peterbaumann
New member
Hi all, I'm new here but have an unusual request! I'm currently doing an MA in Music Production, for which I'm carrying out a study into machine learning/AI's role in music production - and the deadline is fast approaching! One of the areas I'm looking into is music generation, which is becoming an increasingly hot topic and I wanted to see if any of you here would be able to spare 5 minutes or so to give your thoughts on whether or not it would be something you can see filmmakers using in the future. Obviously, a large part of the survey is specifically aimed at music producers/composers, so if you're not musical yourself, the survey shouldn't take very long at all but your voice is incredibly valuable.
It's open to anyone 18+, regardless of previous experience with machine learning technology.
What is the study investigating?
The core aim of this research is to develop an understanding of the successes and pitfalls of current machine learning uses within the feature sets of music production software from the perspective of the end user, as well as gauging the level of consumer uptake of this technology. These results will then be used to highlight where future implementation might be most useful for music producers, enhancing both efficiency and quality of output. Alongside this, the research will be presented to filmmakers and game designers to understand the viability of automated music generation systems.
Who can contribute?
The study is looking for participants aged over 18 who are involved in any aspect of music production, filmmaking and game design. This could include those working in areas such as composition, mixing, mastering or audio restoration, visual media editing, at an amateur or professional level. Although the main focus of the research is aimed at those who have access to machine learning software, input from participants who do not use software with machine learning is also important in order to gauge consumer uptake, register concerns or communicate other relevant opinions.
What is machine learning?
In simple terms, machine learning is the process of presenting an algorithm with a set of sample data, which it then uses to calculate the ideal result of a given challenge. This training process might involve taking a collection of music for media and tasking the algorithm with creating its own music inspired by the training data. For those wanting a more complex explanation, the algorithms use deep neural networks to calculate the ideal result of a particular task using weight and bias values to assign importance to aspects of the input data, in order to achieve the most consistent, accurate results.
This technology can be implemented into numerous systems, such as facial or handwriting recognition, catalogue suggestions for music or visual media providers, online translation services, medical diagnoses and self-driving cars. Within audio processing, current implementation can be found in systems intended for audio restoration, mixing, feature extraction and composition. Companies such as iZotope (RX6 ‘De-Rustle’, Neutron ‘Track Assistant’), Sony (Flow Machines), Jukedeck and Melodrive all have products which are either in development or already available to consumers – all of which incorporate machine learning.
How will the research be conducted?
All responses will be collected via this online questionnaire hosted by Qualtrics. The results will then be analysed and used as part of a dissertation project into the implementation of machine learning technology in music production workflows.
Who is carrying out the research?
The research will be conducted by Peter Baumann, a student at the University of York studying for a Masters in Music Production under the supervision of Dr. Jez Wells (jez.wells@york.ac.uk).
How will the research findings be used?
Upon completion, the dissertation incorporating research findings will be made available to all participants.
Will my contribution be anonymous?
All participation is voluntary and participants may ask to withdraw their responses at any stage before September 1, 2017. All information provided by participants will be stored securely and handled in accordance with the ethical guidelines from the University of York. If you wish for your responses to be processed anonymously, please indicate this on the consent form below.
For more information, please contact Peter Baumann at pb718@york.ac.uk.
It's open to anyone 18+, regardless of previous experience with machine learning technology.
-------------------------------------------
-------------------------------------------
ABOUT THE STUDY (THE BELOW APPEARS AT THE BEGINNING OF THE SURVEY)
What is the study investigating?
The core aim of this research is to develop an understanding of the successes and pitfalls of current machine learning uses within the feature sets of music production software from the perspective of the end user, as well as gauging the level of consumer uptake of this technology. These results will then be used to highlight where future implementation might be most useful for music producers, enhancing both efficiency and quality of output. Alongside this, the research will be presented to filmmakers and game designers to understand the viability of automated music generation systems.
Who can contribute?
The study is looking for participants aged over 18 who are involved in any aspect of music production, filmmaking and game design. This could include those working in areas such as composition, mixing, mastering or audio restoration, visual media editing, at an amateur or professional level. Although the main focus of the research is aimed at those who have access to machine learning software, input from participants who do not use software with machine learning is also important in order to gauge consumer uptake, register concerns or communicate other relevant opinions.
What is machine learning?
In simple terms, machine learning is the process of presenting an algorithm with a set of sample data, which it then uses to calculate the ideal result of a given challenge. This training process might involve taking a collection of music for media and tasking the algorithm with creating its own music inspired by the training data. For those wanting a more complex explanation, the algorithms use deep neural networks to calculate the ideal result of a particular task using weight and bias values to assign importance to aspects of the input data, in order to achieve the most consistent, accurate results.
This technology can be implemented into numerous systems, such as facial or handwriting recognition, catalogue suggestions for music or visual media providers, online translation services, medical diagnoses and self-driving cars. Within audio processing, current implementation can be found in systems intended for audio restoration, mixing, feature extraction and composition. Companies such as iZotope (RX6 ‘De-Rustle’, Neutron ‘Track Assistant’), Sony (Flow Machines), Jukedeck and Melodrive all have products which are either in development or already available to consumers – all of which incorporate machine learning.
How will the research be conducted?
All responses will be collected via this online questionnaire hosted by Qualtrics. The results will then be analysed and used as part of a dissertation project into the implementation of machine learning technology in music production workflows.
Who is carrying out the research?
The research will be conducted by Peter Baumann, a student at the University of York studying for a Masters in Music Production under the supervision of Dr. Jez Wells (jez.wells@york.ac.uk).
How will the research findings be used?
Upon completion, the dissertation incorporating research findings will be made available to all participants.
Will my contribution be anonymous?
All participation is voluntary and participants may ask to withdraw their responses at any stage before September 1, 2017. All information provided by participants will be stored securely and handled in accordance with the ethical guidelines from the University of York. If you wish for your responses to be processed anonymously, please indicate this on the consent form below.
For more information, please contact Peter Baumann at pb718@york.ac.uk.
Please, if you can spare a few minutes, offer your voice to this project!
-------------------------------------------
-------------------------------------------
Last edited:
