cancel
Showing results for 
Search instead for 
Did you mean: 

Understanding Audio and Video Metrics in Prognosis for UC Cloud

Highlighted
Community Manager

Understanding Audio and Video Metrics in Prognosis for UC Cloud

This article explains audio and video metrics shown in the Prognosis for UC Cloud product and how to interpret them when troubleshooting a conversation.

Background

Packets and unreliable networks

Real-time audio and video applications divide their data transmission into smaller units known as packets. Transmission of data across "best-effort" networks i.e. those that don't guarantee a minimum upload and download speed such as those used by Skype for Business can mean that packets are not guaranteed to be delivered.

Transmission of packets across the network entails them travelling through one or more routers, where each leg is known as a "hop". The time taken for a packet to go from one end of the network connection to another (one-way) is known as latency; round-trip time (RTT) is the time taken for a packet to be transmitted and return.

When a network is busy or congested, the routers responsible for forwarding network packets to the next "hop" attempt to store them in memory until they can be transmitted (the delay from their queuing is known as latency). The memory used by routers can become full, at which point they will start to drop packets, causing packet loss.

Packets may also not follow the same route, which along with changes in network congestion, can mean that packets do not always arrive with the same latency. This variation in latency is referred to as jitter, and is an another undesirable characteristic of networks used for real-time applications.

TCP vs UDP

IP-based networks offer TCP as a way to guarantee packet delivery (through retransmission) and delivery in sequence (via sequence numbers), but it is not the preferred delivery mechanism for real-time applications. TCP adds extra overhead, and it's retransmission of lost packets exacerbates delays on congested networks as the number of lost packets accumulates and causes new packets, which are being generated at a constant rate, to queue longer.

TCP is used for non-realtime applications such as web sites (HTTP) or instant messaging applications which far prefer guaranteed delivery over delays in transmission. UDP is the preferred transmission protocol for Skype for Business Online.

Skype for Business Online clients will revert back to TCP when they cannot get a UDP connection due to the routing arrangements on their corporate network.

Audio Metrics

General

Codec

The algorithm used to compress the audio for network transmission.

A codec is used to minimise the amount of network bandwidth required for transmission. The codec selected by Skype for Business will depend on the headset used, capabilities of the computer or mobile device, client version and nature of the audio being transmitted.

Network

Network MOS

MOS (Mean-Opinion Score) is a rating of the perceived listening quality of a call. It was originally derived from aggregating using user input surveys, but automated MOS scores are automatically calculated based on monitoring that takes place during the call.

Network MOS examines only those metrics affected by the network which would impact listening quality. A score above 4 points is generally considered to be "good".

Jitter

Network jitter measures the variance in the latency during a call.

Time-sensitive applications like audio and video need to transmit on a regular cycle, so variance in the transmission periodicity can adversely affect call coherence and quality.

Latency

Latency is the delay in transmission between sender and receiver. It is primarily the result of the geographical distance between participants, but congested or unreliable network conditions (such as WiFi) can also increase latency.

Microphone

Certification

This field specifies whether the microphone was certified for use by Skype for Business Online.

Microsoft maintains a list here.

This field relies on data provided by the device drivers on the user's computer or mobile device, which is often subject to inaccuracies and may not always match the list kept by Skype for Business Online Troubleshooting. If you have a device incorrectly reported as un-certified that you believe is on the Microsoft certified devices list, please raise a support ticket specifying the device name as it is reported by Skype for Business Online Troubleshooting.

 

Speaker

Certification

This field specifies whether the speaker was certified for use by Skype for Business Online.

Microsoft maintains a list here.

This field relies on data provided by the device drivers on the user's computer or mobile device, which is often subject to inaccuracies and may not always match the list kept by Skype for Business Online Troubleshooting. If you have a device incorrectly reported as un-certified that you believe is on the Microsoft certified devices list, please raise a support ticket specifying the device name as it is reported by Skype for Business Online Troubleshooting.

Video

General

Codecs

The video compression algorithm(s) used in the video transmission.

A codec is used to minimise the amount of network bandwidth required for transmission. The codec selected by Skype for Business will depend on the camera used, capabilities of the computer or mobile device, client version and video resolution.

Network

Jitter

Network jitter measures the variance in the latency during a call.

Time-sensitive applications like audio and video need to transmit on a regular cycle, so variance in the transmission periodicity can adversely affect call coherence and quality.

Latency

Latency is the delay in transmission between sender and receiver. It is primarily the result of the geographical distance between participants, but congested or unreliable network conditions (such as WiFi) can also increase latency.

Packet loss

Packet loss refers to the percentage of data packets that were lost during the conversation. High packet loss occurs on congested or limited bandwidth network routes.

Bandwidth

This is an estimate of the available network bandwidth on the network device of the participant. It is not the same as the bit rate, which refers to the amount of bandwidth required (in bits per second) to transmit the data.

Encoding and Transmission

Resolution

The width x height (in pixels) of the video. This can vary depending on the camera used and the size of the window on the participant's device (such as expanded or full-screen mode). It can also vary during the call; usually the last used resolution is captured.

Frame Rate

The frames per second measurement of the video. This may vary in response to network conditions, and is usually bounded by the capability of the camera used.

Bit Rate

The number of bits per second needed to transmit the video, usually expressed in Kbps (kilobits/second) or Mbps (megabits/second). This varies throughout the call and depends on a number of factors, such as the codec, frame rate, resolution and variability of the video stream.

Frame Loss Rate

The percentage of the call that experienced a loss of video frames. This might be seen as similar to packet loss, but in terms of video frames that were dropped in order to maintain a consistent video rate under difficult network conditions.

Low Frame Rate

The percentage of the call that experienced a low frame rate. Skype for Business will attempt to maintain a practical frame rate for video (usually above 14fps), so this reflects the portion of the call that this could not be reasonably archieved.

Low Resolution

The percentage of the call that was spent in low resolution. Skype for Business will target a particular resolution based on the window size or full-screen mode, so this measure should reflect the portion of the call that this resolution could not be maintained due to network conditions (or otherwise).

Quality Ratio

This is a high level metric that measures the amount of the call that was spent in three quality zones (from lowest to highest):

  • CIF (resolutions less than 
  • VGA
  • HD 720

This metric needs to be interpreted with reference to the resolution, as a high percentage in one zone may just be reflective of the target resolution of the call.

Dynamic Capability

Dynamic Capability refers to percentage of a call that a user experienced a high CPU load. Other applications running on the end user's device that consume large amounts or spikes in the CPU load will degrade audio and video quality. Any real-time audio and video process (such as Skype for Business) needs a constant level of access to the CPU to provide smooth rendering to the user.

Post-FEC PLR

Forward Error Correction is a compensation mechanism for packet loss. It does not rely on packet re-transmission, which makes it suitable for real-time applications like audio and video delivery.

Post-FEC PLR is the amount of packet loss after FEC has been applied.

Webinar: Keep the modern workforce connected

Unified Communications has always been an important part of companies' digital transformation efforts due to its ability to enable rich virtual collaboration and communication. But with COVID-19, we've reached a break-through point.

Join Bill Haskins, Sr. Analyst & Partner, Unified Communications at Wainhouse Research, and John Ruthven, CEO at IR discuss UC challenges companies are experiencing due to the COVOID-19 crisis.

Join webinar
Top Liked Members