by Craig Anderton
Iris is iZotope’s first foray into musical instruments, and in typical iZotope fashion, let’s just say they’re not doing a subtractive virtual analog synth. Iris will make your head explode twice—first when you’re trying to figure it out, and second, when you come up with your first patch that makes you go “wow.”
If you’ve done spectral editing with Adobe Audition, Steinberg Wavelab, or Roland’s R-Mix, or used iZotope’s RX 2 audio restoration program, the principle behind Iris is much easier to understand—imagine spectral editing that changes over time, under your control, instead of the standard-issue spectral editing of “I want to remove just a cough” or “I want to remove the kick drum and nothing else.” And if you’ve worked with Photoshop or similar paint programs, then you’ll understand the toolset. If you’re not familiar with these concepts, Iris (the visual reference is not a coincidence) may be daunting . . . but don’t worry, we’ll take care of that.
Let’s start by dealing with the part that’s easy to understand because it’s based on the usual synth- or sampler-based frames of reference. Iris can layer up to three samples (Fig. 1), along with a “sub” static waveform (pulse, saw, sine, sine clean, pink noise, and filtered pink noise).
Fig. 1: Iris’s main screen. The three sampled layers are toward the top, with the static Sub waveform at the bottom.
Sub doesn’t mean sub-bass, although it could be; instead, you can do such wild manipulations with the samples that it’s possible to lose any melodic root. Layering the sub along with the samples brings the combined sound back to reality, or at least, back to something with a defined pitch.
Each sample, as well as the Sub, has various controls (Fig. 2). The Mix window shows all these controls at once, which is convenient, although you can also see a selected sample’s controls on the main screen.
Fig. 2: The “control panel” for an individual sample. Note the choice of LFO waveforms.
Samples can play as a one-shot or reverse one-shot, as well as loop independently with a choice of forward, backward, forward/backward, or backward/forward. Sample transposition can be resampled (in this case duration relates to pitch), have a fixed pitch, or use iZotope’s Radius algorithm that maintains duration despite pitch changes (this is the same algorithm used in Sonar for their high-quality, offline pitch transposition functions). Each sample also has fine and coarse tuning, gain, and pan controls.
For modulation, an ADSR amplitude envelope complements a multi-waveform LFO with tempo-sync and restart options. The LFO can modulate either pitch, amplitude, or pan, but no more than one destination at a time.
Effects are Distortion, Chorus, Delay, and Reverb (Fig. 3) with a choice of Send or Master global effects modes. In Send mode, each sample has its own set of effect send controls.
Fig. 3: The four effects, configured as Send effects. Fig. 2 shows the Effects in Master Effects mode.
Master mode precludes send effects, as all four effects are applied to the composite sound of the mixed samples. As you can see from the screen shot these have a pretty complete parameter set (what you can’t see is that Distortion has six algorithms), and if you know iZotope, you know they have the recipe down for effects.
The Mix window also has a master section (Fig. 4).
Fig. 4: The main window’s Master section.
This section includes a multimode filter with eight modes, ADSR envelope, global LFO with choice of amp, filter, or pan destinations, along with host tempo sync; this is also where the Master Effects depth amount controls become operative if Master Mode is selected for the effects.
But one of the coolest Master section attributes is the ability to assign controls to MIDI Learn and Macros. Macros are sort of like “mini-presets” for most parameters (mixer page and main page), with the added advantage that the parameters assigned to macros can be controlled by Macro knobs, and one knob can control multiple parameters (Fig. 5).
Fig. 5: Macros are all about adding real-time control. Macros 1 and 2 are currently assigned, and the Macro 1 knob for the Sample 3 Amp Envelope attack knob is about to have its minimum value updated. Note that Sample 3’s LFO Depth control is assigned to MIDI controller 16, as indicated by the associated number.
Not all parameters can be assigned to Macros, but most can; you can also impose maximum and minimum limits to the control range. Additionally, there’s an alternative way to manipulate Macros 1 and 2 via a virtual X-Y pad. One point of confusion: For most parameters, you click on a knob to assign it to a macro but with the envelopes, you don’t click on the envelope graphic, but on its labels.
MIDI learn is really easy: Click on a parameter name, diddle the hardware control you want to assign. Done. To unlearn, click on an X next to the parameter name. You can’t set minimum and maximum values (unless this is doable at the controller), but as with Macros, a single MIDI controller can control multiple parameters at once—like control all sample release times with a single fader. As expected, you can also assign the Macro controls to MIDI controllers. (iZotope should strike a deal with Korg to bundle the nanoKONTROL 2 with Iris. Just sayin.’)
Finally, the main window has a virtual keyboard where you can route modulation and mapping, which also affects any hardware keyboard you’ve connected. Iris responds to velocity and aftertouch, as routed to one of four destinations from a pop-up window (Fig. 6).
Fig. 6: Keyboard modulation sources can route to various destination.
This area has an option to pop up the Macros window, adjust Glide, and tweak keyboard tracking. This is also where you can map samples to keyboard ranges if you want to set up splits and layers (Fig. 7).
Fig. 7: Splits and layers are possible for the three samples and Sub waveform.
Also note the strip above the keyboard, where you can edit several characteristics for a selected sample (loop start, loop end, loop crossfade, delay, etc.).
