Fetched bunch of species-specific song snippets.

Worked those into LogHP analysis.
Worked results into fig_invariance_log-hp.pdf.
Put details into new fig_invariance_log-hp_species.pdf (appendix).

python/save_snippet_data.py

@@ -7,9 +7,10 @@ from IPython import embed
 ## SETTINGS:
 
 # General:
+search_target = '*'
 input_folder = '../data/raw/'
 output_folder = '../data/processed/'
-stages = ['raw', 'filt', 'env', 'log', 'inv', 'conv', 'bi', 'feat', 'norm']
+stages = ['raw', 'filt', 'env', 'log', 'inv', 'conv', 'bi', 'feat']
 if False:
     # Overwrites edited:
     stages.append('songs')
@@ -30,7 +31,7 @@ config.update({
     'rate_ratio': None,
     'env_fcut': 250,
     'db_ref': 1,
-    'inv_fcut': 5,
+    'inv_fcut': 10,
     'feat_thresh': np.load('../data/kernel_thresholds.npy') * 0.2,
     'feat_fcut': 0.5,
     'label_channels': 0,
@@ -40,7 +41,7 @@ config.update({
 ## PREPARATION:
 
 # Fetch WAV recording files:
-input_paths = search_files(ext='wav', dir=input_folder)
+input_paths = search_files(search_target, ext='wav', dir=input_folder)
 path_names = crop_paths(input_paths)
 
 # PROCESSING: