Glossary: Difference between revisions

From VoIPmonitor.org
No edit summary
 
(6 intermediate revisions by 2 users not shown)
Line 2: Line 2:
{{DISPLAYTITLE:Glossary of VoIP & Monitoring Terms}}
{{DISPLAYTITLE:Glossary of VoIP & Monitoring Terms}}


'''This glossary defines key terms and metrics related to VoIP quality and network performance, explaining how each is measured and utilized within the VoIPmonitor application.'''
'''Quick reference for VoIP quality metrics and monitoring concepts used in VoIPmonitor.'''


== Core Network Quality Metrics ==
== Network Quality Metrics ==
These metrics describe the fundamental health and stability of the network path carrying the VoIP traffic.


=== Packet Loss ===
=== Packet Loss ===
Packet loss occurs when one or more data packets traveling across a network fail to reach their destination. It is a critical issue in VoIP, as lost audio packets can result in audible gaps, clicks, or dropouts in the conversation. Common causes include network congestion, faulty hardware, signal degradation, or misconfigured network devices.
Lost data packets cause audible gaps, clicks, or dropouts. Common causes: congestion, faulty hardware, misconfiguration.


==== In the Context of VoIPmonitor ====
'''VoIPmonitor approach:''' Records loss distribution (how many consecutive packets lost in various intervals) rather than just average percentage. This matters because 2% random loss is far less noticeable than a 2-second burst of 100% loss.
VoIPmonitor detects packet loss for each call direction (caller and callee) and stores a detailed distribution of these losses. Instead of just a single percentage, it records how many consecutive packets were lost in various intervals. This is crucial because a 2% loss spread randomly is far less noticeable than a single 2-second burst of 100% loss, even though the average percentage might be similar.


=== Packet Delay Variation (PDV) or Jitter ===
=== Packet Delay Variation (PDV) / Jitter ===
In networking, Packet Delay Variation (PDV) is the measure of how much the arrival time of packets differs from their expected, consistent interval. For VoIP, this is commonly referred to as **jitter**. High jitter means packets are arriving in erratic, unpredictable bursts, which can severely degrade voice quality even if no packets are lost.
Variation in packet arrival times from expected intervals. High jitter = erratic packet bursts = degraded quality even without packet loss.


==== In the Context of VoIPmonitor ====
'''VoIPmonitor approach:''' Measures packets exceeding delay thresholds:
VoIPmonitor measures PDV by comparing the arrival time of each RTP packet against the expected interval (typically 20ms for most codecs). It records the number of packets that exceed certain delay thresholds. This detailed breakdown is more valuable than a single average jitter value, as it helps identify specific patterns of delay. The default PDV intervals measured are:
* 50–70ms, 70–90ms, 90–120ms, 120–150ms, 150–300ms, >300ms
* 50–70ms
 
* 70–90ms
{{Note|1='''Constant jitter values''' (same value throughout call) indicate clock mismatch in device, not network issues. '''Zero jitter with large initial delay''' = one-time buffering spike, not ongoing jitter.}}
* 90–120ms
* 120–150ms
* 150–300ms
* >300ms


=== Post-Dial Delay (PDD) ===
=== Post-Dial Delay (PDD) ===
Post-Dial Delay is the time a user waits from the moment they finish dialing the last digit to the moment they hear feedback, such as a ringback tone or a busy signal. Long PDD can create a poor user experience, as callers may think the call has failed and hang up prematurely.
Time from last digit dialed to first feedback (ringback/busy tone). Long PDD causes users to hang up prematurely.


== Mechanisms to Handle Network Issues ==
== Mitigation Mechanisms ==
These are techniques used by VoIP endpoints to mitigate the effects of poor network quality.


=== Jitter Buffer (or De-Jitter Buffer) ===
=== Jitter Buffer ===
A jitter buffer is a temporary storage area at the receiving end of a VoIP call. It intentionally delays incoming audio packets slightly, allowing them to be reordered and played out in a smooth, continuous stream, effectively hiding the negative effects of network jitter (PDV). The size of the buffer determines the maximum amount of jitter it can compensate for. Some systems use **adaptive jitter buffers** that can dynamically change their size based on current network conditions.
Temporary storage at receiving end that delays and reorders packets for smooth playback. Types:
* '''Fixed:''' Constant buffer size
* '''Adaptive:''' Dynamically adjusts based on network conditions


=== Packet Loss Concealment (PLC) ===
=== Packet Loss Concealment (PLC) ===
Packet Loss Concealment (PLC) is a technique used to mask the effects of lost audio packets. Since retransmitting lost packets is not feasible in real-time voice conversations, the receiving device must intelligently "guess" what the missing audio sounded like.
Masks lost packets since retransmission is not feasible in real-time voice:


==== Common PLC Techniques ====
{| class="wikitable"
* '''Zero Insertion:''' Lost audio frames are simply replaced with silence. This is the crudest method.
|-
* '''Waveform Substitution:''' The system repeats the last known audio frame to fill the gap. This is a common and effective technique, an example of which is defined in ITU recommendation G.711 Appendix I.
! Technique !! Description
* '''Model-Based Methods:''' Advanced algorithms use speech models to interpolate or extrapolate the missing audio, providing a more natural-sounding result.
|-
| Zero Insertion || Replace with silence (crudest method)
|-
| Waveform Substitution || Repeat last known frame (common, per G.711 Appendix I)
|-
| Model-Based || Interpolate missing audio using speech models (best quality)
|}


== Voice Quality & Performance Measurement ==
== Voice Quality Metrics ==
These metrics are used to quantify the quality of a call or the performance of a network.


=== Mean Opinion Score (MOS) ===
=== Mean Opinion Score (MOS) ===
MOS is a standardized numerical rating of perceived voice quality, ranging from 1 (bad) to 5 (excellent). Originally a subjective test where human listeners would rate call quality, it is now typically calculated objectively using algorithms like the one defined in the ITU-T P.862 (PESQ) standard.
MOS is a standardized numerical rating of perceived voice quality, ranging from 1 (bad) to 5 (excellent). Originally a subjective test where human listeners would rate call quality, it is now typically calculated objectively using algorithms like the one defined in the ITU-T P.862 (PESQ) standard.


{| class="wikitable"
{| class="wikitable"
|+ MOS Ratings
|-
|-
! MOS !! Quality !! Impairment
! MOS !! Quality !! Impairment
Line 64: Line 64:
|}
|}


==== Codecs and MOS ====
'''Codec baseline scores:'''
Different audio codecs have different maximum theoretical MOS scores due to their compression methods.
{| class="wikitable"
{| class="wikitable"
|+ Example Codec MOS Scores
|-
|-
! Codec !! Data rate [kbit/s] !! Typical MOS
! Codec !! Bitrate !! Typical MOS
|-
|-
| G.711 (ISDN) || 64 || 4.1
| G.711 || 64 kbit/s || 4.1
|-
|-
| iLBC || 15.2 || 4.14
| iLBC || 15.2 kbit/s || 4.14
|-
|-
| AMR || 12.2 || 4.14
| AMR || 12.2 kbit/s || 4.14
|-
|-
| G.729 || 8 || 3.92
| G.729 || 8 kbit/s || 3.92
|-
|-
| G.723.1 r63 || 6.3 || 3.9
| G.723.1 || 6.3 kbit/s || 3.9
|}
 
==== VoIPmonitor MOS Calculation ====
VoIPmonitor calculates '''parametric MOS''' from network metrics (packet loss, PDV), not the audio signal. It simulates jitter buffer performance:
 
{| class="wikitable"
|-
|-
| G.726 ADPCM || 32 || 3.85
! Score !! Jitter Buffer Model !! Use Case
|-
| '''MOS F1''' || Fixed 50ms || Very sensitive to jitter
|-
| '''MOS F2''' || Fixed 200ms || Moderate tolerance
|-
| '''MOS adapt''' || Adaptive up to 500ms || Real-world endpoint simulation
|}
|}


==== VoIPmonitor's Parametric MOS ====
Default calculation uses G.711 codec with PLC for consistent cross-call comparison.
VoIPmonitor calculates a **parametric MOS score** based on network transport quality (Packet Loss and PDV), not the audio signal itself. It simulates how a jitter buffer would perform and derives a MOS score from the simulated packet loss. It provides three MOS scores based on different jitter buffer models:
* '''MOS F1:''' Simulates a fixed 50ms jitter buffer. It is very sensitive to jitter.
* '''MOS F2:''' Simulates a fixed 200ms jitter buffer.
* '''MOS adapt:''' Simulates an adaptive jitter buffer that can grow up to 500ms.


[[File:mos.png|VoIPmonitor calculates MOS based on simulated packet loss and PDV, using a pre-calculated surface for G.711 with PLC.]]
[[File:mos.png|VoIPmonitor calculates MOS based on simulated packet loss and PDV, using a pre-calculated surface for G.711 with PLC.]]


By default, VoIPmonitor uses a calculation based on the G.711 codec with PLC for all calls. This provides a consistent baseline for comparing network quality across all calls, regardless of the actual codec used.
=== R-Factor ===
{{Warning|1='''VoIPmonitor does NOT calculate R-Factor.''' R-Factor (ITU-T G.107 E-Model, 0-100 scale) is redundant because MOS provides equivalent information with direct mathematical correlation.}}
 
'''Recommended approach instead:'''
* Track '''MOS percentiles''' (%95, %99) not averages
* Monitor '''changes over time''' against historical baselines
* Use VoIPmonitor's aggregation by source IP/number


=== RTCP (RTP Control Protocol) ===
=== RTCP ===
RTCP is a companion protocol to RTP that provides out-of-band statistics and control information for a media stream. It gathers metrics directly from the endpoints, such as transmitted packet counts, lost packet counts, jitter, and round-trip delay time. VoIPmonitor can parse these RTCP reports and store the statistics for each call, providing an alternative, endpoint-reported view of call quality.
RTP Control Protocol provides endpoint-reported statistics: transmitted/lost packets, jitter, round-trip delay. VoIPmonitor parses RTCP reports for alternative view of call quality.


=== Answer-Seizure Ratio (ASR) ===
== Carrier-Grade Metrics ==
ASR is a carrier-grade metric for network quality, defined by ITU E.411. It measures the percentage of successfully answered calls out of the total number of attempted calls (seizures). It is calculated as:
`(Total Answered Calls / Total Seizures) * 100`
A low ASR can indicate network problems, but it is also affected by user behavior (e.g., busy signals, unanswered calls).


=== Network Effectiveness Ratio (NER) ===
=== ASR (Answer-Seizure Ratio) ===
NER is a similar metric to ASR but is designed to measure only the network's ability to deliver a call, excluding user behavior. For NER, calls that reach the destination but are rejected (e.g., busy, no answer) are still counted as "successful" from the network's perspective. VoIPmonitor allows you to configure which SIP response codes are considered successful for NER calculation in the system settings.
Percentage of answered calls from total attempts (ITU E.411):


=== Average Call Duration (ACD) ===
<code>ASR = (Answered Calls / Total Seizures) × 100</code>
ACD is the average length of all answered telephone calls. A very low ACD, especially when combined with a low ASR, can indicate problems with call quality (e.g., users hanging up due to poor audio).
 
Low ASR can indicate network problems but is also affected by user behavior (busy, no answer).
 
=== NER (Network Effectiveness Ratio) ===
Like ASR but measures only network capability—calls reaching destination but rejected by user (busy, no answer) count as "successful." Configure which SIP codes are successful in Settings.
 
=== ACD (Average Call Duration) ===
Average length of answered calls. Low ACD combined with low ASR often indicates quality problems (users hanging up due to poor audio).


== Statistical Concepts ==
== Statistical Concepts ==


=== Percentiles ===
=== Percentiles ===
A percentile is a measure indicating the value below which a given percentage of observations in a group of observations falls. VoIPmonitor often uses the 95th or 99th percentile (shown as `%95` or `%99`) for metrics like MOS.
Value below which a given percentage of observations falls.


For example, a `MOS %95` score of 3.2 means that 5% of all calls had a MOS score of 3.2 or worse. This is a much more useful indicator of systemic problems than an average, which can be easily skewed by a large number of good calls.
'''Example:''' <code>MOS %95 = 3.2</code> means 5% of calls had MOS of 3.2 or worse. More useful than averages for identifying systemic problems.
 
== See Also ==
* [[Comprehensive_Guide_to_VoIP_Voice_Quality]] - Detailed voice quality analysis
* [[Alerts]] - Configure quality-based alerts using these metrics
* [[Charts]] - Visualize metrics over time
* [[Silence_detection]] - Additional audio quality analysis


== AI Summary for RAG ==
== AI Summary for RAG ==
'''Summary:''' This document is a glossary of key performance indicators (KPIs) and technical concepts related to VoIP quality analysis. It defines terms such as Packet Loss, Packet Delay Variation (PDV), Jitter, and Post-Dial Delay (PDD). It also explains technologies designed to mitigate network issues, like Jitter Buffers and Packet Loss Concealment (PLC). A significant portion is dedicated to the Mean Opinion Score (MOS), detailing how it is calculated, the differences between subjective and parametric MOS, and how VoIPmonitor specifically simulates three types of MOS scores (F1, F2, adapt) based on network conditions. Finally, it defines carrier-level metrics like Answer-Seizure Ratio (ASR), Network Effectiveness Ratio (NER), and Average Call Duration (ACD), and explains statistical concepts like Percentiles.
 
