Street Patcher

A patcher is a musician who manipulates live sound using a DSP modular environment like Max/MSP. Street Patcher was a software designed for a duo - a musician (patcher) and an interviewer. The duo would go in a public space and interview people. The vocal signal would be recorded in real time. The program allowed to cut snippets of the recorded vocals (people speaking) and to sequence them in a rhythm while applying sound effects to make a live composition. The key element was surprise. People would think they simply answered a question, but then they would hear their voice turned into a deep-house beat. Never performed with it but I had a lot of fun building it.


For demonstration purpose, there is a 'fake input' button in the top-left Bpatcher which simulates the sounds that would come from the microphone in the real performance. The big toggle located in the top-left corner starts the entire patch. It is recommended to press the 'init' button before starting.


On the top-middle Bpatcher the input signal can be viewed in two types of waveforms differentiated by two colors: the white waveform indicates the amplitude envelope of the recorded signal whereas the yellow waveform shows the level noisiness which was analysed during the recording. High values indicate harmonic spectrum, low values show inharmonic sounds. Press the 'Y' key (or the bottom-left pad on Novation LunchPad MK2) to stop the auto- strolling waveform, therefore being able to select a short segment from the SELECT section. Click and drag over the waveform the region that you would like to process. You can reselect by clicking and dragging again or by holding shift and adjusting the range limits. When the desired segment is selected press 'Y' again. The waveform will retake the auto-scroll mode, therefore making visible the current sounds that com from the microphone and allowing another selection.


Within the SEQUENCER area, the selection has been copied allowing the performer to make another three selections by clicking and dragging over the horizontal range sliders. These mini-selections can be programmed to be played in any desired rhythm by clicking on the 3X16 grid matrix located below the waveforms. Additionally, three freezing points (fft-frames) can be chosen by holding the control key (on Mac) and moving the other 3 horizontal sliders that overlap the waveforms. These three sonic points can be programmed in any desired rhythm by clicking on the 3X32 grid matrix located at the bottom of the SEQUENCER section. On the right hand side there are envelopes for setting the pitch modulation of these mini-segments, the option for applying a envelope that modulates the cutoff frequency of a lowpass filter, the range of this movement, and two general 2-way axis slider that control two lowpass filters. The horizontal movement represents the resonance while the vertical movement represents the cutoff frequency. Also here the 'Chordy' button is located, which triggers, whenever pressed, a random chord one randomly chosen mini-selection in the desired length. In the right-bottom corner of the SEQUENCER area there is an envelope that controls the amplitude of a 'frozen' sound, which is extracted from the first slider on the left-hand side of this area. This sound will be triggered every 16 bars and will last as selected in the tab above it's amplitude envelope.


The main selection is automatically copied in the waveform located in the SNIPPETS area. Over this waveform two sliders named ORIG and MOD allow the performer to select two starting points for two 'snippets' that will be played according to the settings made in the SNIPPET section. By default these snippets will be repeated in loop, but this behaviour can be adjusted by controlling the tabs on the top side of the SNIPPETS section where the length of the snippet can be chosen, as well as the length of the pauses between the snippets. The difference between the ORIG and MOD snippets is that the MOD can go through a speed-modulation generated by the analysed noise level of that particular snippet area. The amount of modulation can be set from the 'depth' knob in the right- bottom corner. At maximum amount of modulation the reading speed will be decreased when the analysis indicates a harmonious spectrum and increased when the analysis indicates a inharmonious spectrum (ex: the consonants 's' and 't').


On the 3x12 grid matrix located in the BEATS area, different pre-made percussive loops can be chosen, as well as chordal loops and bass loops. All beats have the length of 8 bars and a maximum of 8 loops can be played simultaneously. Notice that all changes within this section will take place only at the beginning of the next 8-bar phrase which is indicated by the slider at the top of the BEATS section. Please notice that if you change the BPM speed from the RECORD section, the beats will not stretch to the new settings, therefore for another tempo it is needed another set of beats.


In this last Bpatcher of the signal process flow vital settings can be done that will affect the sonic output. On the 1x8 grid the performer can choose which 'engines' they what to work with. This grid functions as an on/off switch for the desired engines and all changes made on the grid will take effect only at the beginning of the next 8-bar sequence. Once selected, the signal for the desired engines will be left through, and the volume for each channel can be adjusted with the yellow dials. The orange dials provide control over an additional lowpass filter for each channel as well as the option to modulate with an LFO their cutoff frequencies, with LFO adjustable speed and direction. A solo option is available below the on/off switch grid. Finally several effects can be added such as reverb and delay, four additional effects offered by Absynth 5 FX from Native Instruments (disabled for this version) and two self-made effects. The one named 'Bellify' is a customised multi-band- pass-filter bank based on analysis of different spectrums of bells. The Q Factor for this effect can be adjusted, as well as the glissando time from one bell spectral image to the other. These bell spectral characteristics are changing randomly every bar. Finally a beat- chooper is placed at the end of the signal flow, allowing the performer to repeat in a loop the current mini segments of the mixed sonic output. The flow ends with a master gain slider.


All controllers can be saved and recalled in the main PRESETS section located in the top- right corner. Exception makes the range slider from the SELECT area, all sliders from the SEQUENCER area and the two sliders from the SNIPPETS area. Press shift-click on a slot for storing a preset and simply click on a slot for recalling.


Many thanks to prof. Johannes Kretz for his techincal support and creative input in creating this patch.