Added Serbian

[mstepanovic]

Hi,

I've added support for the Serbian language.

I've also taken the liberty to add another optional field in the mappings array: articulatory description, e.g.

{
    "phone": "ŋ",
    "phoneme": "n",
    "environment": "before velar stops [k, g]",
    "articulatory description": "voiced velar nasal"
}

I believe this could help differentiate phones whose broad IPA transcription is not specific enough. For example, phones such as [t, d, n] can be realized/described as dental, alveolar, or post-alveolar depending on the language or dialect, but the IPA alphabet often doesn't distinguish between these three places of articulation.

If you think articulatory properties are useful to include, I'd be willing to add them for all the languages and phones in the database.

Best,
Marija