Tips To Tame the DX7II
Programming your own DX sounds is not so difficult when there's someone to guide you. Kevin Stratton, product specialist with Yamaha USA, explains how to create vocal-like sounds and reverb with this FM synthesizer.
As a product specialist with Yamaha, I'm often asked whether programming the DX7II is more difficult than programming the original DX7. The truth of the matter is that most of the new features don't add any more surprises to the elusive 'Edit Zone', but do give more of the features we asked for: better MIDI control, alternate tunings, stereo, more memory, better sound quality, and so on.
Contrary to myth, programming the DX7II is not at all impossible for mere mortals. In this article, we'll cover a few of my favourite programming tips, and open the doors to a more realistic approach to achieving FM sounds never before heard.
Before we can understand FM we must first understand how our ears perceive sound. In looking at how humans hear, we'll see that the DX creates sounds according to these very laws.
In the human hearing process, our ears are sensitive to only two acoustic qualities: pitch and level (volume). In the DX, most of the parameters will control one of these two qualities. For example, pitch is affected by parameters such as Tune, Ratio, and Fixed; level is affected by the Output, Scaling, and Amplitude parameters. Changing one of these parameters will affect Pitch or Volume accordingly.
No matter what word describes the particular kind of pitch or level change taking place, any sound editing parameter change is going to represent either a change in pitch or a change in level. This has always been helpful for me in understanding what is going on inside the DX.
The DX uses 'sonic building blocks' called Operators. Each Operator (Op. for short) consists of a sine-wave generator (with an associated envelope generator to control the sine-wave generator's output); the Operator can serve as either a carrier or a modulator. A carrier provides us with an audio output. A modulator injects rapid frequency fluctuations into the carrier frequency, which affects the output and imparts a particular tone colour, or timbre. Changing either the amplitude (volume) or frequency (pitch) of the modulator affects the timbre of the carrier output.
Multiple Operators (FM instruments typically use four or six Operators; the DX7II uses six) can be combined into what's called an algorithm. Think of the algorithm as a number of vertical stacks of Operators. Carriers (ie. the Operators at the bottom of the algorithm - see Figure 1) have their outputs mixed together to provide the audio output and constitute the fundamentals of our sound. Modulators are stacked vertically on top of the carriers or on top of other modulators. In Figure 1, Operators 1 and 3 are carriers. Operator 1 is modulated by Op.2, and Op.3 is modulated by Operators 4 and 5. In addition. Operator 5 is being modulated by Operator 6. (That line around Op. 2 indicates a feedback path, and is yet another way to create timbral changes.)
Now let's consider what happens when we change the frequency of the carrier or modulator. Changing the carrier's frequency will change the perceived pitch of the sound. Changing the modulators frequency doesn't change the perceived pitch, but does change the harmonics that will be emphasised. Regarding level, increasing the carrier's level increases the overall level of that Operator, or stack of Operators. Increasing the modulator's level increases the amount of Frequency Modulation that takes place, and therefore increases the amount of harmonics (also known as 'sidebands') produced by the carrier. These are really the only major rules of thumb you need to know when forming a sound on the DX. There's a good reason why the algorithms are printed on the DX front panel; they give you an idea of how the sound is constructed, not unlike the block diagram for an electronic circuit.
Let's go back for just a moment to the years when giant subtractive modular synthesizers roamed the earth, in the days of Voltage Controlled Oscillators and Filters. In those days, tweaking the knob of a filter or changing the waveform of an oscillator gave us timbral changes. Today, we can still obtain these kinds of sounds with ease, and this is a good place to begin your experiments in FM programming.
In creating the five basic waveforms (sine, triangle, sawtooth, square, and noise), the only components needed are a single carrier and a modulator stacked right above the carrier. By placing the DX into initialise voice mode, we are ready to go; here's the 'template' from which we can create the five basic waveforms.
Operator 1 (carrier) output level: 99
Operator 2 (modulator) output level: XX
Operator 1 frequency ratio: 1.00
Operator 2 frequency ratio: Y.YY
Now all we have to do is substitute different values for Z, XX, and Y.YY according to the following table:
|Waveform||XX value||Y.YY value||Z value|
|Sine||40 or lower||Any value||0|
|Triangle||40 to 70||1.00||0|
|FM Sawtooth||70 or higher||1.00||Any value|
|Square||40 or higher||2.00, 3.00, 5.00, etc.||0|
|Noise||93 or higher||Any value||7|
Note that the third waveform in the table is labelled 'FM Sawtooth', this is because another sawtooth exists inside the DX7II - not one generated by FM but, actually, a pure digital sawtooth wave. Like the TX81Z, the DX7II has a different waveform to add to the FM process, and this wave is found in the Feedback loop. Try this for yourself:
1. Initialise a voice in your DX.
2. Set the Algorithm to 32.
3. Turn off all Operators except for Op.6.
4. Set the output level of Op.6 to 99 and Feedback to 7.
The sound you obtain is unlike any standard FM sound. Thinking of the feedback as a separate waveform can be very helpful when using Algorithms 4, 6, and 32. Try them out, and while you're at it, change the Key Mode (button 23) from polyphonic to 'Unison Poly' and set a unison detune of your choice. Thoughtful use of this sawtooth waveform can embellish and thicken sounds.
