It's funny, I've been discussing them in the order in which they're listed in the menu, and I only now just realized that this is in fact alphabetical order! Maybe I'm not the best person to be methodically documenting things, if I can completely overlook something that obvious...
Anyway, the subject of this article is the penultimate synth engine, Pulse. (I'm going to leave the synth-sampler for later.)
Pulse is described in the manual (source of the above diagram) as a "Dual Pulsetrain Oscillator" synth. As you might imagine from this description, each period of the waveform is made up of two spikes -- dual pulses -- whose amplitude and relative position can be manipulated in order to produce a range of timbres which fall somewhere in the vicinity of sawtooth/square. To my knowledge this is a fairly novel model.
There seems to be some sort of hardcoded decay applied to the sound: over time the oscillator softens and reduces in volume. The effect is similar to a weak non-resonant lowpass filter whose cutoff is lowered, rounding out the pulses. It's impossible to tell whether this is actually the result of an explicit filtering step, or whether the change is achieved implicitly by altering the shape of the pulses from spikey to more triangular; either way, it gives all Pulse-based sounds a sort of "plucked" character.
Visually, there's quite a lot going on; the bulk of the display is taken up by two periods of the "pulsetrain", above which a pair of dual graphics are located: blue and orange parameter-value indicators, each with a matching coloured waveform which animates in realtime. The blue waveform alters to reflect changes to any of the four parameters.
Blue ("Filter"):
This parameter seems to control the cutoff of a lowpass filter; possibly this is the same filter whose cutoff is ramped downward at the start of each note to produce Pulse's decaying timbre.
It's quite difficult to discern exactly what the nature of the relationship between Blue and the hypothetical lowpass filter is -- perhaps instead of controlling the filter's cutoff, Blue instead controls its slope. The net result is that turning Blue CCW removes high frequencies until only the fundamental is left -- and sometimes even that can be filtered out.
There is definitely some curious behaviour happening in this filter section; in addition to controlling the cutoff, Blue may also control the resonance of the filter, and possibly rate/depth of the decay. Moreover, it's possible that the filter cutoff doesn't track the keyboard (or tracks it at less than a 1:1 rate); the net result is that closing the filter tends to severely attenuate high notes, while severely distorting lower notes.
This distortion in the low end is what suggests some sort of increasing resonance as Blue is rotated CCW; it's quite intense, and sounds like internal clipping. Possibly an incredibly short and fast cutoff sweep might also produce a similar distorting effect? I'm afraid I'm not very knowledgeable about the behaviour of digital filters; regardless, Blue seems to control several things at once.
Visually, the current value of the parameter is indicated by a pair of blue circles in the top-left of the screen; as Blue moves from min/CW to max/CCW, these circles move outward from completely overlapping to separated by a circle-sized gap. The blue waveform becomes more rounded and sine-shaped as Blue is moved towards fully CW.
Green ("Amplitude"):
This parameter seems to control the level of the pulses: higher values lead to a much louder sustained volume once the hardwired decay has ended. Whether this is direct (because Green is affecting the depth or rate of the decay ramp) -- or indirect (because the decay effect is fixed-depth so that louder input results in more signal passed through) is unclear.
Past the midpoint, this seems to add some high-frequency energy to the pulses. Possibly this is due to internal distortion as the pulses are clipped, or perhaps this is again the result of Green affecting the decay (i.e increasing the initial cutoff or decreasing the depth of the downward sweep); regardless, it changes the timbre subtly, making it brighter and sharper.
Visually, this parameter is illustrated by the height of the leftmost "first" pulse in the lower region, and the height of the first rectangle in the upper blue waveform. Note that contrary to what these graphics suggest, the level of both pulses are affected by Green.
White ("Second Pulse"):
The White parameter alters the timing or position of the second pulse relative to the first (which always starts each period of the oscillator). The effect sounds similar to adjusting the pulsewidth of a rectangular wave, but not exactly the same.
