Processor selection in the VoIP Solution

As VoIP continues to replace analog phones in the enterprise voice communication market, this technology is also rapidly adopted in residential environments and Small and Medium-Sized Enterprises (SMBs. The advantages of VoIP, such as multiple lines, the use of cheap digital media voice communication systems, and significantly reduce the communication rate, not only for large enterprises, but also for small customers are quite attractive. As the demand for VoIP devices, hosts, IPPBXs, gateways, and other devices continues to grow, the requirement for higher sound quality is clearly the top priority. market feedback has shown that, if the quality of VoIP is lower than that of typical POTS, mainstream customers will not turn to VoIP. The challenge for VoIP hardware developers is to select a platform to achieve an appropriate combination of system performance and total BOM costs, it provides a successful solution for a competitive market. The core processor in the VoIP system is a key component in the design, which greatly affects the quality, feature set, and material list of the speech processed. This article will examine the hardware requirements of VoIP and the compromise that must be considered when selecting an effective method.

VoIP Application Requirements for processors

The main problems that VoIP hardware developers need to face are the collection of features, voice quality, product costs, development plans, and design scalability. The ideal way is to develop a public hardware/software platform to meet the needs of multiple different products. However, each terminal device (IP Phone, IAD (integrated access device), home gateway, and IP-PBX) has different requirements, choosing a processor platform that meets all these requirements is a big challenge. Most terminal devices have a very short life cycle and will soon be replaced by new versions with new performance or different features. This trend requires a way to make them more flexible, it also helps with rapid design and ensures that the time to market for new products is minimized.

The following table describes the optional VoIP Hardware Work platforms. All these solutions weigh the material cost list, design flexibility, Development Plan, and time to market for products with known performance levels and synthetic speech quality.

Comparison of VoIP device processors based on key decision criteria

Different methods have their own advantages and disadvantages. The product and market priorities must be taken into consideration in the decisions made. The microcontroller-based platform provides a familiar and inexpensive method for embedded designers and a friendly development environment. It is also a typical low cost and can easily process IP packets. However, for voice encoding/decoding and echo elimination signal processing, the microcontroller is not a good choice, and its evolution to VoIP products is a limited platform. DSP architecture is better at data stream processing, which is easier for speech processing. For this reason, it is more reasonable to add a DSP to the design, which makes the processing task assigned to the processor more suitable for execution. But this will increase the material list, not just additional processors, but more memory and peripherals. Multiple chips (MCU plus DSP or MCU plus ASIC) will add more materials and more work.

Dedicated, fixed-function hardware is another method, which has a significant negative impact on design flexibility. If it is customized chip design, there is also NRE (one-time engineering cost) and time to market. If it is a commercially available ASSP (standard product for specific applications), the ability to design differentiated products will be significantly affected. Another good way is to build or purchase a SOC that integrates MCU and DSP on a single chip, which can reduce the number of product packages and cause some loss in flexibility, however, it provides the programmable capability. MCU and DSP programs can be applied to different target applications.

The last method discussed is to use a single "aggregation" processor to process control logic and voice engines. Aggregation processor is a single processor with signal processing performance in the independent DSPs field. Different MCU manufacturers have integrated some signal processing functions, such as instruction set extension and MAC unit on the MCU core. However, this method lacks the fundamental structure required for high-performance signal processing, to handle Advanced speech quality enhancement, such as high-performance echo eliminators and noise suppression.

ADI's Blackfin Processor has high performance in signal and control program execution and complies with "aggregation" processor standards. As the "aggregation" processing satisfies all the key success factors of the VoIP hardware platform and has obvious leading advantages, the rest of this article will discuss this method.

1. VoIP status quo, technical platform, and Development Trend
2. Application of Wireless LAN VoIP technology