Vocalpy

Fixed
- Fix how we compute cepstrum in `vocalpy.spectral.sat`
[123](https://github.com/vocalpy/vocalpy/pull/123).
Fixes [122](https://github.com/vocalpy/vocalpy/issues/122).

Added
- Add `trough_threshold` parameter to `feature.sat.pitch` and `feature.sat.similarity_features`
[117](https://github.com/vocalpy/vocalpy/pull/117).

Fixed
- Fix how `find_hits` matches hits
[120](https://github.com/vocalpy/vocalpy/pull/120).
Fixes [119](https://github.com/vocalpy/vocalpy/issues/119).

Changed
- Vendor code from evfuncs, reduce number of dependencies
[121](https://github.com/vocalpy/vocalpy/pull/121).
Fixes [118](https://github.com/vocalpy/vocalpy/issues/118).
- Bump minimum required version of crowsetta to 5.0.2 (that also vendors code)
[121](https://github.com/vocalpy/vocalpy/pull/121).

Added
- Add convenience function `vocalpy.spectrogram`, that computes a spectrogram from
a `vocalpy.Audio` using a specified `method`, and then returns it as a `vocalpy.Spectrogram`
[109](https://github.com/vocalpy/vocalpy/pull/109).
Fixes [88](https://github.com/vocalpy/vocalpy/issues/88).

Fixed
- Fix hard-coded value in `vocalpy.spectral.sat` that should have been `n_fft` parameter
[111](https://github.com/vocalpy/vocalpy/pull/111).
Fixes [100](https://github.com/vocalpy/vocalpy/issues/100).

Added
- Add `feature` module with functions to extract
[Sound Analysis Toolbox](http://soundanalysispro.com/matlab-sat)
features [85](https://github.com/vocalpy/vocalpy/pull/85).
Fixes [3](https://github.com/vocalpy/vocalpy/issues/3).
Also adds a `spectral` module with a `sat` function that computes
the spectral representations used by the functions in
`vocalpy.features.sat`.

Fixed
- Fix return values for `vocalpy.metrics.segmentation.ir.precision_recall_fscore`
[78](https://github.com/vocalpy/vocalpy/pull/78).

Changed
- Rename `vocalpy.segment.energy` to `vocalpy.segment.meansquared`,
and rename `vocalpy.signal.audio.energy` to `vocalpy.signal.audio.meansquared`
[76](https://github.com/vocalpy/vocalpy/pull/76).
Fixes [75](https://github.com/vocalpy/vocalpy/issues/75).

Fixed
- Add back `vocalpy.signal.audio.bandpass_filtfilt`;
have `vocalpy.signal.audio.meansquared` use this function
(and by extension `vocalpy.segment.meansquared`)
[76](https://github.com/vocalpy/vocalpy/pull/76).
Fixes [74](https://github.com/vocalpy/vocalpy/issues/74).