We’re about to head out on our final test drive of detailed, step-by-step sound design in Dexed. This section of the Dexed Owner’s Manual was written to offer deep insights and tricks into the steps of voice design in Dexed. We’ve come to the end of the road of this chapter, at least in our blog. In this episode, we’ll look under the hood at “Charmed” and woody, Wurly-type voice with a slight, metallic ping for each note. This “tine” is noted specially in the lower registers where, many times, a synth sound will get lost in the mix.
The Basics
Charmed offers several intriguing aspects, so let’s begin with the algorithm. Algorithm 18 is a single audio path output algorithm, where everything above affects OP1. We’ll set the feedback level at a reasonably high 6, as OP3 plays a significant role. More on this later.
Our primary tone in Charmed resembles a square wave created by OP2 modulating OP1, as we’ve seen in previous installments. OP2 is set two octaves above OP1, with a frequency ratio of f=0.5. A slow decay and release are programmed into OP1, providing the main amplitude envelope for the overall voice.
Getting Interesting …
As discussed in the last “Under the Hood,” OP2’s EG features a slightly faster decay time, allowing us to appreciate its unique contribution. The release rate is likely programmed longer than that of most electric piano emulation sounds. However, since many EP sonic options are already available, I decided to enhance playability and sound by applying a bell-like envelope to a tone reminiscent of an electric piano. You are, of course, welcome to adjust it as you see fit.
OP3 plays an interesting role in this voice. With a relatively low programmed Level and a tuning of f=5, it adds the “reedy” portion of the primary waveform as it modulates the carrier. OP1. If you want to add a bit more edge to this reed, you can change the Coarse tune to a higher number and perhaps increase the programmed level of the Operator. That sums up the primary Wurli-type tone of “Charmed” – OP2 creates a time-variant square wave, and OP3 provides a compressed reediness. Next, we’ll explore the “tine” portion of the sound.
Tine be tine
Operators 4, 5, and 6 are primarily assigned to create the “tine” aspect of the sound. While Wurlitzer emulations typically do not include a tine, I was eager to avoid producing another “Wurly,” so I introduced a hint of metallic “plink” at the beginning of the voice. Interestingly, OP6 takes a break in this patch, and its volume is set to zero. I could have easily turned it off – honestly, I’m not sure why I didn’t.
That leaves Operators 4 and 5 to do the heavy lifting. Operator 4 is set to a fixed frequency of 1 Hz, essentially DC, while the EG guiding the volume of Operator 6, set at a ratio of f=30, shapes the modulation. This is a good opportunity to experiment with various settings in this approach. Please take a few minutes to try different values for the fixed frequency of OP4 and its level. Also, try adjusting the level of Operator 5 and its coarse frequency. To my ears, things get out of hand and somewhat noisy quickly. It took a lot of iterative knob-turning to reach this final setting, as I recall.
Mods
The final step is to program that classic Wurlitzer-like tremolo. Decide on a speed you like( I chose LFO Speed = 29). Remember, the only audio pathway out in this Voice is through OP1, so it is the only Operator whose A Mod Sensitivity needs adjustment. For me, an A Mod Sensitivity value of 2 is like baby bear – it’s just right. One is too little, three is too much.
I also enabled OSC Key Sync for this voice. If the LFO is free-running, you never really know where it is in phase when the note-on event occurs. While this may not seem like a big deal, what can happen – and I encourage you to try the sound with this parameter turned off- is that the “tine” portion of the sound disappears. The LFO can be at a point in its cycle where it subtracts from the volume, effectively removing the programmed attack. With OSC Key Sync turned on, the LFO always starts at a modulation level of zero. This neither adds nor subtracts from the volume of the attack, so you hear the “tine” every time.
That sums up the modulation of this sound: only amplitude modulation affects volume. No pitch modulation from the LFO or pitch EG is necessary. This is pure FM, so there’s no need to adjust the filter cutoff or resonance, but they remain for your experimentation.
Work Under the Hood is done
In the past weeks, we’ve explored FM step-by-step sound design in Dexed from the very beginning. We’ve examined every control on the UI, investigated the algorithms, and conducted programming experiments. My perspective on FM programming may differ from most. I focus how the Operators interact with various modulation settings and how the algorithms provide different pathways to create distinct components of the voice.
I’ll admit, there’s another programming school that may or may not generate better sounds by examining frequency ratios and the mathematical relationships between the operators to create sound. My approach takes a different angle on FM. I hope that this approach enhances your understanding of FM programming in any case.
Looking ahead, we will examine the practical aspects of using Dexed. Beyond programming, there is a whole world of creating custom “Carts” of sounds, MIDI integration and control within DAWs, adding effects, and more. Please consider signing up for our “Become a Dexpert” mailing list on the Dexed Owner’s Manual page to ensure you don’t miss any updates. In the meantime, we will continue to explore Dexed and how to use this fantastic tool in your music-making. Thank you, as always, for reading!
Resources:
- The Dexed plug-in:
- Our custom programmed Flywheel: Dexed bank:
- Our custom programmed Crankshaft: Dexed bank:
- SoundEngine’s Complete Dexed Owner’s Manual Series:
- FM Synthesis:
Leave a Reply
You must be logged in to post a comment.