Ensure float operations in SoundSynthesizerEffects & CompassCalibrator.

[microbit-carlos]

Related PR, which includes files in codal-core like LevelDetectorSPL, and together with this PR it saves >400 bytes of flash:

• Ensure float literals and float versions of math functions are used. codal-core#176

Edit: Also ensured float literals in a couple of other files to avoid implicit double promotion.

[github-actions]

Note

This build diff was generated when this branch included for testing commit 5c04d7c , which updates codal-core in target-locked.json to the commit from PR lancaster-university/codal-core#176, and it shows dropping log10() completely for log10f(), which results in almost half a KB of flash savings.

This was reverted and the "real" build diff can be seen in #454 (comment).

Build diff

Base commit: 6ac8666c0100b19551ddb7c2acf007b8950fe9b3
Action run: https://github.com/lancaster-university/codal-microbit-v2/actions/runs/12017300166

    VM SIZE    
-------------- 
 +1.9%     +24    /home/runner/work/codal-microbit-v2/codal-microbit-v2/libraries/codal-core/source/streams/LevelDetectorSPL.cpp
 -2.1%    -488    [section .text]
 -0.1%    -464    TOTAL

[github-actions]

Build diff

Base commit: 6ac8666c0100b19551ddb7c2acf007b8950fe9b3
Action run: https://github.com/lancaster-university/codal-microbit-v2/actions/runs/12036161069

     VM SIZE    
 -------------- 
  [ = ]       0    TOTAL