In more complex FM configurations, where there is more than one modulator, all the above rules still apply. The versatility of a complex stack can even help place your sounds in an environment, so get ready to extend your sound into space as well as time.
First, let's look at two effects that play a big role in the ambience component of a sound: diffusion and delay. When a sound reverberates in a room, waves bounce off every hard surface there and return to the listener at different times. The final effect causes the reflected waves to cancel themselves out at different times. This happens so quickly that we perceive the delayed part as a shadow of the original or direct sound.
To hear this in action, set up your DX like this:
|Algorithm = 9, Feedback = 7|
|*Note: Op. 5 is in Fixed mode and can have a Frequency from 1.000 to 2.884.|
We have yet to talk about envelope generators, but bear with me and just enter the numbers as shown.
If you look back at Algorithm 9 (Figure 1) you will notice that there are two basic stacks; Operators 1 and 2, and Operators 3, 4, 5, and 6. The first stack (1 and 2) provides a basic FM sawtooth wave, with the modulator's envelope acting like the envelope of a filter as it varies the effect of the modulator on the carrier. This creates a simple Brass voice. The second stack (3, 4, 5, and 6) contains an identical frequency ratio between Op.3 and Op.6, with a fixed frequency sandwiched in between. The fixed frequency causes the harmonics generated by the 1:1 ratio (Op.3 and 6) to wash out, thus leaving a mirrored, diffused version of the first stack that works well to create our 'ambience'. Envelope generators in the second stack create the exact types of delay properties desired. For example, you can change Rate 4 of Op.3 to lengthen or shorten the reverb time, and change the Fine Frequency of Op. 5 to alter the pre-delay. With a little time and effort, you'll soon be adding reverb to any waveform you like.
A wavetable is nothing more than a specific waveform(s) that precisely duplicates a sampled or acoustic event. So far we have seen how to create basic waveforms, which can be considered wavetables in their own right. What we're about to cover forms the foundation of an acoustic sound - the human voice. This particular sound, as most programmers agree, is one of the most difficult to master and always sounds different to the next listener. The wavetable of this sound is not a patch, but a structured foundation with which to begin your own patch.
The vocal wavetable requires a minimum of two Operators, and a maximum of three Operators if you desire a broader bandwidth. First let's start by initialising a voice, selecting Algorithm 3 (see Figure 2), and switching off Operators 4, 5, and 6. Next, raise the output level of Op. 1 to 99 and the output of Op.2 to 87. Set the frequency ratio to 2:1 (in other words, the carrier [Op. 1 ] to 2.00 and the modulator [Op.2] to 1.00). This creates the resonance of the throat. By adding a Pitch Modulation sensitivity of 2 and Pitch Modulation depth of 9, and selecting a triangle wave LFO in Multi-LFO mode with a speed of 29, we add movement to this static waveform. Movement becomes a very important factor because the human voice has the unique quality of changing harmonic content without altering the pitch (this is known as 'Chroma Shift'). The best way to create this Chroma Shift is by setting the Key Mode to 'Unison Poly' with a detune value of 6. Now add in Operator 3 at an output level of 44 and a frequency ratio of 6:10 - this gives the voice its nasal resonances. From here, experimenting with the envelopes of the modulators can give the voice different pronunciations, while changing Op.3's Fine Frequency setting produces different voice qualities and bandwidths.
Fractional Scaling, a new feature of the DX family, allows control of the volume (or output level) of each Operator across every three keys on the keyboard. Notice that the range of the output level is extended from the original 0 to 99 to 0 through 255 on the DX7II - this certainly lets you control those small differences "between the cracks". Think back about what changes in modulator output do to the sound, and imagine what fractional scaling can do to give you truly tight control over the harmonic content in a sound (ie. the cracks become canyons, or mole hills turn into mountains, depending on what kind of fanciful constructions you want to create). Try experimenting with different volume levels for the modulators across the keyboard; see if you can take the vocal wavetable and develop areas on the keyboard that form independent bass, tenor, alto, and soprano sections.
Well, it's hard to believe, but I'm just as long-winded on my word processor as I can be in a demo. Still, there are many more features on the DX7II that call out for recognition: microtuning, dual and split modes, panning, and so on. But for now we are out of time and space, which means it's time for me to get my daily dose of "Green Screen" (Yamaha Specialist jargon for programming time!).
One last bit of advice on programming the DX7II: be a leader, not a follower. In these days, where the sound is your signature, sign your own name and dare to dwell on the edge.
© 1987 Electronic Musician magazine ((Contact Details)). Reproduced with the kind permission of the Publishers.
Gear in this article:
Feature by Kevin Stratton
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!