Reason 's BV512 is pretty sophisticated as vocoders go, and is capable of much more than Sparky's Magic Piano-style voice effects.
Back in the middle era of Reason 's evolution (2003), Reason 2.5 arrived with a Cambrian explosion of new effects devices. The BV512 vocoder might seem like an anachronistic device in Reason: an effect traditionally used on vocals in a music package that you can't record into or use to process external signals? There are, in fact, ways to bring recorded vocals into Reason, but in any case the BV512 can be used for a range of creative and unusual purposes with Reason 's other sources.
Let's have a look at traditional vocoding before we get into some of the more unconventional uses for the BV512. A vocoder imposes the spectral 'fingerprint' of one sound onto another, so most vocoders require two audio inputs. The exception is if the vocoder has some kind of internal synth to generate one of the sources (Prosoniq's Orange Vocoder is an example of this). As you'll see later, Reason 's Combinator means you can build a device that operates in a similar way.
The screen below shows the rear panel of the BV512 vocoder, with the various input and output connectors arranged around a schematic diagram. Ignore the Shift and Hold inputs for now, as they are CV inputs. We're mainly interested in the stereo 'carrier' inputs and mono 'modulator' input, which are where our audio sources will be routed. Both sources contribute to the sound of the output in different ways. The carrier is the 'raw material', or basic tonal texture of the sound. This is filtered based on the spectral content of the modulator signal. Imagine the vocoder as a spectrum analyser linked to a graphic equaliser (or, strictly speaking, a series of envelope following multi-band filters). The level of the modulator signal is analysed at various frequency bands and the resulting EQ curve is applied to the carrier wave. So if your modulator is a static, unfiltered white noise signal (which contains all frequencies at equal levels) the result will be a nearly flat frequency plot and the carrier wave will pass through the effect almost unchanged.
The screen above shows two snapshots of the front panel of the BV512. The top graph, labelled 'Modulation Levels' displays the levels of the modulator signal measured at each of the frequency bands. The more bands you choose with the knob to the left of the graph, the higher the resolution of this frequency curve. In the top panel, the modulation input is a white noise source and the graph is almost flat (it's not completely flat because the noise source is not perfect). In the second panel, the modulation source is a sawtooth wave, giving a characteristic series of harmonic spikes with the fundamental frequency determined by the pitch. The top example doesn't do much, but the bottom already has the interesting effect of applying an unusual filter curve to the carrier signal that you are processing. In this way, you can use the vocoder to produce an unlimited number of new filter shapes. Things get even more interesting if you actually play the saw wave in the second example. Any pitch change in the saw wave will shift the whole frequency plot (and therefore the filter effect) up and down. Anything you do to change the modulator has a result on the main carrier signal, so you can add filter and amplitude envelopes, pitch bend, LFOs and even a step sequence to the modulator and create more and more movement in the carrier. This is why vocoders are so effective when you use a voice as the modulator. The ever-changing sound of the voice makes the carrier sound shift in interesting ways that are instantly recognisable as speech or singing.
Let's start with an example of traditional voice vocoding. You will need a vocal sample loaded up and playing back from the sequencer. See the box on the next page for advice on how to set this up. In this example I have a female vocal which I'm playing back from an NNXT sampler. The carrier wave (which will determine the tonal starting point for the sound) will be provided by a Malström synth. The screen above shows the necessary connections. The vocals are connected to the modulator input (the microphone graphic is a handy reminder for which way around to connect the sources), and the Malström goes into the carrier inputs. When you press play, the vocals start playing, but you won't hear anything; you'll just see the frequency display on the BV512 front panel start dancing away. Now, if you play something with the Malström you will still hear very little, maybe just a very low and dull signal coming through. This is because the Malström 's default patch is a sine wave, which by definition has no sonic texture beyond a single frequency. What we need is a very rich carrier sound that can be shaped by the vocoder's filtering. Try switching Osc A to a simple sawtooth. The result is instantly recognisable as the classic 'talking synth' vocoder sound. Try playing some chords higher up for a shiny resonator type sound, or single notes very low in pitch for the raspy robotic effect. You can also impose an entirely new melody on the vocal by playing the Malström.
Now that we've established the basic effect, let's have a look at some of the front-panel controls on BV512. First try pushing up the HF Emph knob. This boosts the high frequencies, improving the clarity and crispness of the effect. Most sound sources roll-off in level towards the higher frequencies, biasing the vocoder's filter effect towards lower frequencies, so the HF Emphasis evens out the frequency response. You can also fine tune (or dramatically alter) the cut or boost of every band by moving the yellow bars in the 'Frequency Band Level Adjust' display. The Attack and Decay controls adjust how quickly the filter bands open and close as they follow the modulation signal. Setting a very low Decay causes the filters to follow changes in the modulator very closely, making the vocal effect very clear and voice-like. Increasing the Decay will smear the sound into more of a resonator effect crossed with a spring reverb. Grouped with the Attack and Decay controls is the Hold button. Pressing this button freezes the vocoder's filters, allowing you to 'capture' a spectrum at a moment in time from a shifting modulator sound. The Shift control adjusts all the frequency bands acting on the carrier up and down. In effect, this makes it sound as though the modulator has been formant shifted and is very dramatic on vocal sources.