At the minimum setting (fully CCW), both pulses are on top of each other; this sounds relatively "sawtoothy". Moving White towards its maximum value (turning CW), the second pulse is delayed by an increasing amount, moving it later in the period. Medium values produce more square-wave-sounding results, and sweeping this parameter in this middle region produces PWM-type sounds.
At the maximum setting, the second pulse is delayed until halfway through the oscillator's period; this results in the perceived pitch of the oscillator jumping up an octave, as the dual pulse oscillator now effectively becomes a single-pulse oscillator at twice the frequency.
You'll notice that playing e.g a C3 with White at maximum doesn't sound identical to playing a C4 with White at minimum; while the pitch is identical in both cases (C4), the timbre of the former note will be duller and more muted than that of the latter -- this is presumably because of the action of the decaying lowpass filter, which tracks the keyboard somewhat so that each of the two identically-pitched notes has a different cutoff value.
Visually, this parameter moves the second white pulse left and right; note that as with Green, the graphic representation of this parameter isn't strictly accurate: if a single period of the oscillator spans from 0 to 1, then the graphic depicts the second pulse's minimum and maximum position as 0.2 and 0.8 respectively, while the actual range of the second pulse's position is 0 to 0.5.
The pulsewidth of the rectangles shown in the blue waveform display alter in response to movement of White; as with the lower graphic, this is more a metaphorical suggestion rather than an accurate representation.
Orange ("Modulation"):
The fourth and final of Phase's parameters controls, as its name might suggest, modulation of the oscillator's parameters by an LFO. I'm unable to determine the shape of this LFO, but I would guess that it is triangular or sine-shaped.
This is a bipolar control: turning CCW ("negative") from center produces one effect while turning CW ("positive") produces another. In both cases, as the value of Orange moves further from the center midpoint, both the rate and the depth of the modulation increase.
Negative values of Orange result in pulsewidth-modulation-type effects: the position of the second pulse is being shifted around. This is akin to manually twisting the White encoder back and forth; sonically this produces the usual lush breathing detuned type of sound typical of PWM.
Positive values of Orange result in a more dischordant, severe type of modulation. This is a complete guess, but I'm going to say that this is most likely modulation of the amplitude of the second pulse; at very low values it introduces a fluttering tremolo effect, while higher values can produce inharmonic overtones similar to audio-rate amplitude modulation -- distorting, obscuring, or altering the perceived pitch of the oscillator.
I'm guessing that this affects only the second pulse, because it doesn't sound obviously like tremolo or AM, it sounds similar but different, which is perhaps because only half of the waveform is being modulated, rendering the effect unfamiliar. Additionally, if this is the case it creates a nice symmetry, where the modulation introduced by Orange affects different aspects of the second pulse -- either its timing or its level. I think it's likely that Teenage Engineering might appreciate and introduce such a symmetry.
Visually, the animated orange waveform in the top-right of the display illustrates the rate and depth of modulation, while the line below indicates the current value of the parameter.
Note that while the blue waveform graphic does change in response to Orange, the result is misleading: negative values of Orange result in the amplitude of the blue rectangles fluctuating, while positive values of Orange result in fluctuating blue pulsewidths. This may be a bug or error, as it seems to be the opposite of the true effect of Orange (i.e negative for PWM, positive for AM).
(This is yet another reason why I guessed that positive values of Orange produced AM -- negative values clearly produced PWM sounds, but visually produces AM of the blue waveshape; since positive values produced visual PWM, I inferred that their actual effect must be AM of some type.)
Conclusion
While it's not the boldest nor the most versatile of the synthesis methods provided by the OP-1, Pulse can at the very least be said to be unique.
Sonically it's interesting, but seems quite confined and limited in comparison with the other synth engines -- there is surprisingly little interesting or useful parameter space to explore. The next post will attempt to dissect and explore these limitations.
As usual, if you have any feedback -- comments, corrections, opinions, observations, etc. -- please visit this discussion thread on the forums: http://ohpeewon.com/discussion/238/op-101-pulse