Iris’s dirty little secret is that if you want to treat it like a conventional sampler and have a multi-sampled bass or whatever, you can do that (albeit with more limitations compared to samplers like NI’s Kontakt, MOTU’s MachFive3, or Reason’s NN-XT, which are optimized for that sort of task). Then again, you can also drive a Porsche without taking it out of second gear, but why? So now that you’re prepped about the standard elements, let’s investigate what makes Iris special.
You can’t understand Iris, or get predictable results from it (her?), unless you understand the concept of a spectrogram. You’re no doubt used to seeing waveforms presented as amplitude vs. time, like the drum loop at the top of Fig. 8; the big peaks are kick drum sounds.
Fig. 8: Three different ways to look at a waveform, from top to bottom: amplitude vs. time, spectrogram display, and a blend of the two. Note that in iZotope, the view you select affects all samples; this was done through cut-and-paste with a paint program.
Below that is the spectrogram display. This is three-dimensional in that the X-axis indicates time, but the Y-axis represents frequency; amplitude corresponds to the brightness. So in the spectrogram display, you can see the kick drums are bright (loud), low-frequency blobs. The thin spikes in the top view are hi-hat hits, and in the spectral view, you can see that they extend way into the high frequencies and because they aren’t as loud, aren’t as bright as the kick.
Iris allows doing a “visual crossfade” between these two views, as shown at the bottom of Fig. 8. This makes it easier to correlate the two ways of looking at waveforms.
Now, here’s the payoff. The spectrogram display is a highly visual way of looking at a waveform, so Iris provides tools that resemble that of a paint program to select, move, erase, and otherwise manipulate different sections of the waveform. For example, suppose you want to get rid of just the kick drum from a file; the spectrogram display makes it easy, because you can remove sounds based on frequency as well as time and amplitude.
One way to do this is to select the entire file, then use the lasso tool to select only the kick drums. There’s a shortcut so that if you hold Alt/Crtl while the lasso is chosen, it will erase whatever you select. In Fig. 9, the upper sample is the original sample, while the lower sample has the kick drums selected (the areas within the dotted yellow lines).
Fig. 9: The kick drums in a drum loop have been selected and erased.
So, you won’t hear the kick drums—here's proof, with two audio examples:
But wait! There’s more!! Of course, Iris isn’t just about editing drum loops—this was just a way to illustrate of how spectral editing lets you “get inside” the waveform, and remove particular frequency and/or time slices. Let’s bring out the heavy artillery.
Here’s an example of how to make a pad transcendent. Check out the original pad, which was created with Cakewalk’s Dimension Pro soft synth.
Now take a look at Fig. 10. I set this to spectrogram display only (no standard waveforms) to make it clearer as to how the pad was edited.
Fig. 10: “Irisizing” a pad.
I used the same pad for all three samples. On the top, notice the arc; the lighter section shows the frequencies that are selected. As the sample progresses, you hear more high harmonics until the halfway point, at which point the high harmonic start to diminish.
Meanwhile, in the middle sample, I’ve traced a path (the wavy line) through the pad’s harmonics. This adds motion. I’ve done the same basic edit in the lower sample, but the path is offset so the two samples go in different directions.
To anchor the melodic element, the Sub waveform is a sine. Note that it doesn’t extend the full duration of the other samples—but that doesn’t matter, as Iris can loop each sample independently. Now listen to the Irisized sound.
The original pad could loop seamlessly, and I wanted the edited loop to do the same. So, I was careful to make sure that my selections were in the same regions of the waveform at the sample’s beginning and end. Iris does include a crossfade function to simplify looping (it’s used in the second sample), but the better the original loop, the better the crossfade.
The editing for this particular sound was done solely with the lasso tool. There’s also a brush and eraser, both with variable size, and even a “Magic Wand” tool that selects regions with similar spectral proﬁles.
This audio example of the modified pad should not only give you an idea of the power of what you can do with Iris, but also, that once you understand how spectral editing works, you can create predictable results—although half, if not more, of the fun of Iris is just loading up crazy sounds, getting out the various paint and selection tools, and going nuts.
If this review has piqued your interest, now would be a good time to hop over to the iZotope site for more details. There are lots of audio examples, so I don’t see a lot of need to post more here; you’ll be amazed at how Iris can transform everyday sounds into sound design nuggets by letting you isolate specific portions of the sound. Of course, you can play these all from your keyboard and create melody lines and chords that have no precedent; Iris simplifies this process by a sort of intelligent root note detection process, although you can edit the root note if needed.
Finally, as iZotope probably recognizes that not everyone is going to start diving in and making their own sounds, you get a 4GB library of samples and over 500 presets as part of the deal. Furthermore, iZotope has already introduced three optional-at-extra-cost libraries of samples and patches—Wood ($29), Glass ($49), and Food ($49).
But really, what with hand-held portable recorders being so inexpensive, go and grab some of your own samples and Irisize them. You might be shocked at how much potential music and sound design lies within everything from ceiling fans, to running water, to the call of a coyote.
Craig Anderton is Editor Emeritus of Harmony Central. He has played on, mixed, or produced over 20 major label releases (as well as mastered over a hundred tracks for various musicians), and written over a thousand articles for magazines like Guitar Player, Keyboard, Sound on Sound (UK), and Sound + Recording (Germany). He has also lectured on technology and the arts in 38 states, 10 countries, and three languages.