Have you heard about speech compression? It is a method used in many communication techniques to transfer audio data. The compression of speech signals has several practical applications. An example is the digital cellular technology where several users share identical frequency bandwidth. Compression permits more users to share the system than otherwise possible. Another such example is in digital voice storage (e.g. answering machines). For a given memory size, compression permits longer messages to be stored than otherwise.
Speech compression involves the compression of audio information in the kind of speech. It is a somewhat distinctive kind of audio information, with a number of desires that should be addressed during compression to confirm that it’ll be intelligible and fairly pleasant to concentrate to. a number of software programs are designed specifically with speech compression in mind, including programs which may perform additional functions like encrypting the compressed information for security.
Speech may be a comparatively simple and widely studied kind of audio information that makes it simple to compress in some ways. However, it’s necessary to make sure that compression retains the integrity of the speech. If the information becomes distorted in some way, it can be tough to know, and it can also be hard to concentrate to.
Thus, speech compression must be performed in a manner that retains the key qualities of the information. It’s easy for speech to song “wrong” to a listener, if interfering with understanding of the transmitted information.
Speech coding differs from different types of audio coding in that speech may be a much easier signal than most other audio signals, and a lot more statistical data is obtainable regarding the properties of speech. As a result, some auditory data that is relevant in audio coding may be unnecessary in the speech coding context. In speech coding, the most vital criterion is preservation of intelligibility and “pleasantness” of speech; with a constrained quantity of transmitted information.
To obtain more compression specific properties of the speech signal should be modelling. The main assumption is understood because the source filter model of speech production. This assumes that a source is passed through a filter to provide the speech. The only implementation of this is often referred to as a LPC synthesizer. At each frame the speech is analyzed to calculate the filter coefficients, a voicing decision, the energy of the excitation, and a pitch value if voiced.
At the decoder a regular set of pulses for voiced speech or noise for unvoiced speech is passed through the linear filter and increased by the gain to provide the speech. this is often a very economical system and usually produces speech coded at 1200bps-2400bps. With clever acoustic vector prediction this could be reduced to 300bps-600bps. The disadvantages are a loss of naturalness over most of the speech and occasionally a loss of intelligibility.0