# Locating the start of a curve

1. ## Locating the start of a curve

Hi,

Im gathering a lot of data for a project currently, we have a lot of excess data at the start of each recording and need to find the start point which changes on each test, iv attached a copy of some data

ST

2. ## Re: Locating the start of a curve

How much of this is generic signal processing (how does anyone using any programming language identify "start of activity" in a signal) and how much is specific to Excel (I know how to identify "start of activity" by other means, but I do not know how to program that algorithm into Excel). You have the point in row 244 marked as "start of activity". How did you decide that this point was the start of activity?

3. ## Re: Locating the start of a curve

As far as im aware there system we are using starts recording when we press start and the there is a period before the data we need is collected which is the noise. each test the "start" point changes on each test

4. ## Re: Locating the start of a curve

That does not tell me how you want to detect start of activity, so I am going to guess that this is more of a signal processing question than an Excel question.

I am not an expert in signal processing, so I can really only comment on some basic ideas. I see a few different simple ways that programmers in signal processing will detect "signal started to do something:"
1) Signal rose above a certain threshold. In your case, if I said that I would declare "the first time the total of the two signals rises above 800, that's where I will declare 'start of activity'." That kind of rule would declare the row immediately below the row you have marked as the start of activity.
2) Signal changed by a certain threshold over its starting value. In your case, the total signal starts around 720, so I might say "when signal rises above 10% above starting value (792), that's where I will declare 'start of activity'." This would declare start of activity about 10 rows before the entry you have marked as start of activity.
3) Or maybe one might look at the "slope" (however you choose to define it for the signal at hand) of the signal, and declare "when the signal's slope changes above a threshold, that's where I will declare 'start of activity'." That might be as simple as adding a column to calculate "slope", then apply either test (1) or test (2) to the "slope signal".

From there, you can get into various combinations of those tests, and various levels of signal smoothing, and more complex logic to detect "start of activity". I would suggest that you step back from Excel and analyze the signals (because I don't know if it is worth the programming effort if this is the only signal you will ever need to analyze) in your own mind until you have at least a conceptual idea of how you want to detect "start of activity". Once you have a more concrete concept in your mind of how you want to detect "start of activity", then we can start working on how to program that algorithm into Excel.

5. ## Re: Locating the start of a curve

some food for throught, thank you for you help its put or a good path to get started

There are currently 1 users browsing this thread. (0 members and 1 guests)

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts

Search Engine Friendly URLs by vBSEO 3.6.0 RC 1