Dream999999
commented
4 months ago Due to uncertainty about whether N has an impact, I tried two scenarios: one without commenting out N=30*fs and the other with it commented out, using N = _next_power_of_2(label_frame.shape[0]) instead. Additionally, I added code in the update_label_frame function to call the _calculate_fft_hr function from the post_process.py file. The waveforms generated in the two different scenarios are slightly different, but the calculated HR values are the same. In the first waveform, the peak does not correspond to the calculated HR value, while in the second case, it does correspond. May I ask if my calling method is correct? Here are my waveforms, with both scenarios indicated in the frequency domain as well. Thank you for your assistance.
Hello, sorry to bother you. I recently encountered an issue while using the code in viz_preprocessed_data.ipynb for visualization. It seems that the setting for the value of N appears to be unreasonable. The function mentioned is supposed to generate data that meets the requirements for FFT input, so why is there a need to set N=30*fs? Could you explain the reasoning behind this setting?