? Can such bite? Perhaps… the key is being careful throughout the entire signal path, and testing. Assumptions more often than not in the analog world worked out fine, and in many cases were a positive. The digital world is different, and if left unchecked, it can bite with a vengeance.