The Importance Of Gender In Communication

Abstract— The Speech is basic part of a human being to communicate with each other. This paper demonstrate the analyzation of gender by the speech and present the certain emotions impact on fundamental pitch range. It differs from gender to gender. In Future speech will be major part to communicate with machine. Now a days as a learner we try to communicate machine with soft voice. This paper take a challenge accepting voice sample with different mood or emotion and detect a specific gender depending on the frequency of the gender.

Keyword— Emotion; Gender; Pitch; Speech Processing.

1. Introduction The communication is nothing but exchanging the information between two or more people. The speech is basic part of communication, by speech one

2. Types of Speech For speech recognition it is majorly categorized in 4 basic category-

2.1 Isolated or Single Word

The speech recognition system which identifies a word with a complete wave, empty head and tail offset. We consider this category to detect gender [1] [3].

2.2 Connected Words

The connected word is similar to the single word, there is minimal pause between two or more sentence [1].

2.3 Continuous Words

The connected words with no pause between two words or in any sentence. There is no empty line in plotted voice pattern between two words

d) Determine each sample is by Human being not by any Machine, and also determine the noise does not affect the quality of speech.

e.) Every recorded speech sample is 1 to 2 sec. and the age range persons belongs to between 20 to 35 years.

4.2 Feature Extracting

Feature Extraction from voice record requires two techniques first Sampling and another is Quantization. We have obtained values from each voice sample The higher sampling frequency consider better quality like 44Khz is considered as good quality, that means 44000 sample for one second of speech. Quantization is nothing but a process where mapping large range’s continues number to range with smaller discrete value, before feature value extraction we need Sampling and Quantization.

Sampling Frequency or Sampling Rate that defines number of discrete values sampled in one second [11]:

Fs = 1/Ts

4.3 Plotting

We have used PRAAT application software for plotting. This software is free of cost and easily available on the net.

4.4 Extracting Values

Using PRAAT tools we have obtained Pitch, Intensity, Formants, Bandwidth, and Time Duration. Every sample speech’s property is isolated and plotted specifically for this