forked from Rudd-O/python-audioprocessing
-
Notifications
You must be signed in to change notification settings - Fork 0
/
CHANGES
32 lines (24 loc) · 1.22 KB
/
CHANGES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Version 0.0.6:
* Apparently, when passing all zero values in the chunk to calculate_dB
makes it fail with a ValueError (math domain error). Now I am trapping
that just like I'm trapping OverflowError.
Version 0.0.5:
* Reorganized the code into separate submodules
* Implemented a streaming model that does not require decoding of the
entire file and does not use temporary files
* Boosted song analysis performance by 2X with the streaming model
Version 0.0.4:
* Tuned default parameters for optimum correlation coefficients
* Cleaned up the library, splitting graphing into a separate workbench module
* Added more options to the command-line programs
* Documented some functions, fixed some bugs
* The API remains unchanged, but it will change in the future
Version 0.0.3:
* Added the ability to plot signatures in 3D (spectral analysis)
* Added a batch operator that can store the results in a GDBM database
* Removed butterscotch-correlate to be replaced in the future by a better
alternative
Version 0.0.2:
* Changed default signature to 32 bands [0..11025] and 12 10-second samples
* Changed correlation functions to return correlation average
* Verified with threshold 0.9 - correlation works reliably as duplicate identifier