map() issues with a temporary solution

Using Arduino Programming Questions
#1

While prototyping a small LED project, I needed to map a potentiometer and found I was having bad flickering issues when I reached the higher end. Running the input/output through Serial, I saw that the distribution in map is horribly uneven. Hunting for solutions, I noticed that many folks have the same issue, so I figured I would post my temporary solution in the hopes it could be improved upon by the community.

NOTE: The flickering was not from bouncing, which I had already compensated for, but the uneven distribution in map() when using smaller numbers (e. g. 3)

unsigned int myMap(int value, int inmax, int outmax)
{

  // divide input max by output max
  int division = inmax / outmax;

  // initialize count
  int count = 1; // note: this method will not work directly if you need a 0 in the map. you will need to -1 the returned value if so.

  // while the value of the sensor is greater than the desired mapping
  while ( value > division * count)
  {
    // add 1 to the count
    count++;
    // if the count is greater than the maximum desired mapping, return the output max
    if (count >= outmax)
    {
      return outmax;
    }
  }
  // return the final count 
  return count;
}

usage:

ledPos = myMap(potVauel, 1023, NUMPIXELS) - 1; // map the led position

Any thoughts on improving this or effectively using map() better are very welcome.

#2

More people might be able to help you if you post your full code.

map() is an integer function so your issues may be coming from truncation if you're trying to divide small numbers by large ones.

#3

I cannot see your code or how you used the map(...) function so I am left to guess that the problems that you had stem from the integer nature of the map(...) function.

You may have better luck implementing a floating point version of the map(...) function. The reference documentation for the map(...) function shows how it is implemented.

long map(long x, long in_min, long in_max, long out_min, long out_max)
{
  return (x - in_min) * (out_max - out_min) / (in_max - in_min) + out_min;
}

You may have better results by implementing a rounded floating point version of map(...). Of course, this will increase the memory usage due to the need for floating point library routines.

Not tested:

long mapRoundedFloat(long x, long in_min, long in_max, long out_min, long out_max)
{
  return long(float((x - in_min) * (out_max - out_min)) / float(in_max - in_min) + out_min + 0.5) ;
}

I believe that floating point arithmetic will be used due to the arithmetic rules of the C language.

#4

Running the input/output through Serial, I saw that the distribution in map is horribly uneven.

It is not. You have (shown) not a shred of proof for that statement.

#5