Over-compression Session:
The parameters for my over-compression session were as follows:
My audio source was a simple synthesized sine wave, either indefinite in length or only as a short impulse of about 200ms length.
I stacked different amounts of copies of the Ableton Glue Compressor. They were all set to the same settings with attack and release both at shortest possible length at .1 and .01 seconds each, and ratio at x10. The threshold was set at lowest possible -40.00db while the makeup gain was maxed out at 20.00db. The dry/wet was at 100% and soft clipping was on.
I first worked with only this setup and then proceeded to experiment with putting different types of effect before this compressor chain.
At first I duplicated the glue compressor well over a hundred times and watched as Ableton informed me that to do so was requiring about 200% of my CPUs capabilities. Aurally there was little more than digital crackle, so I therefore began gradually deleting them until I started hearing desirable sounds at around a 120% usage of my CPU.I recorded some impact sounds with these settings, but upon listening back realized that Ableton is unable to record its own overloading. I would therefore need an external recorder to capture sounds created using this method.
I continued to delete compressors and at my next stop they were requiring around 65% of my computers CPU.
As a starting point I was using the lowest possible sine waves that Ableton can produce, which are at the bottom of the piano roll at 11 Hz. This will usually render inaudible sounds, perhaps perceivable as vibrations on a large subwoofer system. However, when running it through all these compressors it instead produced a rich, pulsating sound.
This highlights a parameter of misuse that is worth pointing out in the context of a thesis, namely that the practice is inherently based on not always understanding why something is happening. But from observing the unprocessed signal through Ableton’s Spectrum Analyzer I can see that the signal peaks at 11 Hz before tapering off upwards in a way that doesn't quite theoretically fit with what a sine wave is supposed to be. In the very highest frequencies the noise floor becomes visible as a sort of bubbling.
This can be contrasted with the spectrum post-compression, where every jitter in the overtone-levels become exaggerated, and part of the noise floor has also become part of the ”tone”, while part of it is too far removed in amplitude levels to even show up on the spectrum.
After recording with this setup for a little while I started experimenting with Abletons Hybrid Reverb. The Hybrid Reverb is, as I interpret it, designed to be a more ”high end” alternative to the standard Ableton Reverb. It contains a few different algorithmically generated reverbs that can be mixed as desired with convolution reverbs recorded either by Ableton staff or yourself.
What is interesting when putting the reverb in this setup is that what should be an impact followed by a fading-out turns into a continuous sound with flattened dynamic range. What is even more interesting is the fact that, on many settings even the ones showing short decay times, the reverb only needs one impulse of input to output sounds that continue for one or several minutes. Even while the output fader is showing that it is not outputting volume the compressors in the chain gradually start reacting, generating what I presume to be an incredibly quiet long background fade.
By using the ”shimmer”-algorithm in combination with the emulated ”old” vintage reverb setting I began generating what sounded like a church organ being covered mixed with exploding noise. I played with this by turning knobs and shifting impulse source pitches. The organ-like sound seemed to be struggling against the roaring noise, and when experimenting with activating it in different octaves I discovered that it would interrupt itself in interesting ways, with dramatic turns.
After this I changed both decay and size settings to minimum, which started generating shorter, morphing metallic sounds. They would shift from side to side and with changes in pitch and small settings have different nuances.
I proceeded to experiment with the Chorus-Ensemble-plugin. Using this tool, I could generate squealing and pulsating sounds with rising and falling pitches. Here the oscillating of the chorus became the main instrument. It had similar qualities to the sounds generated by SOPHIE that had peaked my original interests.
I then continued with the Ableton Vocoder. This is a tool I often use to generate faint high frequency crackles to mix with the source sound, a method of introducing ”crispness” in a mix, and it definitively belongs to the earlier mentioned category of ”unusual tips” one would find in mixing and production tutorials online.
When I flattened the crackle however, the results were strong impact sounds with moving textures within them. This was another perspective on how these crackles work, sort of blocks of aural content that are obscured via their usual difference in volume.
The final plugin that to me generated desirable results was the Redux tool. Redux can lower either the sample rate or bitrate of any given signal, and it is used for grit or a lo-fi aesthetic. Using this tool easily devolved into pure noise, however I had an interesting finding using very specific settings.
Experimenting with the built in filter and using 100% jitter on a sample rate between only 21-30 Hz, I was able to generate a sort of digital bubbling. It had a strange plucking feel, organic yet digitally crackling.
A few challenges presented themselves during this workshop. One is that the flattened dynamic range produces very similar-looking wavefiles, meaning that finding my way around the recorded material proved quite difficult. Another was a strange phenomenon where several plugins would generate some sort of sound before going silent, yet instead of the waveform returning to zero it would continually be pushing at either +1 or -1. This, as I discovered, causes the rest of the mix to go silent as the waveform is telling the speaker to remain at an extreme position, and would need to be very carefully edited out or dealt with in ways presented in the composition descriptions later in this text.
Finally, as is audible in the examples, the milliseconds where the compressors are not yet active generates clicking in the sounds. This makes them useful for things like kick layering, however in other situations the click will often need to be edited away. In later workshops we will explore ways to get around this by replacing the compressor with the utility plugin.
"A vocoder works by analyzing the sound of a modulator signal, which is usually a human voice. The modulator signal is split into many frequency bands. The level of each band is sent as a signal to a corresponding bandpass filter. The filter is set to the same frequency that was analyzed."2
Attack determines how quickly the compressor begins affecting the signal, release determines how quickly it stops doing so when signal is below threshold and ratio determines how much gain reduction the signal above the threshold is given
The Hybrid Reverb, a plugin using both artificial and convolution (an impact recorded at a real location) reverb
Chorus, an effect that multiplies the signal and plays them all back with slight time shift and a modulating pitch shift.
The modulation is controllable in the bottom panel and there are three modes:
- Classic, two signals
- Ensemble, several signals
- Vibrato, only pitch shift
Tool that flattens the dynamic range of a signal, often to keep the volume even or make an instrument sound more powerful by bringing forward low-volume material
