Speech coding is essential for the functionality of VoIP, PSTN, video conferencing systems and digital cellular communication technologies. The purpose of speech coding is to demonstrate expressions in a digital way by producing standardized and intelligible speech applications. The advantage of waveform coding is that it minimizes bit rate and optimizes bandwidth.
Waveform coding is one of the technologies applied to support VoIP telecommunication services. The technology converts analog waveforms into digital forms. Waveform coding can change original signals without any knowledge of how the signal was processed. One of the waveform examples is Pulse Code Modulation (PCM). The PCM is a digital scheme for conveying information. It has binary signals that are demonstrated by logic 1 (high) and logic 0 (low). PCM is beneficial in these modern days because it is possible to digitize all forms of analog data. PCM can digitize full-motion video, music, voices, virtual reality, and telemetry.
The PCM technique involves a three step process:
- The analog signals are sampled by PCM at the rate of 8000 times per second. The sampling is utilized to produce a pulse amplitude modulation wave or PAM.
- Each PAM sample produced can be converted into a digital value. It should be noted that PCM does not use one-to-one assignment values. This is because one to one assignments can cause high bandwidth signals that are unnecessary.
- PAM signals are coded using compression techniques into a particular digital value. The process takes into account the necessity to reduce bandwidth in order to maintain effective techniques. The process produces digitalized speech at the rate of 64Kbps transmission.
Another type of waveform coding technique used by VoIP technology is Adaptive Differential Pulse Code Modulation (ADPCM), Standardized in ITU-T’s. 726. ADPCM is similar to PCM, and it is the most common speech transmission technique used nowadays. ADPCM uses binary signals that are represented by logic O’s and 1’s. The ADPCM can take numerous samples of the voice and demonstrates the value of the sampled voice modulation in binary terms. The technique is used to transmit sound on fiber-optic long –parameter lines as well as to keep music together with the text images and codes on a CD-ROM.
ADPCM does not convey speech signal directly, but instead, it quantifies the differences between predictions of speech forms. It is possible to quantify the speech signal and prediction of the speech because the next samples are similar to each other.
ADPCM can predict the value of the current sample by using the value of the adjacent sample. The variation value between the current sample and adjacent sample is then encoded. ADPCM also transcribes the number of bits utilized to encode a sample based on the range of amplitude taking place in the analog signal over time. The process requires fewer bits to encode the signal. The signal is produced in lower transmission rates ranging from 16 Kbps to 40Kbps. This is a little bit less than in the PCM technique process.
ADPCM has APC devices that can analyze noise. The APC was added to the shape noise spectral. The formation of noise spectrum allows the sound power to be transmitted across the frequency band. The sound is then masked by the input speech spectrum. The APC device shapes sound spectral filters and carries them to the perceptual weighting filters through the process of analysis–by–synthesis. The analysis–by-synthesis process also facilitates the formant filter shaping in post filters. ADPCM and PCM are important technologies to be used in the field of VoIP telecommunication.