Friday, June 22, 2018
Publication Date: 10/1/2008
Archive >  October 2008 Issue >  Tech Watch > 

Successful Mobile VoIP on iPhones and Smartphones

Providing VoIP (Voice over IP) on a mobile device is a technical challenge, as small devices have limited processing capabilities, and can be fairly expensive with short battery lives. Therefore, VoIP processing software must be efficient in terms of memory usage and computational complexity. Some optimization to the processor is usually needed and most devices require fixed point implementation of the signal processing algorithms.

These tasks are demanding and require both programming skills, and signal processing expertise. In addition, the quality of the network connection presents a unique challenge to application development. Mobile devices, by definition, operate over wireless links. For a VoIP call, the typical wireless access methods are Wi-Fi, 3G or Edge packet data connections. Many factors can cause disturbances in the transmission of these wireless links. For instance, low signal strength when the device is far away from base stations or access points can lead to a low signal-to-noise ratio as the background noise level becomes significant.

Interference from Other Users
Interference from other users communicating on the same or neighboring frequencies is another major factor degrading quality, which can lead to bit errors that trigger discarding of packets. Plus, for any real-time communication to be effective, delay must be minimized, eliminating the viability of retransmission, thus leading to packet loss. The receiver software needs to mask the error in the output signal through packet loss concealment (PLC) techniques. For mobile solutions, the PLC methods must be able to handle peak packet loss rates up to 30 percent.

Also, bottlenecks exist in mobile VoIP. Too many users sharing the same frequency band or access point — especially for Wi-Fi — leads to variations in throughput and transmission times. The resulting phenomenon is known as jitter and is evident when the receiver experiences irregular intervals between packet arrivals. The best way to combat jitter is to add a buffer at the receiver side to smooth out the delay variation. The amount of buffering should be adaptive and roughly follow the envelope of the jitter curve. Solutions should minimize delay while accounting for jitter, yielding the clearest and most consistent conversations.

The launch of the iPhone brought with it ground-breaking design and multimedia capabilities to the mobile world. The phone's open APIs and flexible design mean that any developer familiar with the Mac OS can now enable high-quality voice calls, with the possibility of adding real-time video in the future. The IP network access consists of both Wi-Fi and 3G/Edge channels, thereby allowing VoIP calling in virtually any location.

Devices such as the iPhone present additional complications for designing VoIP applications. For instance, smartphones introduce extra jitter when they need to simultaneously perform several tasks for the end user. For example, the iPhone may try to check for e-mails during a call, leading to delays in a VoIP call. The effect is the same as network jitter, further emphasizing the need for a solid jitter buffer.

Hands-Free Operation
Also, hands-free operation is a typical usage scenario. The further a user moves from the microphone, and the more amplified speaker volume, the greater the acoustic echo becomes. This heightens the already lofty requirement for efficient Acoustic Echo Cancellation (AEC). AEC is a tough problem on most platforms and it's tougher on mobile devices with limited processing power.

These unique requirements demonstrate why having high-quality voice and video processing capabilities is important. Since iPhone application developers are typically not audio or video experts, it is Global IP Solutions' (GIPS) goal to deliver real-time IP multimedia communication to the iPhone developer community. GIPS has been developing voice processing solutions since 1999, and has extensive experience with mobile VoIP, particularly in the areas of voice-quality, packet loss, delay and jitter. GIPS developed such innovative solutions as the iLBC codec (an IETF standard) and the de facto standard wideband codec — iSAC, and its NetEQ jitter buffer. In 2002, the company launched the first version of its voice processing package called VoiceEngine for the PC and Windows, and provided a complete package of VoIP sound processing. Soon thereafter, VoiceEngine Mobile for Windows CE on the iPaq was developed, providing breakthrough high-quality VoIP over Wi-Fi. VoiceEngine has since been the most widely deployed voice engine globally, enabling applications such as Skype, Yahoo Messenger and AIM with voice capabilities.

VoiceEngine is platform-independent and has been available for Mac OS for many years. The long mobile history of VoiceEngine, as well as GIPS' Mac experience made porting VoiceEngine to the iPhone a natural next step. GIPS also offers a VideoEngine Mobile product, and as the iPhone APIs become more open, developers will soon be able to create real-time voice and video solutions for the iPhone. The platform-independent code base, combined with a library of optimized code components for the ARM processor used in the iPhone make it possible to quickly release an efficient high-performance software package without having to worry about CPU constraints. In addition, VoiceEngine contains a highly adaptive jitter buffer called NetEQ, which manages the network impairments and additional jitter introduced on smartphones. It adapts very quickly to handle the jitter and minimize buffering delay.

Finally, VoiceEngine utilizes advanced AEC and noise suppression solutions to minimize the effects of echo and background noise. All of these capabilities were optimized specifically for the iPhone, making VoiceEngine the ideal solution for developers of mobile VoIP applications. VoiceEngine Mobile for iPhone also provides a rich API that enables developers to integrate very high-quality VoIP into their applications, while delivering the flexibility to adjust to a variety of parameters, such as echo cancellation.

As the market for real-time voice and video communications on mobile devices grows, developers and manufacturers need to be aware of the unique demands presented by mobile devices. Factors such as jitter, delay, acoustic-echo and CPU limitations must all be taken into consideration in order to produce a high-quality mobile voice and video experience for consumers and the enterprise.

Contact: Global IP Solutions, Inc., 642 Harrison Street, Second Floor, San Francisco, CA 94107 415-397-2555 fax: 415-397-2577 E-mail: Web:

search login