Input for Online Onset Detection

[KimayaKerkar]

Hi, I have gone through the resources for online modes of onset detection, and I have tried to implement the online method proposed in "Evaluating the Online Capabilities of Onset Detection Methods" Sebastian Böck, Florian Krebs and Markus Schedl. Proceedings of the 13th International Society for Music Information Retrieval Conference (ISMIR), 2012.
I'm running into 2 errors

1. No onsets are getting detected.
2. I can't control the time the program runs for. With offline methods, the program runs till the entire audio file is processed. But I'm not sure how to control it for an online implementation.
   I'm sharing my code below, if anyone has any advice please do share.

#!/usr/bin/env python3
from madmom.audio.spectrogram import FilteredSpectrogramProcessor, LogarithmicSpectrogramProcessor
from madmom.audio.filters import LogarithmicFilterbank
from madmom.audio.signal import Stream
from madmom.features import OnsetPeakPickingProcessor, SpectralOnsetProcessor
import numpy as np

stream = Stream(
    sample_rate = 44100,
    num_channels = 1,
    frame_size = 2048,
    fps = 100
)
'''
filt = FilteredSpectrogramProcessor(
	filterbank=MelFilterbank,
	num_bands=12, 
	fmin=30, 
	fmax=17000,
	norm_filters=False
)

log = LogarithmicSpectrogramProcessor(
	mul=1
)
'''
odf = SpectralOnsetProcessor('spectral_flux', filterbank=LogarithmicFilterbank, num_bands=12, log=np.log10, sample_rate=44100, frame_size=2048, hop_size=512
)

peak = OnsetPeakPickingProcessor(fps=100, threshold=1, pre_max=0.01,
                                            post_max=0.00, pre_avg=0.15,
                                            post_avg=0, combine=0.03, delay=0)

onsets = []

for frame in stream:
	of = odf(frame)
	print(of)
	o = peak(of, reset=False).round(3)
	print(o)
	onsets += o.tolist()
print(onsets)