Mammut is a strange and unpredictable piece of software. It basically does a Fast Fourier Transform (FFT) of a sound file but unlike Paulstretch, which uses slices of the sound, Mammut uses the whole thing at once, creating more drastic results.
Mammut is not (in any way) a commercial tool but more of an experimentation one, so I won't go into detail about what is doing under the hood. Instead, I will focus on how it can be used to create interesting and cool sound design. If you want to follow along, you can download it here.
Mammut has many processing tabs but I will only cover some of the most interesting ones.
Loading & Playing sounds.
Mammut works as a standalone software. You need to load a sound (using the browse button) to be able to start fooling around. The "Duration Doubling" section adds extra space (technically, FFT Zero Padding) after the sound. This extra space give some of the effects (like stretching) more time to develop and evolve.
Play and stop the sound on the Play section. There is also a timeline of sorts. Now that our sound is loaded, let's see what we can do with it.
It creates a non-linear frequency stretching with frequency sweep effects. All frequencies are raised to the power of the selected exponent, so small changes are enough to produce very different results. Because of the frequency sweeps, it sounds quite sci-fi, like the classic star wars blaster sound. Here are some examples at different exponents:
As you can hear, as the values get further away from 1, the effect is more pronounced and it also starts sooner. Here are some results with values higher than 1:
And here are some interesting results with a servo motor sound.
This sounds remind me a bit of japanese anime or video games, maybe this could be one of the steps for achieving that kind of style from scratch.
This stretches and contracts the frequency spectrum following a sinusoidal transfer function. You can control the frequency and amplitude of this change.
This one is weird (no surprise) and it doesn't really do what I was expecting. It tends to create sounds that are increasingly dissonant and "white noise like" as you go to more extreme parameters. Here are some examples:
Quite cool. Removes all the frequencies below a certain intensity threshold. This means that you can kind of "extract" the fundamental timbre or resonance of a sound. Used on ambiences (3rd example below), it sounds dissonant at first and then, once you remove almost all frequencies, kind of dreamy and relaxing.
This one basically divides the frequency spectrum in chunks and interchanges their halves a given number of times. Hard to wrap your head around but it produces interesting results. First, the number of swaps seems to make the sound more "blurry" and abstract as you can hear:
Then, the block size seems to create different resonances around different frequencies as you increase it.
Simple but hard to predict. It reflects the whole spectrum around the specified frequency. The problem with this is that when you flip the spectrum around a low frequency, everything ends up under it and is mostly lost. On the other hand, if you use a higher frequency, too much of the energy ends up on the harsh 5-15 KHz area.
A couple of examples:
This one doesn't even have controls or an explanation on the documentation. It seems to extend the core timbre of the sound across time which can be pretty useful. When using this option, the duration doubling function is specially handy.
Mammut is certainly original and unique. Since it only works standalone and is rather unpredictable and unstable, I don't feel it would be very easy to include in someone's workflow. Having said that, is definitely a nice wild card to have whenever you need something different.