[asterisk-users] Calculating R Factor and MOS metrics for VoIP

[Steve Underwood]

On 03/09/2010 12:30 AM, Kevin P. Fleming wrote:
> Steve Underwood wrote:
>
>    
>>> MOS and R factor are the two QoS parameters used to estimate VoIP call
>>> quality.
>>>        
>> You can't calculate MOS. Its an assessment based on a lot of human
>> hearing. Nonetheless, there is a profitable industry in pretending to
>> calculate MOS.
>>      
> Right, computers don't have 'opinions' (the 'O' in 'MOS'). However, it
> seems that many people use PESQ scores as a MOS-equivalent for test and
> planning purposes now. However, that requires running predefined samples
> through the system under test, not just calculations based on network
> effects of real calls.
>    
PESQ is just a marketing trick by its perpetrators.

1. Cook up an algorithm. Any old algorithm.
2. Patent some key element of it.
3. Get the algorithm into an ITU spec, so managers think it has 
substance, and want its output in nice reports.
4. Profit.

Steve