deprecate scikits.audiolab in favor of librosa

[dmtrs]

.//datasets/music/new_experiment16k.py:import sys, os, subprocess, scikits.audiolab, random, time, glob
.//datasets/music/new_experiment16k.py:arr = [(scikits.audiolab.flacread(p)[0]).astype('float16') for p in paths]
.//datasets/music/new_experiment16k_conditioning.py:import sys, os, subprocess, scikits.audiolab, random, time, glob, math
.//datasets/music/new_experiment16k_conditioning.py:samples = np.array([(scikits.audiolab.flacread(p)[0]).astype('float16') for p in paths])
.//datasets/music/new_experiment32k.py:import sys, os, subprocess, scikits.audiolab, random, time, glob
.//datasets/music/new_experiment32k.py:arr = [(scikits.audiolab.flacread(p)[0]).astype('float16') for p in paths]
.//datasets/dataset_conditioning.py:            #data, fs, enc = scikits.audiolab.flacread(path)
.//datasets/dataset.py:            #data, fs, enc = scikits.audiolab.flacread(path)