Different between GUI representation and OpenBCI dataset

amukamuk lexington,ky
edited August 2020 in OpenBCI_GUI

I use the Notch filter(60 Hz) and Band pass filter(1-100 Hz) (butter worth order 4) in OpenBCI generated data
set. After applying filter in data I got a plot and try to have the same configuration in the GUI. When i match my both plot piece by piece I found there are some differences.
Can you please describe what thing I have to modify to get the same thing from the dataset?
I attached my plot figure here

Comments

  • wjcroftwjcroft Mount Shasta, CA
    edited August 2020

    Amuk, hi.

    Not all DSP (digital signal processing) filters are constructed identically. I believe GUI is now using Brainflow for filtering, is that what you are using? Example in Python (see menu at left for other languages):

    https://brainflow.readthedocs.io/en/stable/Examples.html#python-signal-filtering

    I'm also not sure what order filters are specified in the GUI. Mentioning Richard @retiutut.

    Note the scale shown on the two graphs is also quite different. Your graph shows about plus and minus 60 uV. Whereas GUI lower right shows about 20 uVrms. Instead of letting the GUI scale use 'Auto', you can set it to a specific value for more ease in comparison.

    Regards, William

  • amukamuk lexington,ky

    Thank you William for your reply.
    I applied filters from python scipy module, I didn't use the Brainflow filters. I will apply filters from Brainflow.
    I have some confusion about the scale also. Is it peak to peak?
    For more clarification when i select vertical scale 100 uV, does it mean +100 uV and -100 uV for y axis or +50/-50?
    Thank you.

  • wjcroftwjcroft Mount Shasta, CA
    edited August 2020

    GUI "Vert Scale" drop down menu, I believe is exactly what it says, the uV per full scale (positive deflection). Similar to what you see on oscilloscope controls.

    uVrms, is an RMS calculation,

    https://en.wikipedia.org/wiki/Root_mean_square

  • retiututretiutut Louisiana, USA

    Maybe we can try to fit some Time Series Y axis labels. B)

Sign In or Register to comment.