Like Parallel Compression, Serial Compression is one of those esoteric terms that seems to pop up in recording magazines from time to time. While the name might seem abstruse and academic, the process is anything but:
“Putting one compressor before another is something that was going on long before it got a fancy name that made it sound like a ‘technique’,” says Joel Hamilton, one of the four NYC engineers we asked to weigh in on the subject.
“But the idea that you can kind of mine different things out of the same signal by chaining devices with different tones or time constants is totally valid.”
Simply defined, Serial Processing is the use of two (or more) similar effects on the same audio track. Most often, you’ll encounter the term as it refers to compression, EQ, and de-essing.
In addition to Hamilton [Tom Waits, Elvis Costello, Dub Trio], we talked to producer/engineer John Agnello [Sonic Youth, Dinosaur Jr., Kurt Vile] as well mastering engineers Randy Merrill and Scott Hull of Masterdisk, about their approach.
WHY DO IT?
“It’s like using shellac,” Hamilton continued. “You can’t buy a bucket of shellac, pour the whole thing out on your tabletop and expect it to turn out extra-glossy. But, by applying it in a dozen tiny layers, one on top of the other, you can bring the surface to a really high shine.”
“Much the same way, you can’t compress 20db with a Neve 33609 and expect it to sound like several devices each pulling back a few db.”
All of our panelists agreed – sometimes, spreading the work across more than one unit leads to better results:
“It’s well-known that in general, the shorter the signal path the better the sound quality,” said Scott Hull of Masterdisk [Miles Davis, Bruce Springsteen, The Ramones].
“That’s true, and I’ll never use more gear than what’s needed to achieve the goal. But you can’t always get what you need from one device. What you may need is the complex interaction between two.”
“I would probably never choose to put two of the same compressors or EQs inline on the same track, but I will often use two different-sounding but similar types of processors if the combined result is better than without.”
When we boiled it down for this article, it became clear that our panelists consistently cited three basic reasons for stacking their effects: Tone, Tweak Points, and Time.
“There are some pieces of gear that just have a great character and I’ll use them when that character is needed,” mastering engineer Scott Hull said.
“What might confuse engineers that use primarily digital processing, is that an analog EQ isn’t always an EQ. And an analog compressor isn’t always a compressor. Running through my compressors with no gain reduction sometimes produces very favorable results from a tone or color standpoint.”
Producer/engineer John Agnello agreed:
“I believe that when you’re in the analog world, different pieces of gear do sound different from each other, even if you’re just passing signal through them. You can patch into a Pultec and it’ll sound completely different than an API 560 before you even do anything. Sometimes you don’t want to do a ton of EQ, but you want the sound of that piece of gear.”
“With a Pultec,” Agnello continued, “you might just add a little bit of low end or a little bit of top-end, and still get the sound you need from it. But you may still want to sculpt it more, so you might go into an API 560 after that and notch out or notch in a bunch of frequencies.”
“So there’s the sound quality of each piece of gear, and then there’s also the practical factor of having access to all the frequencies you want to get to. You may want the sound of a Pultec but the flexibility of a graphic EQ.”
“I’ll do that a lot on snares and vocals. I’ll usually go through a 560, and then at the end of my chain I’ll have a nice fat Pultec, or maybe a Daking, just to give it a little size on the back end – just to take that sound and make it sound 10 percent bigger at the very end.”
Hamilton had similar thoughts: “There’s a difference between boosting 3k on a Neve 1084 and boosting 3k on an SSL EQ. On an SSL, I know that frequency is going to hurt me a little. The same way, I might want to boost 8khz on a Pultec instead. So that way you can end up with a few EQs on the same source pretty quickly.”
Merrill had similar points to make: “I use multiple EQ’s in series a lot. The curve shapes and phase responses of each of my EQ’s is different. I’ve found that in some cases, several small, incremental adjustments across multiple EQ’s gets me the sound I’m looking for, as opposed to adding more EQ on one unit in the same range. Other times this isn’t the case, and I’ll do a lot of EQ with one unit. It always depends on the mix, but I’d say more often than not, I’m using multiple EQ’s.”
With compressors, there’s another crucial factor: attack and release times.
“When it comes to the ‘how’ part,” said Hull, “ I find it’s simply a matter of putting the compressors with longer attack times first in the chain and faster attack compressors and limiters later in the chain. This isn’t brain surgery.”
That kind of stacking was the first thought to come to mind for Agnello and Hamilton as well:
“As far as time constants, you could put on a very slow, low-ratio compressor first, and send that into a fast limiter that’s catching just the top of the peaks,” said Hamilton.
“That can make it feel like the track is plowing through molasses to get to the limiter. With that approach, you can take something that’s very lightweight and stringy, like an arpeggiated nylon-string guitar, and get some real heft out of it. It’s almost like adding a sense of inertia; some real weight in a mix.”
“It can help get rid of that really unencumbered and pointalistic sound that people associate with a straight up Pro Tools mix, where you have all these really spikey transients. For me, it could be a slower, gushier compressor first, like a Collins 26-1U, followed by a Neve 33609 set to a really fast attack and release.”
Agnello had a similar approach on electric guitars:
“A lot of times I’ll go with a tube compressor and put a solid state compressor on the back end, or vice versa, depending on what sounds I’m going for. If I’m having trouble getting a guitar to sit in a mix, I might put it through an LA-2A to give it a big tube action, but at the end put it through an 1176 and compress the sh*t out of it, to make it really like a brickwall.”
Of course, there are times when the opposite approach can work. When tracking an especially dynamic vocal or bass part, it can be advantageous to set up a compressor with a fast release time first in the chain. A busy and dynamic part can wreak havoc on a slowly responding compressor, allowing some peaks to go by uncompressed, while low-level parts of the performance wind up quieter still, buried in the trough of the compressor’s lazy recovery.