I haven't found this mentioned anywhere in the documentation: is there a minimum length of sound that can be matched by ShazamKit with a custom catalog? Is it possible to have it match a single spoken word?
Is there a minimum length of a sound sample for matching?
I have discovered that the minimum length to create a signature is 1 second, but so far I haven't been able to get anything shorter than 2.5 seconds to be recognized by ShazamKit. Does this sound like an expected result?
Check out the documentation for SHCatalog... there's a property called minimumQuerySignatureDuration that might help...