"The Loudness War" or Whatever

As someone who likes to write about, think about and - above all - listen to sound, I think the so-called "loudness war" is a fascinating subject. For those of you not in the know, the idea is that musical recordings have gotten louder over the past 20-30 years as producers have started using more dynamic range compression. The result, some claim, is that recordings are being "ruined," and sound "less musical" because they have less dynamic range. Or so the critics allege. This awesome Sound on Sound article points out that, to quote Albert Nimzicki in Independence Day, "that's not entirely accurate."

SOS's research found that true, proper dynamic range isn't actually changing at all, but that crest factors seem to be decreasing; that is, the overall loudness of recordings is closer to their peak loudness. To put it simply, recordings are louder overall, as opposed to being loud in certain sections. So the critics are still somewhat right. But what does that really mean for sound quality? Do these metrics mean music is being ruined? I propose that we approach sound quality... well, qualitatively, dammit.

While there are certainly egregious abuses of loudness, the ultimate goal of any engineer or producer is to make something that sounds good. So I think loudness, timbre and dynamics should all be judged contextually. Some records make sense as loud, distorted assaults on the ears, like Jay Reatard's Blood Visions or anything by Death Grips. Justice uses absurd limiting on their basslines, but they do it in an artistically interesting way. For me, limiting goes wrong when the music is more nuanced, like My Beautiful Dark Twisted Fantasy or Sleigh Bells' Treats. Treats is an interesting case, because you could make the argument that - like the others I've mentioned - it's supposed to be loud and aggressive. But Treats also has a certain rhythmic, stomp-and-clap sensibility, and I think the excessive limiting detracts from that. Similarly, I think too much limiting ruins the dynamics of the intricate arrangements and variations in Kanye's most recent effort. While it's nice not to have to adjust your volume, there's a balance that needs to be struck and sense to be made.

Love/Hate: Rustie

I don't know that Rustie really falls under the love/hate category for me so much as a guilty pleasure. But even then, why the guilt? What is there not to love about songs like this? As an appreciator of subtlety and subdued sounds, sometimes I'm slow to open up to really poppy, energetic music. There's just something overwrought about it to me. But Rustie's proven in the past that he can make more subtle music, like his remix of "Spliff Dub" by Zomby. So why the distrust?

What's saddest about my hesitance to absolutely love this album is that it's incredibly unique. The non-stop euphoria is a little over the top for my tastes, but the sounds on it are unlike anything out there. Glass Swords is a mixture of cheezy '80s synths, weird voice-like basslines and MIDI slap bass samples. Who uses a slap bass and makes it sound cool?! It spans a variety of genres (house, dubstep, "trap", other unidentifiable things) with a sound evocative of Scuba's Personality, or Ford and Lopatin's Channel Pressure, but somehow more accessible than either (and less goofy than the Ford and Lopatin album). Maybe these basslines just sound too much like Skrillex.

Production Tips Vol. 2: Compression

Novice music producers and engineers get exposed to the idea of dynamic range compression early on, but because of its subtlety and tempo-based nature it takes a long time to master; I've been producing for over five years and my approach to using it is still evolving. When used correctly, compression can give your tracks punch and help smooth out inconsistent elements, but when misused it can ruin your tracks. In the constant pursuit of loudness, the latter is usually what happens. My aim here is to provide some conceptual pointers on using compression, so that you can understand how it works and use it to suit your needs.

What exactly *is* compression?
A compressor, to get all technical on your ass, is a non-linear amplifier. The easiest way to conceive of a compressor is as an automatic volume control. When a signal exceeds a certain level (called the threshold), its gain is reduced, effectively turning down the volume at superhuman speeds. You might also think of this as a kind of "soft overdrive" or waveshaping. The amount the signal is reduced by is called the ratio of the compressor. The attack and release time settings determine how aggressively this reduction kicks in and out, respectively. Many compressors have adjustable or soft knees as well, which let you control how gradually the volume reduction happens. Soft knees tend to sound more transparent, but hard knees give more aggressive volume control.

Cool, so what do I do with that?
Well, the answer is, "it depends, what's wrong with your signal?" Compressors can either be used for leveling (making slow volume adjustments) or limiting (lowering the volume of loud events in a signal), and they can be applied anywhere from individual tracks to masters. Knowing what you need is the first step. If a track is too "spiky" or "jumpy," you probably want limiting. If something is more subtly uneven and just needs some extra thickness, you want a leveler.

Setting up a limiter
I find the easiest way to set up a limiting compressor is to start by maxing everything out: shortest attack, longest release, highest ratio, threshold at 0 dB. This makes the effect of the compressor painfully obvious. From there, adjust the threshold so that only loud spikes are cut off; parts of the signal below that should sound normal. Next, shorten your release time as much as possible without inducing any uneven pumping (volume changes that don't match the rhythm of the music). After that, increase your attack time so that an appropriate amount of punch and detail comes through before volume reduction. Finally, adjust your ratio to get the desired amount of gain reduction; typical limiting ratios are 4:1 and above. More than 3 dB of gain reduction tends to sound unnatural, but sometimes unnatural might be what you want.

Setting up a leveler
Leveling compressors are less dramatic to set up, but harder to tweak, as they have a much more subtle effect. I start by setting the threshold just below the lowest signal level, with a 50 ms attack and 1 s release. From there, set your ratio between 1:1 and 2:1 for the desired effect and adjust your attack and release as needed.

Am I doing this right?
Compressors can be overwhelming at times, so here are some typical settings to see if you're on the right track:

• Vocals:  3:1 - 4:1 ratio, .5 ms attack, 80 ms release, -24 dB threshold
• Drums:  5:1 ratio, 40 ms attack, 200 ms release, -18 dB threshold
• Mix Buss:  1.5:1 - 2:1 ratio, 50 ms attack, 180 ms release, -6 to -12 dB threshold
• Leveler:  1.1 ratio, 100 ms attack, 1500 ms release, -35 dB threshold.

20-Minute Mix: February 2013

I recently made a mix of some awesome future garage/bass tunes from fellow soundclouders. Check it above or on soundcloud. Artists include Londy, Husky, Morning High, Subjam, Kid Simpl and myself.

On Repeat: FaltyDL ft. Ed MacFarlane - "She Sleeps"

My first exposure to FaltyDL was 2009's wonky, funky Bravery EP, and my favorite track of his until today was probably "My Friends Will Always Say." Despite his sonic experimentation, I've always thought of him as a garage (as in U.K.) producer, because his garage tunes are usually my favorites. No more. While the jury's still out on his reserved, nuanced full-length Hardcourage as a whole, "She Sleeps" is breathtaking.

Certain tracks have an infectiously rhythmic quality without being forceful, and "She Sleeps" nails that mood. I was struck immediately by how delicate it felt; a deep kick drum and occasional subbass drives the beat, but the other elements - shuffling snares and hi-hats, vocal pads and Ed MacFarlane's abstract falsetto vocals - are mixed back and seem to just float along, with no snare on two or four. If you listen closely there's actually a lot going on, but it feels incredibly sparse due to the lack of reverb on everything but the synths and vocals.