The second article in a series looking at the new version of ITU-T Rec. P.863 POLQA. Speaker dependency in the Narrowband Model is the focus here.
Aspects of Skype performance between two LTE phones on the network, over WiFi and between WiFi and the network
A new version of ITU-T Rec. P.863 POLQA is now included in MultiDSLA. Discover what has changed in v2.4.
Measuring the performance of Skype and other VoIP clients is easy with MultiDSLA. Identify poor speech quality, long delay and other issues quickly, accurately and repeatably.
LinkedIn Twitter Facebook Google+ Weibo VKontakte Viadeo Flipboard SkypeThe time alignment algorithm in POLQA was designed to be much better than that in PESQ. This example comes from a G.711 connection through an ATA in a test lab. There are no audible impairments but why is POLQA narrowband (NB) showing a score of 4.18 compared to the PESQ P.862.1 score of 3.58? …
LinkedIn Twitter Facebook Google+ Weibo VKontakte Viadeo Flipboard Skype50% of POLQA Sales made to Operators in the First Year. POLQA (ITU-T P.863) was published by the ITU in January 2011. We released one of the first commercial implementations of POLQA in August of the same year. The year that has passed since we released POLQA has certainly been a year of learning. …
Understand how POLQA time aligns signals before analysing surface views. POLQA and PESQ align reference and degraded signals in time using rather different philosophies. Read more before you start analysing results in detail. If you’ve been using PESQ for years, you may be confused.
Expect reporting of small delay variations which are not actually there. POLQA has a sophisticated time-alignment algorithm to account for both step-delay changes and time-warping. Read this article to avoid being confused when POLQA reports small delay variations on a perfect network.
Ensure POLQA reference files pass the transparency test. Here we explain the hows and the whys of checking that your reference speech material is appropriate for POLQA testing. Read this article to avoid running hours of worthless testing because your reference speech files were not good enough.
Don’t use more than 10s of speech in narrowband mode. For long recordings of, say, 32 seconds POLQA can show a drop in score of around 1.1 compared to the average of four 8s measurements.