Final year project - Not sure which project to choose

This thread is probably for I.T people only

I have been looking at the list it includes some interesting things and HIFI releated stuff:

One of them is to create an algorithm for an audio compression format then use C++ to write a program that will compress .wav file. This is far beyond abilities

Another one is to write a website that calculates the optimium position of speakers in a room according to where the listener is sat

These both seem quite hard, another is to produce a site teaching you how SACD (partly invented at my university for Sony) and DVD-A.

However all these a bit too difficult, one that cought my eye was to produce a simple robot that can be controlled over the internet the robot must use the USB interface and be driven by a stepper motor. The program basicaly controls the stepper motor but must be controlled over IPV4.

One project that cought me eye was to create a virtual learning evironment that can also be operated offline and then uploaded to the central database later. This seems to have some difficulty but perhaps dosn't involve as much maths.

The other projects are stuff like create an online shop in ASP.NET but I don't this will be much of a challange.

Basicialy should I go for an easy project that I will do well in or do somthing more ambigitous but not do so well? The later will have scope for more marks but I am not too sure.

 

I'd do the audio compression one. The framework is trivial, MS do most of it for you, and I can think of three of four straightforward compression algorithms right now. Even if they don't work you can do the experiment with a range of source material, produce the result and explain why it didn't work. You also have professional lossy and lossless to compare to. Nothing real time and no hardware to source or have flake out on you.

Paul

 

O'Reilly used to publish a very good book about compression algorithms, from the basics to some really interesting stuff. Don't know if it's still in print. Seems like the most interesting project out of that lot.

-- Ian

 

I'd go for the compression one too. It's the mostly tightly defined and relatively easy to do. Afterall, you don't have to write something that competes with MP3 - as long as it does some useful compression I'm sure that'll be acceptable.

Michael.

 

i taught myself c by writing a huffman compressor / decompressor. took about a week to write it and get it working.
cheers


julian

 

Me and a work mate one wrote an LZW compressor/decompressor in two weeks - in 68020 assembler! It's not as hard as you think.

 

[/geek mode]
mmmmm, 68k series processor assembly language, gggggg (homer style drool gargle)
i miss the days of huge instruction sets, all this risc bollocks is tedious.
[/geek mode]
cheers


julian

 

I once wrote a basic RLE (run length encoding) image compressor/decompressor in 6502 assembler for the BBC Micro. I was running out of space to store my Mandelbrot fractal images I was generating.

Julian, I bet today's RISC processors still have a larger instruction set than the old 6502 which wasn't a RISC processor! I remember when I did a bit of Z80 and then later 8086 that their instruction sets were HUGE in comparison.

Michael.

 

Writing in assembly language sucked. I hated it! Bleeding impossible to debug.

Mind you, I hate programming full stop. If something's been written to do a job, why re-invent the wheel and do it a different way (and balls it up and then have to debug it!?)?

That said, I'd go for the compression algorithm too, although...

...IF you can write a site that can explain why cloth-eared non-audiophool people should ditch their CDs and upgrade to a format that sounds the same (to them anyway) which will actually convince them to change, that'd be worth a degree on its own! You might make some dosh out of Sony

 

yeah 6502 was only 255 instructions iirc - i went from loverly 68k on the amiga / genesis / st to 65816 on the snes which was a nightmare.
dom,
the tools got pretty good there for a while before c took over. i can remember debugging with nothing other than some colour bars and some nops. when we got a dissasembler that allowed single stepping it was amazing. it's all a matter of perspective i guess.
cheers


julian

 

He he he - I hope so, 'cos coding by hand in hexi (iirc) and then trying to debug using the stack registers etc was a nightmare.

Programming a UART was pretty cool though

The best bit of uni was programming one of the infamous Reading Uni Seven Dwarf robots. The damn thing got stuck in a loop (I was trying to be overly clever) and started chasing it's own tail, round and round in circles. Pure class!

 

I have 'The Data Compression Book' by Mark Nelson. Published by M&T in 1991 and complete with two 5.25" floppies...

Not great on audio compression.

Paul

 

Have you considered searching through the open source projects held on SourceForge ? I reckon they have a whole mass of compression and audio related projects that might give you a few ideas about what others are doing in that field of work.

 

Look, if it's going to be important to your grades and is going to occupy your time for some weeks then my advice would be to go for whatever floats your boat.

If the compression thing is your thing, then do it. From what people have been saying, there are ways of making it less difficult. If you don't fancy it go for the virtual learning environment or the robot.

The worse thing would be to get involved in a project that doesn't interest you.

 

If you have some idea of the kind of work you like to do after uni then do a project that will impress a future employer.

 

It turns out both of the robot and compresison thing is done by a supervisor who won't help you much. Basicaly I have been told to avoid him at all costs by a lecturer and fellow students. I have invented my own project now which is a sales website where you select PC conponents such as a CPU then it lists the compatable motherboards etc it will eventualy give you the price of the PC. I will add more complex stuff into it like maybe webservices which work out a say home installation costs according to your postcode etc.

I have to wait for it to be approved though. I want to do it in ASP.NET and MS SQL Server probably in XML format.