Mean Opinion Score (MOS) has become a very popular indicator of perceived media quality. While there is a clear benefit to such a “reference quality indicator” and its widespread acceptance, MOS is often applied without sufficient consideration of its scope or limitations.
Basically, eMOS (estimated MOS) is just a different parameter that evaluates MOS with a different approach. An important point to highlight is that these quality indicators cannot be interchanged one to one. As follow we will describe the eMOS evaluation.
How eMOS can be estimated?
An estimation of the human based MOS score (estimated MOS – eMOS), can be pursuit with algorithms that focus on modem response time, codec speed and other tests that predict how the voice quality would be perceived by the human ear. This approach, compared with the standard one (MOS), is more practical and scalable.
In fact, based on this scalability and simplicity, most VoIP measuring applications use MOS scoring based on the e-Model (defined in ITU-T G.107).
The e-model is an algorithm which is designed to evaluate factors of a transmission which lead to a deterioration of the quality of that data traffic, specifically in a cumulative or composite measure, and produce a result – called “Rating Factor” (R-factor) – in the overall quality of a conversation.
The E-model is based on a reference connection that, as shown below, is split into a send side and a receive side.
The model estimates the conversational quality from mouth to ear as perceived by the user at the receive side, both as listener and talker.
The resulting rating factor R, which combines all transmission parameters relevant for the considered connection, is obtained as following:
- “R0” is the basic signal-to-noise ratio (SNR), including noise sources such as circuit noise and room noise
- Factor “Is” is a combination of all impairments which occur with the voice signal
- Factor “Id” represents the impairments caused by delay
- The effective equipment impairment factor “Ie-eff” represents impairments caused by low bit-rate codecs. It also includes impairment due to randomly distributed packet losses
- The advantage factor “A” allows for compensation of impairment factors when the user benefits from other types of access to the user
The default values for all input parameters used in the algorithm of the E-model, are listed in the ITU-T G.107 reference. If all parameters are set to the default values, the calculation results in a very high quality with a rating factor of R = 93.2. For a more detailed description of these parameters refer to the ITU-T G.107.
From E-Model to eMOS
After obtaining the R value, a relatively standard formula to convert that to a MOS score is often applied:
1. Start out with an R-value of 93.2, and deduct from there, based on network conditions. Take the average latency, add jitter, but double the impact to latency, then add 10 for protocol latencies:
VoIP calls often are in the 3.5 to 4.2 MOS range. The following chart can be used as a guide for VoIP MOS testing and a good comparison for voice quality.
|Maximum for G.711 codec||4.4|
|Some users satisfied||3.6-4.0|
|Many users dissatisfied||3.1-3.6|
|Nearly all users dissatisfied||2.6-3.1|