Help me come up with an idea for my final year university project pls

#1
It's embarrassing that it's got to the stage where I feel like asking the pit for assistance is an option worth exploring, but I've had over a year to think about this and I've got nothing.

So here's the situation:
I start my final year of uni in like 3 weeks and I need a reasonable idea for a project. It's equivalent to two modules but really it's worth more than that. The reason for that is that I'm doing my BA in Music Technology, and let's face it, it's a real bullshit course and gives me basically zero prospects for the future, not to mention I have very little interest in music any more. So what I'm hoping is that I can do a project which overlaps between music and something better so I can do my masters in something more interesting, and something that'll get me on a decent PhD.

The only requirement for this project is that it has to have something to do with sound. Like it can be pretty much anything as long as it deals with sound in some way. And my requirement is that it involves something more interesting in the vague region of computer science, maths or physics (but really I'm interested in any idea at all). I'm leaning towards research as my future career so if it's something I could write a paper on that'd be great. I'd rather not have to make anything significantly large, like an entire program, but ideally I'll be creating something as part of it.

Here's an email from my tutor just to give you an idea of the scale of project they're looking for:
For the Individual Project, the possibilities are endless! This is a very demanding module, yet most students get fantastic marks as they can use our facilities and staff support to make what they want to do in life for real! We need you now to think of what you want to do. At the moment, we need a one-liner project description from you, with a supervisor if you really have a strong opinion on who you want.

The most popular option is the EP of original music, usually 24 to 30 minutes long. Styles vary, from technical metal to glam rock, from experimental noise to singer-songwritter.

The second most popular option is a recording portfolio as a producer, consisting usually of two to three projects for a total of 40-45 minutes.

Other popular options are original AV projects (film, video, animation) of about 12-15 minutes, technical design of electronic tools, or software design... but feel free to propose anything for now, just tell us what you want to do.

So here is a couple of examples of what I need from you on that day:

- I want to record a nuevo-flamenco EP with my band. I composed all the music and will engineer it. I want Father Ted as tutor if possible.

- My project is a new software design, based on a new approach to granulation. It will be a standalone for all platforms and I want to commercialise it.

- I want to make a portfolio of musical genre I can compose for image. I plan to take 10 three minutes long extracts from commercial films and record music to them.

-My project is to be a producer on 3 projects. I'll record a singer-songwritter, a rock trio and a pop quintet, and will be assessed on their musical improvement, as well as the sound quality of the recording and the mix.

I hope you get the point by now ;-) On the day, I will introduce you to assessment criteria and what you can do during the summer to get ahead of the schedule. Summer is a great time to get inspiration…


I don't even need a full idea from you. Any little spark of inspiration you might have could help me.

So far I've been leaning towards something artificial intelligence related, specifically machine learning. In English: teaching a computer to do something, I just don't know what. For example, at the place I worked in Paris, there was a project that aimed to create a series of statistical models of sound textures like wind and rain (I'm not doing this because a whole team of researchers couldn't even get decent results so fuck that). There was also a project that I was on, that aimed to identify specific sound events in mixes by learning what they sound like from a set of samples.

So hmu if u hav any ideas yeh
#2
Do the project on soundwave interpretation. That would be as simple as writing an AI that can translate the soundwaves into letters. Given enough time and development, its accuracy could become an auto-writing AI, which would make typing redundant. Might also force teachers and (primarily) parents to do their ****ing job and teach the kids pronunciation.
#3
I think you work with Max/MSP and Csound correct? You should create an interactive device for music/sound generation via Kinect or a similar device. I think I showed you some videos of my colleagues who did similar things but if you want I can pull one up for you.


edit: how about something with Leap?

...modes and scales are still useless.


Quote by PhoenixGRM
Hey guys could you spare a minute to Vote for my band. Go to the site Search our band Listana with CTRL+F for quick and vote Thank you .
Quote by sam b
Voted for Patron Çıldırdı.

Thanks
Quote by PhoenixGRM
But our Band is Listana
Last edited by Xiaoxi at Sep 3, 2013,
#5
Try to explain what Boss Metal Zone does to your guitar sound.
Quote by ChemicalFire
You get my first ever lolstack






The image in my head is just too funny for words at this point


Aw yeah.
#6
Quote by Lord_Doku
Do the project on soundwave interpretation. That would be as simple as writing an AI that can translate the soundwaves into letters. Given enough time and development, its accuracy could become an auto-writing AI, which would make typing redundant. Might also force teachers and (primarily) parents to do their ****ing job and teach the kids pronunciation.

I assume you mean speech recognition? It's a pretty complicated thing, and it's been done by tons of people much better than I could hope to achieve in a year, unfortunately.

Quote by Xiaoxi
I think you work with Max/MSP and Csound correct? You should create an interactive device for music/sound generation via Kinect or a similar device. I think I showed you some videos of my colleagues who did similar things but if you want I can pull one up for you.

edit: how about something with Leap?

Aye I do max and such, and kinect/wiimote/arduino things are one of the things they suggest. The problem with something like that is it'd have to be something really creative and I can't think of anything. Also, just making a straight Some sort of interface that learned gestures of some form maybe (not necessarily visual). I'll think about it and how feasible it could be.

