Hi all,
In a first experiment we have been recording nirs data with a sampling frequency of 10 Hz. To correct from movement artefacts and so on we have applied the typical band pass filter [0-3]. In a new experiment, we ve doubled the sampling frequency (20 Hz) so we are going to work with a band pass filter [0-10]. Someone who has faced the same situation has a suggestion?
Thanks in advance =)
Guillermo
Dear Guillermo
I have seen same problem in my thesis, If your are designing your filter with matlab functions such as "remez" or you are using FDA TOOLS,do not change any parameter in your filter except sampling frequency or fs, after that MATLAB apply required effects automatically
Thanks for your answers, but could you explain a bit more? I d like to understand how the filter change according to the different sampling rates
Bandpass filter allows a band of frequencies to pass through, hence bandpass filter.
Your current bandpass filter is set to 10 Hz, which means it will allow any signal occurring at 10 Hz, to pass through unhindered while signal at other frequencies will be reduced/attenuated. (How much attenuation depends on the Q factor of the filter).
As you have moved from 10Hz to 20Hz, the previous settings will no longer be suitable and hence why you need to change the centre frequency of the bandpass filter to match the sampling frequency as otherwise, the filter may artificially reduce the signal that you want to observe.
A question that we should have asked is why 20Hz sampling rate was chosen, was it because 10 Hz was undersampling or for another reason? Ideally, the sampling rate should be 3 times or more than the actual frequency of the thing you wish to observe.
It is clearer to me now, I was indeed wondering about the need to change the centre frequency when recording at 20 Hz to avoid interference with the final output. On the other hand, the reason we ve increased the sampling frequency this time is to improve the power of posterior connectivity analysis.
Thanks
It depends on the frequency of the signal output, which hasn't been mentioned. The centre frequency of the bandpass filter should be the same as the frequency of the signal. The sampling rate should be at least double the frequency of the signal. However, as I have not seen your experiment, your mileage may vary.
The link will explain sampling rate and why it needs to be 2.7 times greater than the frequncy of the signal. https://en.m.wikipedia.org/wiki/Sampling_(signal_processing)
- Badges
- Science topic