I've taught myself to use the IPA as an aide for pronunciation training, but I'm well aware that other phonetic alphabets are in widespread use in language courses.
Where does the IPA fall short where other alphabets presumably excel? Why do some language courses use other alphabets?
The only phonetic alphabet I know of that is comparable to IPA in universality and language coverage is X-SAMPA. The advantage of X-SAMPA compared to IPA, is that it contains only ASCII characters and is therefore easy to type. The disvantage of X-SAMPA is that it's harder to read, and there are definitely much fewer people who read it fluently than those who can read IPA. Apart from that, there is no difference, as these two phonetic transcription systems can be automatically converted into each other.
Apart from IPA and X-SAMPA, we have a lot of systems that cover specific langauges or groups of languages. Their main advantage is usually the fact that they are adapted to these languages, and are therefore easier to read. Moreover, they are usually clearly phonemic - they express only contrasts that are perceivable by native speakers of a particular language. IPA, on the other hand, may express pronunciation on the phonemic level, but can also show non-phonemic details. This versatility is an advantage, but it may also be a problem: given only an IPA transcription without any other information, you can never know how crude or how detailed the transcription is.
My advice would be learn whatever phonetic alphabet is most common in your target language, in order to use pronunciation information from dictionaries, etc. Apart from that, learn IPA in order to catch similarities between different languages you know, but be aware that we cannot be sure if a phoneme expressed with a particular IPA symbol in one language is exactly the same as a phoneme expressed with the same symbol in another language. The usage of the same IPA symbol does indicate, however, that the sounds are at least similar. Finally, if you're planning to type IPA a lot, learning X-SAMPA will be definitely useful.
IPA is the best choice for pronunciation transcription. IPA has a long history of development (since 1886) and represents best scientific practice. This means the system is concise, unambiguous and accepted word wide.
Currently IPA is maintained by International Phonetic Association and described in book Handbook of the IPA.
Unfortunately quality of linguistic sources differ across publisher and teachers. You can see in typical dictionary:
little [ˈlɪtəl]
But it is hard for non-professional to read dialectic variants:
[ˈɫɪɾɫ] - General American
[ˈlɪʔo] - Cockney
[ˈɫɪːɫ] - Southern US English
Many important dictionary publishers switched to IPA but unfortunately they use wrong IPA transcription just to obey tradition.
For example modern English has [ɹ̠] (LATIN SMALL LETTER TURNED R) but that sign can scare readers and dictionary publishers often use [r] (LATIN SMALL LETTER R). Compare sound fragments from:
Another example is dark l (velarized alveolar lateral approximant). l in feel is not the same as in lamp and some English learners may have a hard time to distinguish them:
lamp [læmp] ⇔ feel [fiːɫ]
On the other hand I don't see in popular English dictionaries usage of ɫ sign (LATIN SMALL LETTER L WITH MIDDLE TILDE).
I see growing practice to use IPA for person or organization name pronunciation (on home pages, Wikipedia, etc) to avoid ambiguity.
For entering IPA into text I uses Emacs X-SAMPA input method and it works great (I build dictionary with it) but unfortunately I don't know any other convenient way to input Unocode IPA sympols.
UPDATE @michau
we have a lot of systems that cover specific langauges or groups of languages. Their main advantage is usually the fact that they are adapted to these languages, and are therefore easier to read. Moreover, they are usually clearly phonemic - they express only contrasts that are perceivable by native speakers of a particular language
It is very true. For example American Heritage Dictionary of the English Language is typically for American dictionaries extensively uses diacritic sign and this system looks like a garbage for me but widespread in USA ((
Examples of inconsistency for pronunciation transcription for English language can be explored here: https://en.wikipedia.org/wiki/Pronunciation_respelling_for_English
l in feel is not the same as in lamp" - as a native English speaker I'm pretty sure it is... but what do I know?
I have seen some other phonetic alphabets (e.g., the one used in the Merriam Webster Dictionary using things like /ᵫ/ instead of IPA /ʏ/, or the Uralic Phonetic Alphabet (UPA) used scientifically for Uralic languages), but IPA is the de facto standard.
It doesn't matter much which phonetic alphabet you use for learning just one language, but you appreciate to have a standard phonetic alphabet once you get at the second language to learn.
I think none can compete, none is as complete as the IPA to describe languages of the world. It has been adapted, revised, so it's also certainly the most accurate.
It's the official scientific "unit" in linguistics and phonology, like the meter/second is the international scientific unit for the speed in physics...
But: if you study only one language, and don't compare several languages, you may find another one more adapted to the language you study.
So, it depends on your goal. If you want to be understood by other people (asking if this word is pronounced this or that in their dialect) or to study comparative linguistics, IPA is the best choice (all books use it).
If you are, like I am, a wannabe polyglot, who study several languages, the best choice is to first spend a lot of time studying the IPA, because it's very well documented, and you will understand all the linguistics and phonological concepts to create your own afterwards.
So, the best is to learn the IPA, then, either:
1 create your own one afterwards. Adapting it to your needs, and you will learn a lot about the phonology of the language you study doing that.
2 use a phonetical conscript if you find one. You can adapt it, as it's not certain that this script would be as scientifically accurate than the (so much verified) IPA is.
You can try existing conscripts:
Conscripts are the equivalent of Conlangs for scripts, they are adapted or invented scripts to write a language, they are not all phonetical, but some of them are, and are often adapted to a few languages only, or only one, and you can find inspiration if you want to create your own system:
-https://www.omniglot.com/conscripts/ -https://www.omniglot.com/conscripts/natlangs.htm
They also have resources to help you create your new font for your invented scripts:
https://www.omniglot.com/links/fonts.htm#software
Phonetical conscripts:
https://www.omniglot.com/conscripts/phonetic.htm
For instance, this one is called "UA" (for universal alphabet), and is phonetical:
https://www.omniglot.com/conscripts/ua.htm
This one has been specially created to write French:
https://www.omniglot.com/conscripts/franabugida.htm
I think today the IPA alphabet is used only in cases where you have a very poor material-base for learning a language (maybe, minority language). I'm convinced that in case of most 30 widely learned world's languages, students don't need to know IPA (they can easily find (hear) pronunciation on YouTube and other internet sites).
But we should keep in mind that in the academic world IPA is very important in case of extinct languages. There are lots of them we have no books about, but during those times scientists used to write pronunciation in IPA.