What's leap?

Quote by WhiskeyFace
Are you me!?

I think so, aye. Got the same problem?
#8
I recently finished a degree in Music and Sound Technology, and for my dissertation I analysed the writing devices and guitar techniques in a couple of select albums from a genre, and wrote, recorded, mixed and mastered my own 5 track EP within the genre using the information I had gathered.

You could do something like that. Other people did genre studies with no artefact, one guy made a plugin using MAX, a couple did soundtrack related pieces but I don't know if that is part of your course.
#9
So my background is linguistics, which has no crossovers with what you're doing APART FROM this guy David Pesetsky (King of Generative Syntax) who wrote a whole load of papers on whether music and language are the same thing / represented the same way in the brain.

Here's one
http://web.mit.edu/linguistics/people/faculty/pesetsky/Katz_Pesetsky_Recursion_Conference_Music.pdf

Any use?
Hahahahahahahahahah hahahaha har har har
#10
Make one of those songs that goes beeeepp bwop bwop boop boop boop ring ring beep beep bwop bwop beep ring boop bop bop dwop beep ring ring bzoongaly zippity boppity bwoop
#11
Try and make a connection between specific sounds and fourier transformations. Not sure there is one, but hey. Or, get a second order ODE, graph it, translate it into a sound wave and see what connection various operators have on the audible sound, curl, divergence, gradient etc. Again, I doubt you'll find anything interesting.
#12
Quote by captainsnazz
The problem with something like that is it'd have to be something really creative and I can't think of anything.
It don't matter what you do in life--creativity is always the moneymaker.

What's leap?

https://www.leapmotion.com/

...modes and scales are still useless.


Quote by PhoenixGRM
Hey guys could you spare a minute to Vote for my band. Go to the site Search our band Listana with CTRL+F for quick and vote Thank you .
Quote by sam b
Voted for Patron Çıldırdı.

Thanks
Quote by PhoenixGRM
But our Band is Listana
#13
Quote by willT08
Make me a suite of plugins I wish I had

It's a shame all DAW plugin APIs are a pain in the fuckin arse because I absolutely would make a better version of every plugin you have with sidechain possibilities on absolutely everything.


Quote by Random3
I recently finished a degree in Music and Sound Technology, and for my dissertation I analysed the writing devices and guitar techniques in a couple of select albums from a genre, and wrote, recorded, mixed and mastered my own 5 track EP within the genre using the information I had gathered.

You could do something like that. Other people did genre studies with no artefact, one guy made a plugin using MAX, a couple did soundtrack related pieces but I don't know if that is part of your course.

Yeah this sounds pretty similar to the sort of stuff people are doing on my course, the issue is that I need something that will help me get into a different field for my masters. Plugin building is the only option for me there (though it'd most likely be a standalone thing because fuck plugin APIs but that's irrelevant really), and it is an option I'd like to consider. It's just what plugin to make that's the problem. If I could find a concept for an instrument/effect that uses machine learning or something in some way that'd be great.

Quote by McTodd
So my background is linguistics, which has no crossovers with what you're doing APART FROM this guy David Pesetsky (King of Generative Syntax) who wrote a whole load of papers on whether music and language are the same thing / represented the same way in the brain.

Here's one
http://web.mit.edu/linguistics/people/faculty/pesetsky/Katz_Pesetsky_Recursion_Conference_Music.pdf

Any use?

This is absolutely relevant to my interests, but I really know very little about linguistics so I don't know what I could hope to achieve with it in a year. Maybe it's worth considering some sort of grammar-informed approach to algorithmic composition. Idk actually maybe that's something I could do. Generative composition is really a lot of work though. See: my mostly failed attempt at making a dubstep generator for a much smaller project a couple of years ago/for lulz.

Quote by FearMyLightning
Make one of those songs that goes beeeepp bwop bwop boop boop boop ring ring beep beep bwop bwop beep ring boop bop bop dwop beep ring ring bzoongaly zippity boppity bwoop

That's will job m8

Quote by conor-figgy
Try and make a connection between specific sounds and fourier transformations. Not sure there is one, but hey. Or, get a second order ODE, graph it, translate it into a sound wave and see what connection various operators have on the audible sound, curl, divergence, gradient etc. Again, I doubt you'll find anything interesting.

Yano the fourier transform of a time-domain signal just gives a frequency-domain representation of it? When analysing a complex sound it's always useful to consider it in the frequency domain anyway. What you're talking about is analysing audio output of mathematical functions and yeah it can be interesting, but it's just getting a specific interesting thing with enough substance to test and write about.

Quote by Xiaoxi
It don't matter what you do in life--creativity is always the moneymaker.


https://www.leapmotion.com/

✿◕ ‿ ◕✿ thank you ✿◕ ‿ ◕✿
#14
Say you want to create some new age relaxation music with the possibility of commercializing (can talk about it's social importance, blah, blah).
Just play a bunch of suspended chords on some synth pads, chuck in some nature samples and call it a day.

pretentious small text, right justified signature
UG's professional coffee nerd
also UG's musical theatre nerd
roscoe's wetsuit