'''Keywords:''' glossary, packet loss, pdv, jitter, jitter buffer, mos, mos score, pesq, pdd, rtcp, asr, ner, acd, plc, packet loss concealment, percentile, g.711, g.729, codec, quality, kpi, metric
'''Summary:''' Glossary of VoIP quality metrics and monitoring terms for VoIPmonitor. Covers network metrics (Packet Loss with distribution tracking, PDV/Jitter with threshold intervals, PDD), mitigation mechanisms (Jitter Buffer types, PLC techniques), voice quality measurements (MOS 1-5 scale with codec baselines, VoIPmonitor's parametric MOS calculation using F1/F2/adapt jitter buffer simulations), and carrier metrics (ASR, NER, ACD). Key clarification: VoIPmonitor does NOT calculate R-Factor - use MOS percentiles (%95, %99) and historical trend monitoring instead. RTCP provides endpoint-reported alternative statistics.
 
'''Keywords:''' glossary, packet loss, pdv, jitter, jitter buffer, mos, mos f1, mos f2, mos adapt, r-factor, pesq, pdd, rtcp, asr, ner, acd, plc, packet loss concealment, percentile, g.711, g.729, codec, quality, kpi, metric, voip quality, parametric mos
 
'''Key Questions:'''
'''Key Questions:'''
* What is the difference between Jitter and Packet Delay Variation (PDV)?
* What is Packet Delay Variation (PDV) and how does VoIPmonitor measure it?
* How does VoIPmonitor calculate MOS? What are MOS F1, F2, and adapt?
* How does VoIPmonitor calculate MOS? What are MOS F1, F2, and adapt?
* Does VoIPmonitor calculate R-Factor? What should I use instead?
* What is the difference between ASR and NER?
* What is the difference between ASR and NER?
* What is Packet Loss Concealment (PLC)?
* What is Packet Loss Concealment (PLC) and what techniques exist?
* How do I interpret a 95th percentile MOS score?
* How do I interpret MOS percentile scores like %95?
* What is a Jitter Buffer and what does it do?
* What is a Jitter Buffer and what types exist?
* What MOS score should I expect for different codecs?

Latest revision as of 16:59, 8 January 2026


Quick reference for VoIP quality metrics and monitoring concepts used in VoIPmonitor.

Network Quality Metrics

Packet Loss

Lost data packets cause audible gaps, clicks, or dropouts. Common causes: congestion, faulty hardware, misconfiguration.

VoIPmonitor approach: Records loss distribution (how many consecutive packets lost in various intervals) rather than just average percentage. This matters because 2% random loss is far less noticeable than a 2-second burst of 100% loss.

Packet Delay Variation (PDV) / Jitter

Variation in packet arrival times from expected intervals. High jitter = erratic packet bursts = degraded quality even without packet loss.

VoIPmonitor approach: Measures packets exceeding delay thresholds:

  • 50–70ms, 70–90ms, 90–120ms, 120–150ms, 150–300ms, >300ms

ℹ️ Note: Constant jitter values (same value throughout call) indicate clock mismatch in device, not network issues. Zero jitter with large initial delay = one-time buffering spike, not ongoing jitter.

Post-Dial Delay (PDD)

Time from last digit dialed to first feedback (ringback/busy tone). Long PDD causes users to hang up prematurely.

Mitigation Mechanisms

Jitter Buffer

Temporary storage at receiving end that delays and reorders packets for smooth playback. Types:

  • Fixed: Constant buffer size
  • Adaptive: Dynamically adjusts based on network conditions

Packet Loss Concealment (PLC)

Masks lost packets since retransmission is not feasible in real-time voice:

Technique Description
Zero Insertion Replace with silence (crudest method)
Waveform Substitution Repeat last known frame (common, per G.711 Appendix I)
Model-Based Interpolate missing audio using speech models (best quality)

Voice Quality Metrics

Mean Opinion Score (MOS)

MOS is a standardized numerical rating of perceived voice quality, ranging from 1 (bad) to 5 (excellent). Originally a subjective test where human listeners would rate call quality, it is now typically calculated objectively using algorithms like the one defined in the ITU-T P.862 (PESQ) standard.

