Studio Procedures 101
I promised I would parse the comments that Paul McGowan posted the other day on his daily blog. And I will in a few moments. But I wanted you to know that the person that told Paul about the problems with PCM compared to analog or DSD was Gus Skinas of the Boulder Super Audio Center and the dealer for Sony’s Sonoma system. Gus is wonderful person and I know that he fervently believes in the advantages of DSD. He responded directly to the controversy that Paul’s comment initiated. Here’s what he claims to be true:
“Well, Paul did indeed get these ideas from me. (Gus). And I stand by them. During the PS Audio visit to our facility in Boulder last week, I relayed an observation that we experience daily during recording sessions at my neighbor’s recording studio. Immersive Studios has a 32 channel Sonoma DSD recorder and a Euphonics CS2000 analog console. This is the configuration they use most often. They also have another popular PCM workstation, which they usually run at 96k 24 bits. Most of the time, they are using one or the other. Sometimes both. And while we have not constructed a scientific test to document these observations, it is just something very obvious that we notice when switching between the two recording platforms. In either case, the mixer is the analog console. What we find is that when tracking to DSD, we have to work far less hard to make instruments sit where we want them in the mix than we do when tracking to PCM. It’s that simple. With the PCM multitrack, EQ might be required to better distinguish one instrument from another panned close to each other. With the DSD, often no EQ is needed to achieve the same (or better) result.”
While I know Gus is an experienced engineer and respect his opinions on many audio production realities, this strikes me as another “it just sounds better” argument…kind the guys that swear by expensive power cords. As I wrote in my response to his comments, I’ve done this exact comparison at a first rate studio in Whitefish Lake, Montana and none of the engineers that compared DSD 64 and PCM 96/24 felt that PCM took more work.
As for the usual procedures that Gus imparted to Paul, the use of DSD requires mixing to be done in the analog domain. The Euphonix CS2000 console in the Immersive Studios room gets its input and output from the 32-track Sonoma multitrack recorder. The CS2000 console has EQ and auxiliaries for routing to reverberation devices and back to be mixed into the stereo output busses. The use of EQ to carve out unique spaces for individual instruments is applied equally to analog, DSD, and PCM tracks. The use or avoidance of any processing is done at the discretion of the engineer, artist, and producer. The choices are made for sonic reasons and not because of the individual format in use.
Most mixing in commercial releases doesn’t happen in the digital domain. The “magic” of older analog consoles still reigns supreme for the biggest names in mixing. But mixing “in the box” is growing, plug ins are getting better and closer to the original analog outboard processors, and headphones are used more and more to finalize projects in home studios. You just need to know what you want and go after that sound.
I fail to understand the value of mixing high resolution (per Mark’s definition) in “the analog domain” OR I fail to understand what is possible for high resolution in “the analog” domain.
I do not accept assumption that “analog electronics” cannot process frequencies in the 20-40+ kHz range but the noise, loss, and other fragility seems almost insurmountable. This is where I don’t understand the value of all the conversion to mix in the “analog domain”. It seems like one needs the audio equivalent of a very expensive “clean room” to preserve the high definition attributes during mixing.
Converting more than twice seems unwise. Convert to digital as close to the source as possible once and only once. Convert to analog as close to the destination once and only once.
Has analog technology radically change and is no longer susceptible to the affects of physical environment variations, interference, etc that introduce noise and signal degradation? If yes (which I doubt) then let’s eliminate digital audio (and video). What?? We now have quantum interconnects? ;-}
Wayne, engineers and producers choose to mix in the analog domain because of the tools that are uniquely available, the “sound” of analog, and the physicality of turning knobs and flipping switches to achieve a particular sound. Analog circuitry can easily pass ultrasonic frequencies AND handle the dynamic range required to be called high-resolution. And, as I’ve pointed out some engineers and producers are forced to work in the analog domain because there are any tools to work in their preferred format…i.e. DSD.
Individual choices are made by producers and engineers per their own requirements.
While I could confirm, that when using text book linear phase alias and anti alias filters for PCM, that I got a different soundstage (mostly depth) with PCM when comparing with the analog signal coming in or with DSD. But when I use minimum phase filters for alias and anti alias filters and running with 88k2 24 Bit, then I got the same (or very similar) soundstage with PCM as with DSD as with the analog input signal (but better focus and detail with PCM, but maybe a more smooth and forgiven treble with DSD). But I prefer PCM (when done with minimum phase filters) and preferring digital mixing (instead of analog mixing). So when comparing PCM, we have to always putting the time behavior of the digital filters into account, otherwise there is no “typical” PCM sound. With changing the behavior of the digital filter I can change the tonal characteristic of the DA and AD converter totally, from flat without depth to “analog”, as it was at the input side.
Juergen
Thanks for the input…I’m with you.