The utilitiy Float2Qpt allow the user to convert floating point numbers to either signed or unsigned fixed point numbers in a Q(QI.QF) format with automatically selected or user definable word lengths, and automatically selected or user definable error tolerances (epsilon). Alternatively a user can pass in optional parameters to force the type of rounding used in the conversion, enable/disable verbose reporting, and even open the reference paper the untilty is based on. The Utility FxdPt2Float converts a fixed point number back to a floating point number given a fixed point number in a Q(QI.QF) format based on the fixed point number and the number of fractional bits QF.
These utilities are useful for development of fixed point versions of algorithms & the referenced white paper provides example of typical multiply/accumulate/saturate types of operations done in fixed point.

function[FxdPt, QI, QF] = Float2Qpt(num,varargin)
% Calculate the Fixed Point Approximation of a Floating point number "num"
% in Q Fixed Point Format - Q(QI.QF)
% FxdPt - is the fixed point coverted value
% QI - is the number of iteger bits required for range of num
% QF - is the number of fractional bits required for a specified epsilon or
% the number of fractional bits implemented when using a shortest standard
% word length and living whith the resolution you get for the required
% range
%
% OptionalInputs
% 'signed' - Create a signed fixed point number *Default
% 'unsigned' - Create an unsigned fixed poin number
% 'ShortestSWL' - Creates the shortest Standard Word Length
% 'WordLength' 8,16,32,64,128,256
% 'epsilon',number - specifies an epsilon error tolerance used to set the
% number of fractional bits QF
% 'QF',number - The user can specify a "number' of fractional bits and
% covert afloating point "num" to fixed point directly
% without use of other parameters other than rouding options
% 'round_closest' - rounds to the closest integer when computing FxdPt
% 'round_zero' - rounds the towards zero when computing FxdPt
% *** round towards zero is default behavior
% 'verbose' - displays conversion process diagnostics based on user inputs
% in addition to coversion range & precision information
% VERY useful to turn on - off by default
% 'Paper' - opens URL to a PDF copy of the paper below:
% "Fixed-Point Representation & Fractional Math Revision 1.2"
% August 30 2007
% DOI: 10.13140/RG.2.1.3602.8242
% Link1 http://www.superkits.net/whitepapers/Fixed%20Point%20Representation%20%26%20Fractional%20Math.pdf
% Link2 https://www.researchgate.net/publication/235791711_Fixed-Point_Representation_Fractional_Math_Revison_12
%
% Written By Erick L. Oberstar 12/15/2015
% U.W. Madison Mechatronics Lab & Intellivention LLC
% oberstar@engr.wisc.edu & oberstare@intellivention.com
%**************************************************************************

function [ FloatingPtNum ] = FxdPt2Float( FxdPt, QF, varargin )
% Calculate the Floating point value of a fixed point number
% in Q Fixed Point Format - Q(QI.QF)
% FxdPt - is the fixed point coverted (integer) value
% QI - is the number of iteger bits used for range of FxdPt (not required)
% QF - is the number of fractional bits used that sets the epsilon
%
% OptionalInputs
% 'verbose' - displays conversion process diagnostics based on user inputs
% in addition to coversion range & precision information
% VERY useful to turn on - off by default
% 'Paper' - opens URL to a PDF copy of the paper below:
% "Fixed-Point Representation & Fractional Math Revision 1.2"
% August 30 2007
% DOI: 10.13140/RG.2.1.3602.8242
% Link1 http://www.superkits.net/whitepapers/Fixed%20Point%20Representation%20%26%20Fractional%20Math.pdf
% Link2 https://www.researchgate.net/publication/235791711_Fixed-Point_Representation_Fractional_Math_Revison_12
%
% Written By Erick L. Oberstar 12/15/2015
% U.W. Madison Mechatronics Lab & Intellivention LLC
% oberstar@engr.wisc.edu & oberstare@intellivention.com
%**************************************************************************