TM4C1294NCPDT: TI 4-wire resistive touch driver source code

If I may - as one w/(some/slight) Display Experience - the following is offered.    It is possible that you seek (only) 'Coordinate Calculations' - I will answer that - yet a 'bit more 'Touch Screen' detail' may prove valuable to others.     

I've created a 'new thread' - more in keeping w/the 'Basics of Touch Screens' - as vendor has just announced that such effort was '6 years past.'

https://e2e.ti.com/support/microcontrollers/other/f/908/t/838837

Now our firm's method of 'Screen Calibration' involves, 'Systematic Touches at each Screen Edge and Screen Center - while one of the Screen Planes is powered.    (i.e. one plane is powered - the other is then read - the pressure of the 'Touch' proves sufficient to impress voltage upon the 'undriven screen.'    (i.e. the orthogonal planes come into 'contact' at the 'touch site!')

In most cases - my firm/others designate the 'Left, Bottom' portion of the Screen as the (near) '0,0' Location.    Thus the Right, Top assumes 'max value' - and you are free to assign value to that.    Note that the 'normal' touch screens may undergo (some) change in resistance with: 'lot, aging, use & temperature.'    Under 'high accuracy demands' - it proves wise to provide regular 'Touch Screen Calibration' - in which the user will 'touch key targets' - which you generate & place 'on-screen' - to receive their (hopefully) overlapped touch.    Your software can then make the necessary modifications (as/if required) to re-establish 'Accuracy of Touch.'

You may have been concerned due to vendor's use of 'FLIP' - imposed upon Landscape & Portrait.    This provides the ability to 'accommodate' a 'reorientation' of the 'Display' - in case it must be rotated - for any reason.   (usually to enable a 'better' fit - or mistaken measurement .)

Now to your 'dedicated question' - those calibration parameters appear in 'Groups of seven' - in my experience '5 such parameters' usually suffice.    (4 identify each display corner, 1 marks display center)      I believe that the 'enormity' of (many) of those (previous) vendor parameters result from the 'scaling up' of both the 'most significant 8 ADC bits' & the (earlier) 'display's pixel resolution.'      The goal (should) be to eliminate the need for 'floats' - while 'Speeding & Easing' the 'ADC Voltage to Screen Pixel Location' Calculations.     (i.e. makes the math easier for (both) MCU & programmer.)    

It is further suspected that the (unusual), 'Use of  Negative Numbered Parameters' is due to 'Designating the exact center of the display screen' as the '0,0' Location.     (both horizontally & vertically)     In this manner - Touch Location starts from and 'radiates out' from Screen Center.   (assume Left & Down - referenced to screen center - yield  those negative coordinates.)      This methodology seems advantaged - as it may improve Touch Accuracy AND yield Faster Touch Calculations.     (as most touch targets are likely to be nearer to screen center than screen edge!)  Note that this is (all)  'cb1 theory' - which may - or may not - correctly describe/represent vendor's implementation...

With the vastly increased pixel field (which your 480x320 display provides) it may be that your 'Touch Screen's Resolution' - becomes challenged.    It is normal/customary to determine the 'reasonable screen area' over which a 'touch extends' - and then to use that measure to determine 'how many & at what separation' (legitimate) 'Touch Targets' may be placed.    You may then 'slide' your  touch targets - either left/right or up/down (or both) - to more precisely overlap your display's 'touch directing pixels.'    (these announce/present the 'screen's target area for the touch'')    

To my mind - the attempt to 'duplicate' the past parameter calculation - may prove w/out substantial merit.   It is believed that the 'general methods listed here' - will most usually - enable you to  'generate & detect multiple & efficiently sized/placed 'touch targets.'     Relating the resulting ADC voltage produced - to each target's touch location - should not prove (that) difficult - and may resolve via a (near) linear formula.