Sound Research WIKINDX
Sound Research WIKINDX |
|
 |
|
Davies, G., Cunningham, S., & Grout, V. 2007, September 27–28, Visual stimulus for aural pleasure. Paper presented at Audio Mostly 2007, Röntgenbau, Ilmenau. Added by: Mark Grimshaw-Aagaard (07/05/2008, 10:27) |
 |
| Resource type: Proceedings Article BibTeX citation key: Davis2007 Email resource to friend View all bibliographic details  |
Categories: General Keywords: Sound mappings Creators: Cunningham, Davies, Grout Publisher: Fraunhofer Institute for Digital Media Technology (Röntgenbau, Ilmenau) Collection: Audio Mostly 2007 |
Views: 13/697
|
| Abstract |
|
Can images be represented by music and if so how? This is one of the principal questions addressed in this paper through the development, implementation and analysis of computer generated music and assessment of the music produced. Images contain a multitude of definable information content such as the colour values and dimensions, which can be analysed in various statistical combinations to produce histograms and other interpretations of the content. This provides music compositionalgorithms with a variety of data sets and values which can be used to inform parameters used in the music generation functions.However, images also contain semantic information which is widely open to interpretation by a person looking at the image. Images, much like music, are able to induce emotions,feelings and other soft responses in a viewer, demonstrating that images too are more than simply the sum of their parts. Therefore, if we are to generate music from image data then a goal of doing so would be to produce music which can begin to invoke similar humanistic responses. The work presented in this paper demonstrates how compositional algorithms can be used to create computer generated music from aset of various images. Several established music composition algorithms are explored as well as new ones, currently being developed by the authors of this paper. We explore different uses and interpretation of the image data as well as different genres of music, which can be produced from the same image, and examine the effect this has. We provide the results of listener tests of these different music generation techniques that provide insight into which algorithms are most successful and which images produce music that is considered to be more popular with listeners. Finally, we discuss our future work, directions and the application areas where we feel that our research would bring particularbenefit. In particular, we seek to incorporate the work presented in this paper with other technologies commonly used to assess visual and psychological responses to images and propose techniques by which this could be harnessed to provide a much more dynamic and accurate musical interpretation of an image. One of the main goals we aim to achieve through further development of this work is to be able to successfully interpret an image into a piece of music or sound, which could be played to a visually impaired or blind listener to allow them to grasp the emotional content and responses which imagery can invoke. Added by: Mark Grimshaw-Aagaard |
| Notes |
|
Potentially interesting because it talks about mapping visual parameters to musical parameters. However, it uses MIDI so infinite variety in an image is shoe-horned into 127 discrete steps. Why not use real-time synthesis such as Pure Data or MAX/MSP?
Added by: Mark Grimshaw-Aagaard Last edited by: Mark Grimshaw-Aagaard |