How to identify the English \t\ consonant in sound recordings?

Question

I am trying to recognize the consonants in a word i.e. if the word spoken is tap, using different tools like Praat or MATLAB etc. I want to implement a system to ensure that the first letter spoken was \t\. Someone please guide me about any pattern related to \t\ consonant which may use the intensity or formants etc. Once I get the pattern of the consonant, I can then assess the quality of the consonant \t\ i.e. whether its correctly spoken or something else was spoken instead of \t\.

What I Know:

1. Its impossible to recognize a consonant alone as there's no information in a consonant and it can only give information with the help of an adjacent vowel.
2. In all the stop consonants, the difference between the formants f2 and f3 is decreased in the stop and then increased when the stop is released.

Answer 1

You could use Acoustical characteristics of selected English consonants, by Ilse Lehiste, to determine what the formant transition patterns are, or the voicing patterns, that would be of interest to you. However, a simpler way to decide is to just listen. If for example a listener isn't fluent in English and has problems in certain contexts, distinguishing "tea, key, pea" then this might be useful. However, nothing much will help you if your underlying question is "is this [t] or is it [t̪]?". A study acoustical properties of sounds in a language will mainly tell you what distinguishes sound a and sound b within the language, but there is no resource for "absolutely identifying" what sound exists in a token, compared to the vast array of sound in all other languages. Absolute identification is impossible, but if you are trying to find a closest match assuming English, then Lehiste's book should help.