Spectral
A brief introduction to spectral synthesis…
Spectral synthesis and additive synthesis represent two different ways of creating complex sounds by summing together a number of simpler components. In additive synthesis (which you can read about here), each component is a sine wave partial tuned in relation to a harmonic series. In spectral synthesis, the audible spectrum is divided into a large number of ‘spectral bins’, and different sounds are represented as different distributions of energy across these bins. Normally, Alchemy plays spectral data by summing together a number of sine waves centered in each bin; it can also fill each bin with filtered noise.
This process uses a model known as a phase vocoder. We use the term ‘spectral synthesis’ to avoid confusion: Alchemy does not provide a conventional ‘vocoder’ and is not designed to impose the spectral properties of one signal on another in real time.
Alchemy lets you resynthesize existing audio files (WAV or AIFF) using spectral techniques as well as additive ones. Your choice may depend on the type of material: additive resynthesis is often preferred for sounds that project a single clear pitch, while spectral resynthesis is well suited to chords and noisy sounds. Alchemy can also perform resynthesis using a combination of the two techniques, which is useful for sounds that combine a clear pitch with a noisy component (e.g. the hammer strike of a piano, or the breath noise of a flute).
See the Import page for instructions on how to perform resynthesis.
Spectral resynthesis allows you to manipulate your source material in unique ways. You can switch between resynthesis using sine waves and resynthesis using filtered noise — this can transform speaking into whispering. And for detailed work, you can edit spectral data by applying various painting tools in Alchemy’s dedicated Spectral Editor.
Unlike the other elements in Alchemy, the Spectral element generates a mono (rather than stereo) signal, so the Source STEREO button has no effect on the Spectral element signal. If you import a stereo file using spectral resynthesis, the left channel will be used and the right channel will be discarded.
Modes of operation
The Spectral element in each source can operate in one of three different modes, selectable via the three Spectral Mode buttons at the left of the Spectral sub-page:
- WHITENOISE. Produces steady white noise.
- RESYNTH. Plays the current spectral data (from resynthesis or painted directly into the Spectral Editor) by summing sine waves.
- NOISE-RESYNTH. Plays the current spectral data (from resynthesis or painted directly into the Spectral Editor) by summing bands of filtered noise. The result is still more or less ‘pitched’, but with a distinctive noisy/phasey quality.
Controls
- KEYTRACK button. Determines whether or not the pitch of the Spectral element tracks the keyboard in RESYNTH and NOISE-RESYNTH modes. For keytracking to occur, the main Source KTRACK field must also be set to ‘Key+PBend’ or ‘Key’.
- Volume. Adjusts the volume of the Spectral element, independent of other elements in the same source.
- High Pass. Sets the frequency below which spectral bands are eliminated (acts like a steep highpass filter).
- Low Pass. Sets the frequency above which spectral bands are eliminated (acts like a steep lowpass filter).
- Amp knob and pop-up menu. Provides control over the Spectral Formant Filter, described in further detail below.
Using the Spectral Formant Filter
The Spectral Formant Filter produces complex filtering effects by adjusting the balance among all the components that are summed together to play a spectral sound.
NOTE. Because the Spectral Formant Filter depends on changing the balance among numerous discreet components, it can be used to filter a spectrally synthesized sound, but it cannot be used to filter the raw white noise signal produced by the Spectral Element in WHITENOISE mode. To create formant-filter effects in WHITENOISE mode, the best approach normally involves using the Source Filters in a parallel configuration (click for more details on this technique).
The Spectral Formant Filter can be applied to the Spectral Element of any Source. To access it, click the Amp pop-up menu, scroll to ‘Formant ModMap’, and choose ‘Fmt 1’ (or create a new one). When you choose Fmt <N>, Alchemy will interpret the Nth ModMap as a description of how to boost/cut various frequencies: the horizontal (x) axis represents frequency, while the vertical (y) axis represents amplitude. For further details regarding the Spectral Formant Filter and the ModMaps that describe its behavior, see the corresponding description of the Additive Formant Filter, which operates similarly.