Information Technology Reference
In-Depth Information
In phonological analysis it is worth noting that in the whole process of creating
phonological oppositions some exceptions can be encountered, namely phones
with a complementary distribution which makes the difference between them
acoustic and articulatory. In addition, these phones do not occur in the vicinity of
one another which means that they are opposed to other phones, but not to one an-
other, so such phones are classified as phones belonging to one phoneme, because
their common features do not characterise any other phone.
A matter of utmost importance in phonological analysis and semantic categori-
sation is the increasingly frequent problem of speech analysis and the analysis of
the speaker, i.e. the person pronouncing specific words. This is why the next chap-
ter presents the problems of speech and speaker recognition.
2.4.1 Using Semantic Categorisation to Analyse Speech and the
Speaker
As an element of semantic categorisation, the structural approach hinges on the
analysis of the input data structure. It has already been said that this chapter will
focus on phonological analysis, and thus speech analysis. The entire process of the
discussed data analysis ends in a stage referred to as recognition, the stage most
important from the point of view of the analysis, because at this stage the question
is to be answered whether we have been able to recognise the analysed text at all
(spoken or written) and whether it is logical from the perspective of the analysing
person.
Two primary subjects of recognition are distinguished in the process of text
recognition by its semantic categorisation: speech recognition and speaker recog-
nition.
The first type of recognition, speech recognition, starts at the speech processing
stage which is to end in extracting significant features and their meanings used in
subsequent analysis. In this case, the analysed speech is equated with the contents
of the text, independent of the person saying the specific words. For this reason,
during the analysis, the speech signal is identified together with all its elements
that can in any way determine the contents of the text.
Speech analysis is conducted very frequently due to rather significant benefits
of speech, which we frequently seem to forget, but which play important roles.
These benefits definitely include the natural way of transmitting the message, in-
cluding requests, instructions, requests, prohibitions and spontaneous content.
Communication itself is a very important aspect and element of speech, as it is
much faster than any written form of text. In addition, it contains some expres-
siveness, because when we want to say a certain sentence, we often add our own
emotions to it and this is sufficient, whereas in a written sentence we need to add
the appropriate number of epithets to at least partly convey the emotions that came
across with the verbal form. The spoken form frequently also dominates in social,
