Onset and offset detection

[ThiCop]

Hello Thijs,

My approach was using something quite closely linked to your linear fits.

It was a combination of  local cross products (on a limited set of data points) between an example transition (which could be a linear or non-linear figure) and normalized measurement (local) data.

This can be sometimes quite sensitive to noise or false/incorrect "transition peaks/valleys".

If it works great, then I'm not going to overflow you with extra work. 😉

[ThijsBoeree]

Can you attach your example in my vi? I am very curious!
Regards,
Thijs

[ThiCop]

Hello Thijs,

I seem to have missed the e-mail notification of this forum post.

My apologies for the late reply.

On my work pc I have already deleted this test code.

If you want me to, then I can have a look at home to see if I backed it up over there.

[ThijsBoeree]

Hi again,

I came to the conclusion that curve fitting only works on certain data, my amplitudes are not always the same height and the steepness of the signal isn't allways the same...

So i've done it on another way, i take the deritative of a piece of the signal then normalise it between -1 and +1 then get the index of the max value, and search in both directions from that peak value when my value reach 5 % of the max value (0,05) i get the indexes and those are my on and offsets...

If anyone still have a better solution...?

Best regards,

Thijs