Food for thought

My wife Terri organizes company field trips. We do two a year and they're always welcome, fun events. This year the entire company invaded the Super Audio Center run by Gus Skinas as well as toured Immersive Studios (the only recording studio in the world with a full 32 track Sonoma DSD recording system). Gus treated us to many sonic delights. It was a real 'ear-opener' to many in the company who had never heard recorded music sound so live. One piece of information that stood out for me, relative to yesterday's discussion about Soft Edges, was a comment Gus made about PCM vs. DSD and analog. It isn't often we hear about the differences between the two formats from the recording engineer's viewpoint and I found it illuminating. His comment concerned multitrack recording and mixing as most studios do today. Multitrack recording incorporates a separate 'track' or, in today's lexicon, 'file' for each microphone used. So, imagine we have a small group of performers: two singers, a drum kit, stand up acoustic bass and a keyboard, each separately mic'd. Each of the five microphone feeds are sent through individual A/D converters and stored on a computer's hard drive. When it's time to playback and mix the five channels to a stereo version, the tracks are digitally mastered according to the recording engineer's mix and then converted to analog through a stereo D/A converter into analog. It is in the digital mixing process where the differences between PCM and DSD really become apparent. When the recording engineer mixes the five mono channels into stereo he does so digitally and places each channel into either left, right, center, or a combination of L and R to replicate the placement of the performers in acoustic space. When the session is an analog or DSD capture, the process is straightforward and everything works according to Hoyle: each instrument occupies the correct space and the listener can easily discern their position in space: left, right, front back. But when the session is a PCM capture, the acoustic space gets muddled and more difficult to separate between instruments, requiring the recording engineer to change EQ settings and manipulate the sound of the track to get it right. What's fascinating about this observation is that analog, with its limited dynamics and frequency response (relative to DSD or high sample rate PCM) does not experience this issue, nor does DSD (which is closer to analog). Interesting stuff for thought.