view examples/iAudioDB/README @ 770:c54bc2ffbf92 tip

update tags

Using CAMUS
===========

1. Installation

To install CAMUS, simply drag the application icon onto your
Applications folder. It should then run from this location.

2. Functionality

Creating a new database
=======================

1. Choose New from the File menu.
2. Select the feature type for the database: pitch allows for
comparison based on harmonic or pitch content; timbre based on a
broader measure (e.g. for non-pitched instruments).
3. Specify the maximum number of tracks you would like to store in
the database.
4. Specify the length, in seconds, of the longest track (e.g. if
your longest track is 5 minutes, choose 300).
5. Specify the slice size in milliseconds.
6. Press Create, and then enter the filename where you would like
to store your database. Press Save to create it. The file will have
a .adb extension, and an associated file with a .plist extension:
if you call your database mydb.adb, there will be mydb.plist alongside
it. myadb.plist contains the database configuration and content
information.

Adding files to an open database
================================

1. Either click 'Add' from the CAMUS toolbar, or 'Import Audio'
from the File menu.
2. Select the audio files you would like to import. You can select
multiple by holding down Alt or Shift when clicking. Audio files
may be MP3, WAV, AIFF, or un-DRMed M4A.
3. Click 'Open' to import.

Querying the database 
=====================

1. Click 'Query' on the CAMUS toolbar.
2. Click 'Select File' and choose the file to query against. This
must already be present in the database.  Once loaded, Query Start
and Query Length are populated with the maximum possible values.
3. Select the query start point, either in vectors or seconds. This
is the time into the track at which the query will begin.
4. Select the length of the query, again in either vectors or seconds.
5. If these values are invalid, the Query button will be disabled.
Press Reset to revert to the very beginning and the length of the
track.
6. If you would like to only retrieve the best match for each track,
uncheck the 'Find multiple matches within tracks' checkbox.
7. Press query to carry out the query.

Exploring the results
=====================

By default, results are sorted in ascending order of distance. As
such, the first match will likely be the query itself. The iPos
column specifies the position in the track in seconds, and the bars
in the left-most column show the distance in graphical form as a
(non-scientific) visual indicator. To play a result, either
double-click its entry, or select it and click 'Play Result' in the
toolbar. To stop playback, press the stop button while playing. To
play the query and the result at the same time, select the result
and click 'Play Both' in the toolbar.

3. Behind the Scenes

CAMUS makes use of audioDB (http://omras2.org/audioDB) for similarity
matching and Sonic Annotator (http://omras2.org/SonicAnnotator) to
extract features from audio via its Vamp plugins
(http://vamp-plugins.org/). librdf and its associated libraries are
bundled to handle the import of features in N3 format.

4. Feedback

To report bugs or feedback, please email [e-mail].

Requirements:

* CAMUS requires BWToolkit, which was not included for space reasons.
It is freely available at http://brandonwalkin.com/bwtoolkit/. The
BWToolkitFramework.framework file should be placed into this folder
prior to recompilation.
* All other libraries are included, with their binaries patched to
allow for their bundling within the CAMUS application.