Skip to content

Releases: TheCacophonyProject/audio-analysis

v1.26.2

13 Nov 14:42
Compare
Choose a tag to compare
fix error returned

v1.26.1

12 Nov 14:04
Compare
Choose a tag to compare
fix docker file

v1.26.0

12 Nov 14:02
Compare
Choose a tag to compare
add github workflow

v1.25.0: re raise ex (#17)

14 Jul 22:25
bbf1835
Compare
Choose a tag to compare
Co-authored-by: gferraro <[email protected]>

v1.24.3: Analyse tracks (#16)

18 Jun 02:25
4e8ac57
Compare
Choose a tag to compare
* load values from metadata and allow for slightly longer last cacophony index bin

* add support for new audio models

* filter morepork for now

* black

* use arg parse to tidy up commands

* bring back ref

* update tensorflow

* add probability

* fix bug with short recs

* add chirps

* fix formatting

* update reqs

* update opencv

* just use ubuntu image

* dont filter moreportks

* dont let chirps overlap, tighten tracks

* check for track index

* should =

* remove print

* change tracks back

* fix short tracks

* fix cacohpony index

* ability to use embeddings model

* multiple bird models

* reomve prints

* remove print for real

* fix end track bug

* return 0 chirps

* yamnet model support

* add new chirps

* make docker image contain everything needed

* load tracks

* remove prints

* make tracks

* enlarge by 1.4

* adjust overlap

* remove small tracks

* fix chirps

* adjust chirps and track generation

* read proper meta

* fix confidences and add max chirps

* add a best tag when no predictions meet the thershold

* use int

* fix confidence

* multiple predictions per track

* fix for embeddings

* update to use multiple models per a track

* remove prints

* update cacophony index

* update index

* update max chirps

* use old cacoph index

* fix no tracks

* fix sample size beeing slightly wrong

* ensure we dont get more than sample size

* add butter worth filtering

* add freq to morepork and only butter low freqs

* reduce min track length

* enlarge tracks to 0.7 seconds if smaller

* make freq an int

* fix name

* check for 0 max chirps

* filter tracks below 1000

* add latest

* also enlarge mel

* analyse tracks

* update for json keys

* dont do index on insect

* fix index

---------

Co-authored-by: gferraro <[email protected]>

v1.24.2

27 May 19:43
Compare
Choose a tag to compare
dont do index on insect

v1.24.1

24 Apr 01:57
Compare
Choose a tag to compare
update for json keys

v1.24.0: Analyse tracks (#15)

16 Apr 06:24
c0a7343
Compare
Choose a tag to compare
* load values from metadata and allow for slightly longer last cacophony index bin

* add support for new audio models

* add chirps

* adjust chirps and track generation

* reduce min track length

* analyse tracks

---------

Co-authored-by: gferraro <[email protected]>

v1.23.8: Generate tracks (#14)

20 Nov 02:19
1383654
Compare
Choose a tag to compare
* use chirps

* generate tracks from audio

* filter tracks below 1000

---------

Co-authored-by: gferraro <[email protected]>

v1.23.7

08 Nov 02:11
Compare
Choose a tag to compare
add latest