Sample-based synthesizers use sampled audio or instruments as the basis for their sound. While they can be used to accurately replicate the sound of traditional instruments, the real fun of sample-based synthesis happens when you mangle existing audio to create unimagined sounds with creative complexity.

Sample-based synthesis breaks the rules of music production and theory. With the right settings, two (or more) samples from different sources and varying tempos, pitches, and tones can sound like they were made for each other.

In this article, I’ll show you how make the most out of sound layering and sample-based synthesis using Iris 2.

Enhancing real-world sounds

Let’s start with the basics of sample layering. Many producers achieve a convincing sound simply by sampling real word audio, like voices, birds, and instruments and enhancing it with electronic processes.

Listen to Jan Jelinek’s Loop-Finding-Jazz-Records to get a better idea. You will hear familiar elements like rhodes piano, vinyl crackle, and vocals, but there are many sounds on the album that could only exist due to technological intervention, for example, the reversed chords and delay-drenched rhythms on “Them, Their.”

The balance between these two sound categories (familiar real-world sounds and new, tech-expanded ones) is a big part of what makes the album, and so many others like it interesting.

Let’s try out something similar with Iris 2. The synth has four sample pools, each of which can be loaded with a different sample from either the built-in Iris 2 library or your own collection. I’m starting with a vibraphone in Pool 1, which is set to loop.

In Pool 2, I copied over the same vibraphone (also set to loop) and reversed it, producing a swelling effect. In Pool 3 and 4, I loaded up rhythmic samples, each panned in opposite directions. They are slightly out of time, and they fumble instead of glide over the vibraphone. These slight irregularities entertain curious listeners.

At the end of the first four bars and again, at the end of the sixth, you’ll notice a pitch freakout. This is done by assigning Pool 1 LFO to modulate pitch and cranking LFO depth way up.

From here, each sample layer can be fine tuned with common synth controls like ADSR, LFO, and effects such as reverb, distortion, and chorus (some of which can be heard in action above).

Spectral shaping from noise   

Iris 2 has a unique set of spectral filtering tools that can be used to locate the recognizable pitches of non-musical audio sources to produce musical results. Why would anyone want to do this, you ask?

Lots of producers rely on the same presets and samples to make their songs. And while it’s OK to use these same tools, the most reliable way to make music that sounds different (which is what listeners love) is to adopt different production approaches.

Here, I loaded up a car engine recording with a recognizable pitch. Looking at the spectrogram, the important tonal information is easily identifiable by the brightest horizontal lines.

Before I program any musical sequences, the general noise in the high end preventing the mid-range harmonics from being heard clearly needs to be removed.

With the magic wand tool, I can isolate and play back only the fundamental frequency (the bottom band) by clicking it once. With a bit of reverb, delay, and chorus, it sounds like this:

Of course, not all noise is a bad thing, so from here, I will actually bring back some of the high-end material that we removed earlier for brightness and air.

There are multiple ways to do this. I’ll start with the time frequency selection tool and draw a rectangular shape over the group of frequencies I want to hear (approx. 3 kHz to 12 kHz). From there, using the eraser tool, I’ll poke holes in that rectangle, so when the playhead reaches them, the noise will drop out, creating more dynamic movement. 

Of course, not all noise is a bad thing, so from here, I will actually bring back some of the high-end material that we removed earlier for brightness and air.

There are multiple ways to do this. I’ll start with the time frequency selection tool and draw a rectangular shape over the group of frequencies I want to hear (approx. 3 kHz to 12 kHz). From there, using the eraser tool, I’ll poke holes in that rectangle, so when the playhead reaches them, the noise will drop out, creating more dynamic movement. 

In this final audio clip, you will hear the new frequencies. You will also hear a much wider mix: I copied the Pool 1 sample to Pool 2, and applied slightly different spectral filtering settings to the second. To enhance this widening effect, the first sample is tuned flat and the second sample is tuned sharp. Panning plays a big role here too.

Remember—this all started with a car engine.

So, as you scavenge through your hard drive looking for sonic loot, don’t be so quick to discard noisy samples. There may be a synth hiding in there!

Check out this video tutorial on making synths from machinery—it’s from the first version of Iris, but the concept still applies.

Make it move

It can be frustrating to find new ways to keep your synth in motion. Modulation takes a while to dig into—not that you shouldn’t. And automation is useful, but a tad boring in process.

As an alternative, add interest to synths with evolving textures and frequency spectrum play. In Iris 2, load a sample with focused harmonic content. Flutes, strings, and square and sawtooth synth samples are perfect source materials here.

In Pool 1, I loaded up a bass synth sample and pitched it up to sound more like a pad. To get things moving, I drew a rectangle covering nearly the entire frequency spectrum, then carved out a path blocking the frequencies in it (in grey) from passing through.

In Pool 2, I am using the same pad sound, but this time only selecting the fundamental frequency and upper harmonics. For a dreamier character, there’s a bit of delay and chorus going.

For ambient sound layers, I have white noise running in Pool 3 and a percussive sequence in Pool 4, set on loop to play forward and backward depending on the MIDI notes played. This adds a great deal audio movement. Again, panning is used to open up the stereo field.

Once your spectral selections are set, experiment by swapping out samples for a different result. Your selections will stay the same.

Go big or go home

When it comes to sound layering and sample manipulation, go beyond simply stacking audio tracks in your DAW. Use the editing capabilities and spectral filtering in Iris 2 to craft layered synths that are full of expression and movement.

Turn conventional sounds like Rhodes and bass into unheard instruments. Or build a synth from the ground up with machinery recordings. There’s always a new avenue to explore.

For more helpful resources, check out our Getting Started with Iris 2 resource page where you can find the 20 minute video guide for Iris 2 along with other help guides, tips, and tutorials. 

Download a free Iris 2 demo here.