MOS Quality Impairment
5 Excellent Imperceptible
4 Good Perceptible but not annoying
3 Fair Slightly annoying
2 Poor Annoying
1 Bad Very annoying

Codec baseline scores:

Codec Bitrate Typical MOS
G.711 64 kbit/s 4.1
iLBC 15.2 kbit/s 4.14
AMR 12.2 kbit/s 4.14
G.729 8 kbit/s 3.92
G.723.1 6.3 kbit/s 3.9

VoIPmonitor MOS Calculation

VoIPmonitor calculates parametric MOS from network metrics (packet loss, PDV), not the audio signal. It simulates jitter buffer performance:

Score Jitter Buffer Model Use Case
MOS F1 Fixed 50ms Very sensitive to jitter
MOS F2 Fixed 200ms Moderate tolerance
MOS adapt Adaptive up to 500ms Real-world endpoint simulation

Default calculation uses G.711 codec with PLC for consistent cross-call comparison.

VoIPmonitor calculates MOS based on simulated packet loss and PDV, using a pre-calculated surface for G.711 with PLC.

R-Factor

⚠️ Warning: VoIPmonitor does NOT calculate R-Factor. R-Factor (ITU-T G.107 E-Model, 0-100 scale) is redundant because MOS provides equivalent information with direct mathematical correlation.

Recommended approach instead:

  • Track MOS percentiles (%95, %99) not averages
  • Monitor changes over time against historical baselines
  • Use VoIPmonitor's aggregation by source IP/number

RTCP

RTP Control Protocol provides endpoint-reported statistics: transmitted/lost packets, jitter, round-trip delay. VoIPmonitor parses RTCP reports for alternative view of call quality.

Carrier-Grade Metrics

ASR (Answer-Seizure Ratio)

Percentage of answered calls from total attempts (ITU E.411):

ASR = (Answered Calls / Total Seizures) × 100

Low ASR can indicate network problems but is also affected by user behavior (busy, no answer).

NER (Network Effectiveness Ratio)

Like ASR but measures only network capability—calls reaching destination but rejected by user (busy, no answer) count as "successful." Configure which SIP codes are successful in Settings.

ACD (Average Call Duration)

Average length of answered calls. Low ACD combined with low ASR often indicates quality problems (users hanging up due to poor audio).

Statistical Concepts

Percentiles

Value below which a given percentage of observations falls.

Example: MOS %95 = 3.2 means 5% of calls had MOS of 3.2 or worse. More useful than averages for identifying systemic problems.

See Also

AI Summary for RAG

Summary: Glossary of VoIP quality metrics and monitoring terms for VoIPmonitor. Covers network metrics (Packet Loss with distribution tracking, PDV/Jitter with threshold intervals, PDD), mitigation mechanisms (Jitter Buffer types, PLC techniques), voice quality measurements (MOS 1-5 scale with codec baselines, VoIPmonitor's parametric MOS calculation using F1/F2/adapt jitter buffer simulations), and carrier metrics (ASR, NER, ACD). Key clarification: VoIPmonitor does NOT calculate R-Factor - use MOS percentiles (%95, %99) and historical trend monitoring instead. RTCP provides endpoint-reported alternative statistics.

Keywords: glossary, packet loss, pdv, jitter, jitter buffer, mos, mos f1, mos f2, mos adapt, r-factor, pesq, pdd, rtcp, asr, ner, acd, plc, packet loss concealment, percentile, g.711, g.729, codec, quality, kpi, metric, voip quality, parametric mos

Key Questions:

  • What is Packet Delay Variation (PDV) and how does VoIPmonitor measure it?
  • How does VoIPmonitor calculate MOS? What are MOS F1, F2, and adapt?
  • Does VoIPmonitor calculate R-Factor? What should I use instead?
  • What is the difference between ASR and NER?
  • What is Packet Loss Concealment (PLC) and what techniques exist?
  • How do I interpret MOS percentile scores like %95?
  • What is a Jitter Buffer and what types exist?
  • What MOS score should I expect for different codecs?
Retrieved from "https://www.voipmonitor.org/doc/index.php?title=Glossary&oldid=10222"