What exactly constitutes “punch”? Find out here
by Craig Anderton
We all know a punchy recorded sound when we hear it—but what exactly constitutes “punch”? It seems that perhaps punch is something that can not only be defined, but quantified.
This all started because years ago, I wondered why seemingly every musician agrees that the Minimoog has a punchy sound. Then, when I started playing a Peavey DPM3, several people commented that my bass patches had a punchy sound, “like a Minimoog.” Clearly, the technologies are totally different: one was analog, the other digital; one used voltage-controlled oscillators, the other sample playback. Yet to listeners, they both shared some common factor that was perceived as punchiness.
Analyzing a Minimoog bass line revealed something interesting: even with the sustain set to minimum, there was about 20-30 milliseconds where the sound stayed at maximum level before the decay began. There is no way to eliminate that short period of full volume sustain; it’s part of the Minimoog’s characteristic sound.
I then looked at the DPM3’s amplitude envelope and it exhibited the same characteristic—a 20-30 ms, maximum level period of sustain before the decay kicked in. Also, both instruments had virtually instantaneous attacks. Could this combination be the secret of punch?
For comparison, I then checked two synths that nobody considered punchy-sounding: an Oberheim OB-8, which is generally characterized as “warm” and/or “fat” but not punchy, and a Yamaha TG55. Both had fixed attack times, even with the attack control set to zero, that lasted a few milliseconds. I also recalled some experiments ex-Peter Gabriel keyboard player Larry Fast ran in the mid-70s, when he was curious how fast an attack had to be for a sound to be “punchy.” His research indicated that most listeners noticed a perceptible loss of punch with attack times as short as one or two milliseconds.
So it seems the secret of punch is that you need an extremely fast attack time, but you also need a bit of sustain time at maximum level. This sustain isn’t long enough to be perceived as sustain per se; it’s more of a psychoacoustic phenomenon.
Wondering if this same technique worked with other sounds, I took an unprocessed snare drum sound and tried to add punch by normalizing each cycle to the highest level possible for the first 20-30 milliseconds. Comparing the processed and unprocessed sounds left no doubt that the edited version had more punch.
When I designed the Minimoog Expansion Pack for Rapture, I made sure that where appropriate, the envelopes had that characteristic Moog attack (Fig. 1). Note that the second node sustains the sound for 27.5 ms. Rapture’s tight attack time and ability to create “high-resolution” envelopes made it easy to add punch.
But what if your synth’s envelope doesn’t have a fast attack and brief period of sustain? There’s not much you can do about the attack, but a little limiting can cut off the top of a waveform and give the required sustain. The key point is to set the threshold high enough so that you clip or limit only a few milliseconds of the sound. If the sustain becomes too long, you don’t hear punch because what happens isn’t quick enough.
In closing, here’s a word of caution: Just as you don’t want to fill up a piece of music to the point where it’s cluttered, you don’t want to “overpunch” every sound you use. For example, suppose a kick drum and bass note hit at the same time. Consider adding punch to the kick to reinforce the rhythm, but leave the bass alone; its attack will probably be masked by the kick drum transient anyway. A continuous succession of punchy sounds can fatigue the ear; use punch sparingly.
Rapture is also included in SONAR X3 Producer