|
Techniques to Reducing Payload Size
|
bet | 8/32 | Sana | 22.07.2021 | Hajmi | 0,89 Mb. | | #15652 |
4.2 Techniques to Reducing Payload Size
Network congestion can be a major problem in data networks such as the internet in which this application may use to transmit its data across. As a data network becomes congested routers begin to drop packets to reduce there congestion. Normally UDP packets are dropped first since unlike TCP, UDP does not detect the congestion and automatically reduce the rate in which it transmits data across the network So there needs to be a mechanism integrated into the application to reduce the data size sent across the network, such that, the network becomes less congested to reduce the probability of packets being lost due to them being dropped by congested network routers.
Compression is such a method which can be used to reduce the data size which will be transmitted across the network; compression basically removes any redundant information within the data to reduce its size. There are a number of different concerns which have to be addressed when choosing a compression codec; a major concern is whether the processor has enough power to perform the compression in real time, normally there has to be trade offs between delay incurred by compressing the data, the quality of the resulting packet and the size of the packet required, to find a balance suitable to the application.
There are two main codec’s which can be considered for voice compression, theirs GSM and G.723. The GSM lossy speech codec namely RPE-LTP (Regular Pulse Excitation Long-Term Prediction), is a codec which was specially design by the mobile network operators for compressing the speech which is sent across their networks. The audio is split into blocks of 20ms before being passed though the speech codec in which results a block of 260 bits which gives a rate of about 13 kbps almost 5 times smaller that the original PCM 64kbps. GSM provides 2:1, real-time compression as long as the system hardware is fast enough which makes it an ideal choice for a VoIP audio compressor.
Alternatively to GSM is the G.723.1 standard which specifies the format and algorithm for a constant bit rate audio codec designed for sending or receiving voice over low bandwidth network connections. The standard provides for two modes of operation a high quality mode and a low quality mode. High quality mode produces a stream at 6.3 kbps whereas low quality produces a stream at 5.3 kbps. G.723.1 works with a fixed block of audio of 30ms and produces blocks of about 160 bits and 190 bits depending upon mode. G.723.1 is used mainly for producing .asf files in Microsoft Windows and within Video Conferencing application as the choice of audio codec.
There are no real places to find a direct comparison in terms of encoding and decoding speeds between GSM ad G.723.1, but the quality of GSM is slightly better then that provided by G.723.1. With the application being developed for Windows CE an operating system with little free developer support it would be more practical to go with the better known codec of GSM with it being more probably to find an implementation of the codec available for the Windows CE architecture. GSM also provides better quality than G.723.1 at a size of just less than five times the original which is idea for the VoIPv6 application.
|
| |