The final control we should look at before moving on is the band selector. This selects how many frequency bands the modulator is split into for analysis, and how many bands of filtering the carrier passes through. The more bands, the higher the resolution of the frequency plot extracted from the modulator. In effect, the more bands you use, the more the output takes on the character of the modulator. In our example of a vocal modulating a sawtooth wave, as you increase the number of bands the output increasingly comes to resemble the original human voice. In FFT (Fast Fourier Transform) mode, there are a whopping 512 bands, so very fine detail can be heard from the modulator, and, in the case of a vocal, you can tell whose voice it is. Lower band settings give more synthetic results, so choose a setting based on the sound you are looking to achieve.
A great use of the BV512 is to create rhythmic synth parts, using a drum machine or REX player as modulation source. A favourite trick is to connect the modulator input of a BV512 to a send from the mixer. You can then feed a signal from a Redrum or a loop player (or both) using the channel send controls. If you then play a synth or sampled sound into the carrier input, it will be gated by the rhythm tracks and also mimic their shifting sonic characteristics. The Reason Operation Manual describes an old-school trick in which you use this same technique, but with a noise source as the carrier, and the result is a remarkable imitation of a reverb that you can colour in all sorts of interesting ways. Using a loop or Redrum source also has the advantage of giving you fast results without needing to sequence anything in the main sequencer.
Reason is very much a synth rack rather than a full-blown production environment, and as such has no ability to record audio tracks or process external audio signals. Generally, if you want to record vocals alongside Reason you do it in another application and run Reason via Rewire. However, if you want to use singing or speech with Reason 's vocoder, you need to get it inside Reason somehow. What you need to do is chop out the phrase you want in your audio-editing program, and export it as an audio file (WAV or AIFF). You can then load this into one of Reason 's samplers. As you can see in the screenshot below, I've created an NNXT sampler and loaded in three vocal phrases, which I've named Chorus pt1, Chorus pt2, and Chorus pt3. I've assigned each phrase to a different single key (C2, C3 and C4), and set their root notes to be the same as the keys they are mapped to. They can now be triggered from the keyboard, or from the sequencer.
This being Reason, you won't be surprised to learn that there are some extra quirky but very powerful ways to use the vocoder. The first concerns the two rows of 16 CV connections on the back panel. These provide a completely different way to control the filter bands that can be used instead of, or alongside of the modulator input. The CV inputs and outputs map directly to the filter bands if you are in four-, eight-, or 16-band mode. In 32- or 512-band mode, each port is shared by bands. Connect a CV signal to one of the inputs, and the corresponding filter's (or group's) level is now controlled by the CV signal, and no longer follows that band of the modulator signal. You can use any of Reason 's many CV sources, such as LFOs or envelope outputs. A particular recommendation (again suggested in the manual) is to use the gate outputs of a Redrum drum machine. You can then create a pattern sequence in Redrum and trigger different frequencies in the carrier signal in a rhythmic manner. The level and length of each 'note' is determined by the velocity and length settings on Redrum 's channel controls.
Finally, you can trigger the vocoder's filter bands to open via MIDI (again instead of, or as well as, using a modulator signal), allowing you to 'play' the vocoder and record sequences of triggers in the main sequencer. MIDI notes trigger the bands to open, starting from C1, and they stay open while a key is pressed. The filter levels are mapped to velocity, and their response times follow the envelope set by the Attack and Decay controls. The Hold button can also be controlled via MIDI note C4. The button only stays active while you hold the key. In order to route MIDI data to the vocoder you will need to create a MIDI track for it (effects devices don't get MIDI tracks by default). Just right-click on the vocoder and choose Create / Sequencer Track.
The Combinator has the potential to make working with the vocoder somewhat simpler (for example, you can save common routing configurations in Combinators for easy recall), and also raises interesting new possibilities. Another big advantage is the ability to play multiple devices within a Combi at once from the same MIDI keyboard. The above screen is an example. This Combinator contains a vocoder and two sound sources (a carrier and a modulator). Because they are both in a Combi, you can play both of them at once, removing the need to sequence one of the sources. You can have one source follow the pitch of the notes, and the other (typically the modulator) only trigger at key presses by switching off keyboard tracking. In this example, the carrier is a bright sawtooth Malström patch again, and the modulator is a Subtractor synth with a dramatic sweep and fast LFO that adds lots of movement to the final sound.
To create a vocoding Combi that gives instant results to any signal you pass through it, you will need to set up a source with a Matrix sequencer so that it plays automatically. There are a couple of presets in the Factory Sound Bank ('Evil Vocoder' and 'Wah Wah Vocoder') that demonstrate this technique, one with a built-in modulator and one with a built-in carrier. You can find them in Combinator Patches / Effects Devices / Vocoder.
The BV512 has an equaliser mode that lets you control the multiple filter bands manually, using the band-adjust bars in the central display area. Used in this mode, the BV512 ceases to act as a vocoder, and is simply operating on a single signal as a kind of graphic equaliser (as shown in the screen below). The signal that is processed is the one connected to the carrier input, and the modulator input is ignored. The number of bands is set as normal with the band-selector knob. Regular readers may remember that the BV512 was used in four-band EQ mode back in the January issue of SOS to provide simple band 'kill' switches in a Combinator patch. One word of warning here, though: if you use the BV512 as an equaliser in anything but FFT (512-band) mode, it will colour the signal even when the bands are all at 0dB. This is because there is some overlap between the bands, and therefore some phase cancellation and reinforcement. You can hear this in particular if you turn the Shift control and sweep the frequencies of the bands. If you need a transparent signal path through the BV512, use the FFT mode.
Vocoding is often an under-exploited tool in Reason, but is a great avenue to explore for finding new and inspiring sounds. Once you're comfortable with its operation, be sure to check out some of the tips and tricks in the Operation Manual.