remove masked arrays from salience

[rabitt]

Removes masked arrays, performs peak masking after harmonic summation.

---

This change is 

[bmcfee]

Review status: 0 of 1 files reviewed at latest revision, 2 unresolved discussions.

---

librosa/core/harmonic.py, line 90 at r1 (raw file):

        S_peaks = scipy.signal.argrelmax(S, axis=0)
        peak_mask = np.zeros(S_harm.shape).astype(bool)
        peak_mask[:, S_peaks[0], S_peaks[1]] = True

peak mask should now be the shape of S_sal/S.

You could also compute this directly by saying peak_mask = util.localmax(S, axis=0)

---

librosa/core/harmonic.py, line 91 at r1 (raw file):

        peak_mask = np.zeros(S_harm.shape).astype(bool)
        peak_mask[:, S_peaks[0], S_peaks[1]] = True
        S_sal[peak_mask] = 0.0

Shouldn't this mask out everything except peak_mask?

Also, the fill value should be user-specified, not hard-coded to 0.

---

Comments from Reviewable

[bmcfee]

Reviewed 1 of 2 files at r2.
Review status: 1 of 2 files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[bmcfee]

Reviewed 1 of 1 files at r3.
Review status: 1 of 2 files reviewed at latest revision, 1 unresolved discussion.

---

librosa/core/harmonic.py, line 44 at r3 (raw file):

    fill_value : float
        The value to fill non-peaks in the output representation. (default:
        np.nan) Only used if fitler_peaks = True.

fitler -> filter, = -> == (and put in scare-quotes)

---

Comments from Reviewable

[bmcfee]

---

Reviewed 1 of 1 files at r5.
Review status: all files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable