Wow! I can't believe all the responses. Thanks so much to everyone!

I'll try to answer everyone's questions. The basic MP3 player is built. I 
created it for my first project in the course, so this final project is meant 
to expand its capabilities. It currently uses getNthFileName and refreshes the 
music library from the set music directory. My group members are unreachable 
this week due to the holidays but I believe we plan on hosting a diverse 
selection of songs online for the player to stream, so technically, it only 
needs to work for those songs.

Our background is VERY limited. One group member has a little experience 
w/ "neural networks"...rather my prof's definition of one. I did a lot of the 
MP3 player functionality, so I'm really good w/ lists. I would consider all 4 
of us novice-intermediate level. Math skills are far from grand...:)

I honestly have no idea what my professor meant by DRM. I'm thinking he must 
have been using the term loosely. As far as security, we just want to make 
sure that the user can only stream and not download the songs.

My professor said comparing songs by genre tags is not an option. His 
expectation for the scope of this project is basically the sound equivalent of 
image recognition in director.

In all my searching, I never came across the Amplitude Xtra or asFFT Xtra. 
I'll have to speak to my professor to see if he has either for our use. It 
seems like it would help our problems a lot. I hadn't even thought of the wave 
visualization. I think that would impress him...he seems to just want to 
be "wowed". Oh, and I believe we have to base the comparison off the entire 
song, so if the song is faster in more spots than slow, i guess the fast would 
win over. It doesn't seem like a live search is possible, so we'd almost have 
to create a DB, which someone suggested, for every song that is searchable and 
have every song pre-analyzed. I have worked with ASP and XML w/ director.

Warren's suggestion for an alternative approach to the project makes sense to 
me in theory, but I'm not sure I completely understand how it would work. Are 
you saying that w/ a wide enough sampling of different genres already included 
in each of the networked programs, there should be a similar waveform that 
could be matched?? I don't know much about waveforms for different genres. Are 
they similar enough in form to get a match? I guess jazz is the only genre 
that seems to have a good structure to it.

As for rights to the project, I don't think my professor plans to steal our 
work. He claims to understand all the CS aspect, but has never shown us any of 
his own work. To my knowledge, he's done no work remotely related to this, so 
he can't be of much assistance. I think you're on to something tho when u said 
he's trying to get us to do things he could never figure out.

Buzz--Could you give more details about what u mean by users querying for 
THEIR assessment of similar songs?? Do you mean have them choose what genre it 
is or something and search based on that? As futile as it seems, I think we're 
required to do the query by analyzing the song playing...

Thanks again for everyone's suggestions. I'm going to have to do a LOT of 
reading for the rest of my thanksgiving break. I'll take all this info to my 
group as soon as I can meet with them, and I also hope to talk to my professor 
on monday. All your information should be helpful when I talk to my professor. 
Hopefully we can work with him to work out a little different project spec 
that makes more sense and is more usable. I'd hate to invest so much time into 
something that doesn't yield a usable project.

I'm trying to think of more alternatives for the project, b/c as i mentioned, 
i think he's looking for the wow factor, and he prolly won't even take time to 
look at our code. Does anyone have any other ideas like album art or 
something?? I've been searching for a large DB that we could query and find 
album art automatically. He also mentioned have skins when we did the first 
MP3 player. He just keeps talking about having a "neural network". I don't 
believe it's possible to learn AI in a week w/ our limited programming 
experience.

Well, I apologize for the massive response, but there were so many different 
emails that i wanted to have the responses all in one place. 

Hope everyone is having a wonderful thanksgiving.
Cheers,
Mindy
[To remove yourself from this list, or to change to digest mode, go to 
http://www.penworks.com/lingo-l.cgi  To post messages to the list, email [EMAIL 
PROTECTED]  (Problems, email [EMAIL PROTECTED]). Lingo-L is for learning and helping 
with programming Lingo.  Thanks!]

Reply via email to