mkb218 / opera-omnia Goto Github PK

Right now this is locked to a single instance. How can I get multiple instances to share a single datastore?

When using ffmpeg->libfaac, sox can't read from stdin.

Can probably just use libfaad2 from go-land for this.

Then, why not use libsndfile for everything else (except MP3, guess I could use libsox or mpg123)

In fact, I want to remove all subprocesses and use library calls for encoding and decoding.

I don't know what this is all about, or if it's even real.

I uploaded PTA's Now's Tomorrow and it never reached the RequestQueue. Haven't figured out why.

Monstrous piece of work, may need to do this on demand.

Allow uploading multiple files at once.

Right now, if you upload a track and mark it both "add to library" and "play this back" it will be played back only with segments that were uploaded previously.

Right now, only Echo Nest data supplies artist and title for the attributions page.

Don't just allow echo nest ID playback (untested)

Also allow track search. Only for playback!

Need to list content that has been added to the library.

Write output to a SoundCloud account, take files from SoundCloud followers.

What you typed out on G+ is probably good enough for now :-)

If a track is uploaded for adding, make sure that we don't re-add the segments if they're already there.

Need to save the segment library and md5 -> id -> analysis mappings to gobs after every update.

Audio Request segments should be:
struct { root, s *Segment }

Right now, for each root segment in the analysis of the target track, we re-sort the set of all segments by the root. This gets ridiculously slow once the library has about 5 tracks.

One approach to make the search space smaller is this:
Pick a few attributes and thresholds. These can be static or dynamic based on the root (three largest pitch or timbre members, let's say). We filter the search space before sorting, by these criteria

As the number of segments grows, the thresholds get tighter.

Echo Nest doesn't do FLAC, will need to rencode.

Listen to all (or most recent) dumped output from a drop down.

Some things have gotten messy.

Why doesn't this work?

If the server crashes, it needs to restart itself.

Should be easy