Differences

This shows you the differences between the selected revision and the current version of the page.

--- software:om-pursuit 2012/06/14 23:59
+++ software:om-pursuit 2013/06/13 05:21 current
@@ Line 1: / Line 1: @@
 ====== OM-Pursuit ======
-Dictionary-Based Sound Analysis/Synthesis in Computer-Aided Composition.
+Dictionary-Based Sound Models for Computer-Aided Composition.
 {{template>:software_summary
-|participants=[[people: Marlon Schumacher]]\\ [[people:Marcelo M. Wanderley]] (supervisor)
+|participants=[[people: Marlon Schumacher]]\\ Graham Boyes \\ [[people:Marcelo M. Wanderley]] (supervisor)
-|funding= CIRMMT student award 2011/2012
+|funding= CIRMMT student award 2011/2012 \\ FQRNT DED scholarship
 |license= GPL
 |period= 2011 - 2012
@@ Line 34: / Line 34: @@
-===== Overview =====
-[[https://github.com/marleynoe/OM-Pursuit|OM-Pursuit]] is a library for dictionary-based sound modelling in computer-aided composition.
-Parametric sound representations provide interesting conceptual models for music composition (see e.g. the French spectralist school). A number of software tools are available allowing composers to integrate signal and symbolic data within a compositional framework, such as extracting frequency structures from a sound file via partial tracking. Most of these tools, however, are based on sinusoidal sound models and hence rely on certain assumptions which are less adequate for non-stationary sounds containing noise and transients, for example. They typically require large numbers of individual components to describe these types of signals and poorly represent their temporal fine-structure. Dictionary-based methods, instead, represent sound as a linear combination of atoms stored in a dictionary. Mostly used for sparse audio coding purposes (compression, transmission), we propose creative applications for this model within the context of computer-aided composition, orchestration and transcription.
-Using sound samples as atoms for the dictionary we create the 'timbral' basis of our analysis/synthesis system. Dictionaries can be built from arbitrary collections of sounds, such as instrumental samples, synthesized sounds, field recordings, etc. Dictionaries can be freely combined and pruned on via sound descriptors.
-Due to the specifics of working with sound samples (as opposed to parametric basis functions), we developed an adapted matching pursuit algorithm (implemented in [[http://www.github.com/gboyes/pydbm|pydbm]] by G.Boyes) for sparse decomposition, which allows approximating a given target sound via combination of the samples in the dictionary. This can in a sense be compared to techniques for creating photo mosaics in the visual domain.
-The resulting model is an abstract description, which can be visualized and edited in OpenMusic using different representations (spatial, tabulated, temporal). The model can then be used to control sound synthesis processes, or for orchestrations and instrumental transcriptions.
-----
-{{software:om-pursuit:dict2.png?320|}}{{software:om-pursuit:om-pursuit-ov2.png?320|}}
-----
-Atomic Decomposition offers interesting possibilities for musical applications:
-  * Iterative Approximation. Control over degree or resolution of the model
-  * Approximation is signal-based rather than descriptor-based
-  * Dictionaries can be arbitrarily designed and combined (sample libraries, field-recordings, etc.)
-  * Model is naturally polyphonic (combination of superposed atoms)
-  * Constrained Matching Pursuit (control of horizontal/vertical distribution of atoms)
-  * Process leaves a "residual" - Multiple iterative decompositions (serial, parallel) of the same target are possible
-Interesting applications are possible in combination with [[software:OMPrisma|OMPrisma]] (dictionary-based spatialization), and [[software:om-sox|OM-SoX]] (batch-processing on the dictionaries, e.g. transpositions, filterings, etc.).
-OM-Pursuit was awarded the [[http://www.cirmmt.mcgill.ca/activities/support/student/past/awards11-12|CIRMMT director's interdisciplinary excellence price 2011]].
-----
@@ Line 84: / Line 65: @@
+===== Overview =====
+[[https://github.com/marleynoe/OM-Pursuit|OM-Pursuit]] is a library for dictionary-based sound modelling in OpenMusic.
