Browse Prior Art Database

User Controlled Voice Quality and Billing

IP.com Disclosure Number: IPCOM000022659D
Original Publication Date: 2004-Mar-24
Included in the Prior Art Database: 2004-Mar-24
Document File: 3 page(s) / 163K

Publishing Venue

Motorola

Related People

Jim Womack: AUTHOR [+2]

Abstract

The levels of quality of service (QoS) in communication networks tend to be based on application. For instance, the voice application is usually given preferential treatment over data applications, though no call is differentiated from the others in the network. This generates few complaints from users as the high-level of service normally assigned to voice provides sufficient quality. However, this comes at a cost not negotiable by the end users. If users and operators were allowed to have more control over the quality of voice that they receive, they would most certainly opt for less (and more) expensive quality under certain circumstances and charge accordingly.

This text was extracted from a Microsoft Word document.
This is the abbreviated version, containing approximately 33% of the total text.

User Controlled Voice Quality and Billing

By: Jim Womack and Jheroen Dorenbosch

1         The problem

1.1        Problem overview

The levels of quality of service (QoS) in communication networks tend to be based on application. For instance, the voice application is usually given preferential treatment over data applications, though no call is differentiated from the others in the network. This generates few complaints from users as the high-level of service normally assigned to voice provides sufficient quality. However, this comes at a cost not negotiable by the end users. If users and operators were allowed to have more control over the quality of voice that they receive, they would most certainly opt for less (and more) expensive quality under certain circumstances and charge accordingly.

1.2        The enablers

There has been considerable work done in creating variable rate vocoders. One of the most notable is the adaptive multi-rate wide-band (AMR-WB) vocoder that is has been designed for 3-G GSM systems [1]. There has also been considerable standardization on the automated means for measuring voice quality by end devices with some trying to measure the mean-opinion scores (MOS) in real time. Some examples of these techniques are the perceptual evaluation of speech quality (PESQ) and perceptual speech quality measurement (PSQM) [2]. Additionally, TIPHON has standardized QoS measurement methodologies [3] and the ITU has created an “E-Model” for use in transmission planning [4].

Hence, there are considerable means from which we can draw a solution. However, no one is currently, to the best of our knowledge, utilizing these techniques for anything other than passive measurements of the network performance.

1.3        The solutions

Given there are vocoders with adaptable rates and means for measuring the voice quality at the endpoints, we propose to use them in an active manner in which the user or operator can set variable targets for QoS. Common voice quality degradations are caused by network impairments such as dropped packets and excessive jitter. These degradations are typically time varying in nature. The idea is to set the target MOS[1] score, that in most cases will not be changed very often, and let the vocoder adjust its data rate such that voice quality is maintained at or near the target level. Below is the outline of three features that can be implemented with this.

1.3.1        User selectable MOS rates

We begin by assuming that there is a suitable interface by which users can select their desired level of voice quality. An example of an inconvenient method would be for the user to call a customer service support line and allow the operator to set the target levels. A more pleasing solution might be to provide the customer with a user interface and negotiated pricing scheme on their device. Alternatively a web site can be used. Another detail would be how to set default levels. It could be an operator default, a user-selected default or an implicit user-sel...