The value of LEVEL_DETECTOR_SPL_8BIT_000_POINT (32) will never be reached

[microbit-carlos]

Because the LevelDetectorSPL minVal is set to 52:

codal-microbit-v2/source/MicroBitAudio.cpp

Line 82 in bd446c4

levelSPL = new LevelDetectorSPL(*rawSplitter->createChannel(), 85.0, 65.0, 16.0, 52.0, DEVICE_ID_SYSTEM_LEVEL_DETECTOR, false);

And LevelDetectorSPL will clamp the lower end of the level value to minVal:
https://github.com/lancaster-university/codal-core/blob/509086cc8590465041b15493ab52b56e7071c110/source/streams/LevelDetectorSPL.cpp#L142-L147

[microbit-carlos]

PR created in codal-core:

• Set LEVEL_DETECTOR_SPL_8BIT_000_POINT to 52.0 dB. codal-core#174

MakeCode and MicroPython already use this value, so it shouldn't have an impact there

• https://github.com/microsoft/pxt-microbit/blob/v7.1.3/libs/microphone/microphone.cpp#L8
• https://github.com/microbit-foundation/micropython-microbit-v2/blob/v2.1.2/src/codal_app/codal.json#L16

[microbit-carlos]

Previous discussion:

• New audio.levelSPL returns fairly high values in a very quiet room. #236

When LEVEL_DETECTOR_SPL_8BIT_000_POINT was introduced it was set to 35.0, but the default value for LevelDetectorSPL.minVal was already set to 52.0, so with the default values LevelDetectorSPL.getLevel() would have never return a value below 52 dB:

• lancaster-university/codal-core@d376f1a
• https://github.com/lancaster-university/codal-core/blob/d376f1a851457effaf34058c62bd6c42c5ada6a0/inc/streams/LevelDetectorSPL.h#L97-L100

However, at that time, the uBit.audio.levelSPL constructor in codal-microbit-v2 was setting minVal to 0, it was only changed to 52.0 a year later, so until then uBit.audio.levelSPL->getValue() could return values lower than 52.0, and that explains why earlier measurements could detect 35-40ish dB:
02a85fa