Unexpected performance in melody tracking

[zumpchke]

Hi, I have an audio clip with a pretty clear synth based melody but the melodia is unable to track it.

Here is my code

from mir_eval.sonify import pitch_contour
import numpy
import essentia.standard as es
import matplotlib.pyplot as plt

sample_rate = 44100
loader = es.MonoLoader(filename="better_off_alone_segment.mp3")
audio = loader()
print(f"Duration of the audio sample [sec]: {len(audio) / sample_rate:.3f}")
# PitchMelodia takes the entire audio signal as input - no frame-wise processing is required here...
pExt = es.PitchMelodia(frameSize=2048, hopSize=128)
pitch, pitchConf = pExt(audio)


time = numpy.linspace(0.0, len(audio) / sample_rate, len(pitch))
## plot the pitch contour and confidence over time
f, axarr = plt.subplots(2, sharex=True)
axarr[0].plot(time,pitch)
axarr[0].set_title('estimated pitch[Hz]')
axarr[1].plot(time,pitchConf)
axarr[1].set_title('pitch confidence')
plt.show()

from tempfile import TemporaryDirectory
temp_dir = TemporaryDirectory()


synthesized_melody = pitch_contour(time, pitch, 44100).astype(numpy.float32)[:len(audio)]
es.AudioWriter(filename = temp_dir.name + 'melody.wav', format='wav')(es.StereoMuxer()(audio, synthesized_melody))

print(temp_dir.name + 'melody.wav')

better_off_alone_segment.mp3.zip

Any ideas what I can do to improve performance?