|
akwok:
Thanks for your interest in the Echo Nest. Here are the answers to your questions:
1.) If a user's library is 100GB, and none of these files have been previously uploaded (by this user or anyone else), is it safe to assume that 100GB of upload bandwidth will be used to analyze this whole library?
Yes
2.) I know there is support for .wav, but is there support for .alac and .flac?
We support the same formats that are supported by ffmpeg. There's a list of supported formats here: http://en.wikipedia.org/wiki/FFmpeg. Flac is listed but alac is not. I'm not sure if we've tested with flac, but it should work.
3.) Is the MD5 hash generated completely independent of ID3 tags? That is, given I have two identical mp3 files, and I change the ID3 title tag of one song, will The Echo Nest give off two different MD5 hashes?
Currently, we use a full file MD5 hash, so differences in ID3 tags will indeed yield different MD5 hashes
4.) Is there a local, standalone version of the Echo Nest I can use for academic purposes? If the worst case requires uploading 100GB+ of music per user, it really isn't feasible to upload -- bandwidth is costly, after all. I read about a standalone Python build that was used with another academic project (Music Box) -- is this still available?
There is no local version available, but we lets talk about how we can make this work. Send me an email at Paul@echonest.com
Hope this helps.
Paul
|