+Parametric sound representations are popular conceptual models in computer-aided composition contexts (e.g. spectral models). A number of software tools are available allowing composers to integrate signal and symbolic data within a compositional framework, such as extracting frequency structures from a sound file via sinusoidal partial tracking. Most of these tools, however, are based on models which typically describe slowly modulating signals -which can be less adequate to faithfully represent non-stationary sounds such as noise and transients, for example. They typically require large numbers of individual components to describe these types of signals and poorly represent their temporal fine-structure. Dictionary-based methods offer a different model, representing sound as a linear combination of atoms stored in a dictionary. Mostly used for sparse representation of signals for audio coding purposes (compression, transmission, etc), we believe this atomic model offers interesting possibilities for computer-aided composition applications, such as algorithmic transcription, orchestration, and sound synthesis.
+Using sound samples as atoms for the dictionary we create the 'timbral' basis of an analysis/synthesis system. The dictionaries used to analyze a given audio signal can be built from arbitrary collections of sounds, such as instrumental samples, synthesized sounds, recordings, etc. Using and adapted matching pursuit algorithm (see [[http://www.github.com/gboyes/pydbm|pydbm]] by G.Boyes), the system will iteratively approximate a given sound using a combination of the samples in the dictionary -in a way similar to photo-mosaicing techniques in the visual domain.
+Communication between OpenMusic and the dsp-kernel (pydbm-executable) is handled via an SDIF interface and scripts which are generated and manipulated using the visual programming tools in the computer-aided composition environment. The decomposition of a target sound results in a sound file (the model), the residual, and a parametric description of the model. We developed a number of tools for visualization and editing of this model using different representations (spatial, tabulated, temporal). The model-data can be used within OpenMusic in a variety of contexts, from the control sound spatialization and synthesis processes, for computer-aided orchestration and -transcription, to graphical notation of musical audio.
+----
+{{software:om-pursuit:dict2.png?320|}}{{software:om-pursuit:om-pursuit-ov2.png?320|}}
+----
+Atomic Decomposition offers interesting possibilities for musical applications:
+  * Iterative Approximation. Control over degree/resolution of the model
+  * Approximation is based directly on the signal rather than signal descriptors
+  * Dictionaries can be arbitrarily designed and combined (sample libraries, field-recordings, etc.)
+  * Model is naturally polyphonic (juxtaposition and superposition of sound grains)
+  * Constrained Matching Pursuit (constraint-based control over distribution of sound grains)
+  * Process leaves a "residual" - Multiple iterative decompositions (serial, parallel) of the same target are possible, e.g. for different orchestrations
+Interesting applications are possible in combination with [[software:OMPrisma|OMPrisma]] (dictionary-based spatialization), and [[software:om-sox|OM-SoX]] (batch-processing for designing dictionaries, e.g. transpositions, filterings, etc.).
+OM-Pursuit was awarded the [[http://www.cirmmt.mcgill.ca/activities/support/student/past/awards11-12|CIRMMT director's interdisciplinary excellence price]].
+----
+In a recent project we have employed OM-Pursuit for the notation of musical audio:
+The notation of musical audio (such as a tape-part in electroacoustic music) is typified by two extremes on a continuum: On the one end are signal-based notations, such as sonograms, waveforms, etc. -which however don’t provide information about musical semantics. On the other end are symbolic notations, such as icons, characters, etc. -which in turn don’t provide information about the sonic details. Both aspects, however, are important for the interpretation of a musical work, in particular for contemporary music, which does not rely on common musical idioms. Using dictionary-based approaches (OM-Pursuit) an audio signal can be analyzed / transcribed based on a user-defined collection of smaller-scale sound files (so-called ‘atoms’). The individual atoms can be associated with arbitrary visual representations (symbolic, iconic, etc.). The user can then design simple programs to algorithmically arrange these visual elements on a 2D-canvas into a kind of cartographic map of the higher-level sonic content in the signal. Below are some example screenshots from early experiments and my piece for 2 pianists and tape for diffusion inside the piano "6 Fragments on 5’05” - 1 Act of cleaning a piano".
+----
+{{software:ompursuit-notation-1.png?640|}}
+{{software:ompursuit-notation-4.png?640|}}
+{{software:ompursuit-notation-5.png?640|}}
+----

Differences

Project Categories

Tags