Main siteMain site  ForumForum  ForumSearch  Private messageEmail contact  RegisterRegister  Log inLog in 
Topic: Standard Deviation
Reply to topic
Author Message
EMebane



Joined: Apr 15, 2019
Posts: 67

PostPosted: Jun 26, 2019 6:19 PM    Post subject: Standard Deviation

I'd like to take the standard deviation of 4 scalars.

My thought was to feed them into a Vector and then apply a VectorToScalar with Standard Deviation. There are lots of RMS-related options, but no Standard Deviation among the options.

What's a good way to take the Standard Deviation of 4 scalars without a Standard Deviation option in VectorToScalar?
jarek



Joined: Oct 22, 2007
Posts: 721

PostPosted: Jun 26, 2019 6:45 PM    Post subject:

There is STANDARD DEVIATION option in VectorToScalar, but only in the Float version (or DoubleFloat).

The Scalar elements is how BioEra started, but they have limitatation (and sometimes are problematic), so recent development is focused more on Float elements.
EMebane



Joined: Apr 15, 2019
Posts: 67

PostPosted: Jun 26, 2019 7:07 PM    Post subject:

Is your recommendation to convert to float to do the calculation (then back to scalar if necessary)?

What's the name of the element? I can't find it.
jarek



Joined: Oct 22, 2007
Posts: 721

PostPosted: Jun 26, 2019 7:22 PM    Post subject:

My recommendation is to use primarily Float elements, and Scalars only when operating on integer (non scaled) values.

Look for "ToScalar" element and select FLOATVECTOR. This is fully Float element, even though is has "scalar" in its name. To determine the type of an element, look at the input/output pins (hover mouse over an input/output pin and it will show its type).
EMebane



Joined: Apr 15, 2019
Posts: 67

PostPosted: Jun 26, 2019 7:41 PM    Post subject:

OK, I see that element. I have been using "VectorToScalar" and didn't know about the ToScalar element.

The aversion to Scalars is important and news to me. I recall reading that you advise favoring Scalars due to faster processing so I've been trying to use them.
Was the Scalar favoritism old advice prior to the shift towards Floats?
Are the processing speeds of Scalar v. Float similar now?
Are there other times when Scalar is favored (other than non-scaled ints)?
jarek



Joined: Oct 22, 2007
Posts: 721

PostPosted: Jun 26, 2019 8:18 PM    Post subject:

I have been recommending Float element since their inception. I don't remember recommending scalars, where is this? Scalars could be faster on older devices (without coprocessor), but I don't believe that is the case on a typical Windows computer.

The "ToScalar" is a category. The elements are still VectorToScalar or F_VectorToScalar.
Reply to topic