Ondolon is a real-time tool to transform natural sounds or musical instruments sounds into invented sounds. Developed since 2011 by Charles-Edouard Platel, its concepts are at the source of recent works.

For example, in "Soft Metal", tibetan singing boils are harmonized and orchestrated while their sub-harmonic resonances stay in synchronous vibrations. In the beginning of "Forest vibes " , we can hear modulations of barytone saxophone which are multiplied in ranges and speeds, so making hear, through the magnification of wooden reed vibrations, like an accelerated perception of the biological shivers of thousand trees. Last, in Migration, most rhythms and sound colors result from an instrumental playing executed on Ondolon.

ondolonThe Ondolon prototype

Download of the Ondolon application and documentation (version 3 Max/MSP/JItter-v7 stand alone, version 0.28 on Pure data):

version 3.04 MacOS 32-64bits

version 3.04 Windows 64bits

version 0.28 Linux, MacOS, Windows


Basic ideas

The prime idea of Ondolon is to handle an audio excerpt like a video shot:

  • to select the part which is interesting to work,
  • to crop it in height (pitch) and in width (speed),
  • to zoom it to emphasize tiny sound details,
  • to change its sound texture,
  • to freeze it on a spot, or scan it at variable speed, including backward.

The second idea is to enable a polyphonic playing of the different voices in phase, in order to respect the articulation and discontinuities of the sound.

The purpose is definitely an instrumental approach : all the user's settings are in a single window and react in real time during the listening in order to provide the musician with a feedback. The technical principles are not new, but the goal is to operate them easily towards an aesthetical intention.

How does it work?

The Ondolon is globally similar to a polyphonic sampler controlled by MIDI, plus specific digital signal processing features:

  1. The digital audio playing can be performed either from a wave table derived from the audio file, or from a phase vocoder. Actually, in a classical sampler (like in a magnetic tape player), to change the pitch, the playing speed is proportionally changed. For example, the playing speed must be multiplied by two to obtain an octave interval from the original sound, thus, the sound duration is divided by two. Thanks to the phase vocoder, speed and pitch may be separately defined, so all the polyphonic notes stay synchronized against the original sound morphological and spectral characteristics.
  2. The sequence mode : each new note is not played from the audio loop start, but after the previous one within the looping process. So the respect of the original sound, by following the concrete natural melody, prevail over the instrumental melodic sequence. Thus in polyphony, the voices play in chord this natural melody. The legato mode is similar to the sequence mode, but the ADSR enveloppe is applied to the global sequence rather than to every individual note.
  3. The canon mode : new notes are automatically triggerred from a defined point in the loop.
  4. Transients processing : just after having loaded a new audio file, a digital pre-processing is launched to anticipate on the further real-time audio processings. Particularly the audio areas containing transients, significant of the recorded sound articulation, are detected and displayed against the mono or stereo signal waveform. This allows interesting real-time processings:
    • the stationary sound areas being possibly strongly speeded down or up, the transient areas may be played with a specific speed close to natural;
    • transient sounds may be accentuated or smoothed, thanks to the ADS envelope;
    • the transient peaks may be accentuated using their shadows, a sort of very next echo doubling the signal.
  5. The AGC option may be used to boost the lower audio level audio areas.
  6. The sound spectral texture may be modified, in separately fading the harmonic components (responsible of the sound pitch) and the anharmonic components (responsible of the sound matter), and in reducing any residual background noise. The output signal spectral envelope may be displayed and modified through a graphic equalizer.
  7. Inversely, the blurring operator can shade the spectral contrasts through a light random jamming.
  8. For each played note, the hold ctl option is designed to freeze the volume and panoramic MIDI controls and the pitchbend at their values present at the note-on time, until the associated note-off, whenever these controls change on this channel.
  9. A tuner enables to check the fundamental frequency absolute pitch in a representative point of the original sound, so allowing to compensate its bia against the desired MIDI pitch.
  10. A low frequency binaural beating may be obtained by a slight shift of the stereo channels frequencies


The user operates a sound recording like a video shot. By listening, he can select a favourable area to shape, according to different looping modes, speeds, pitches and textures. Then he can play this area from a MIDI keyboard or the space bar, while modifying the previous settings through the mouse. The audio output may be directed towards other softwares or simply recorded into an audio file.

According to the original sound character and the applied settings, the digital signal processing stretches or compresses the phases in order to obtain the intended result. This may produce some aesthetically amazing sound phenomena, but also some undesirable acoustical warpings, particularly in stereo.  Thanks to the real-time settings, Ondolon helps to grope for one or several interesting configurations.

Works are re-usable:

  • in-water-flow audio output recording feature;
  • original audio files and user setting configurations can be archived in presets, for further use.


Charles E.Platel,
(August 2013, revised July 2016)