Silence detection: Difference between revisions

From VoIPmonitor.org
(Add section explaining how to use silence detection for identifying calls with background noise and differentiating between caller/called party noise)
(Rewrite: consolidated structure, added parameter table, See Also section, improved formatting)
 
(One intermediate revision by the same user not shown)
Line 2: Line 2:
[[Category:GUI manual]]
[[Category:GUI manual]]


'''This guide explains how to enable and use VoIPmonitor's advanced audio analysis features for G.711 codecs, including silence detection and audio clipping detection. These tools are invaluable for diagnosing audio quality issues that are not caused by network packet loss, such as one-way audio, muted microphones, or faulty endpoint hardware.'''
'''This guide explains VoIPmonitor's audio analysis features for G.711 codecs, including silence and clipping detection. These tools diagnose audio quality issues not caused by network problems, such as one-way audio, muted microphones, or faulty hardware.'''


== Overview & Use Cases ==
== Overview ==
Standard network metrics like packet loss and jitter can only identify problems that occur during network transport. However, many audio quality issues originate at the endpoint itself. The Silence & Clipping Detection engine analyzes the decoded G.711 (a-law/u-law) audio stream to identify these problems.
 
Standard network metrics (packet loss, jitter) only identify transport-layer problems. Many audio issues originate at endpoints. The Silence & Clipping Detection engine analyzes decoded G.711 audio to identify these problems.


<kroki lang="plantuml">
<kroki lang="plantuml">
Line 29: Line 30:
</kroki>
</kroki>


Enable this feature to diagnose issues like:
'''Use cases:'''
* A user complaining about "dead air" when the network statistics look perfect
* "Dead air" complaints when network statistics look perfect
* One-way audio scenarios where one party is not transmitting any sound
* One-way audio (one party not transmitting)
* Identifying calls with audio clipping, which is often perceived as clicks or distortion
* Audio clipping (clicks or distortion)
* More accurately assessing call quality by penalizing MOS scores for periods of absolute silence
* Accurate [[Glossary|MOS]] calculation by penalizing silence periods
* '''Detecting calls in noisy environments''' - using silence threshold to identify calls with background noise (see [[#Detecting_Calls_with_Background_Noise|Noise Detection Workflow below]])
* Detecting calls in noisy environments (see [[#Detecting Calls with Background Noise|below]])


== Prerequisites ==
== Prerequisites ==
Before enabling this feature, please ensure the following requirements are met:


;1. Codec Limitation:
{| class="wikitable"
:This feature currently only supports calls using the '''G.711 a-law/u-law''' codec. Audio streams from other codecs will not be analyzed.
|-
! Requirement !! Details
|-
| '''Codec''' || G.711 a-law/u-law only. Other codecs are not analyzed.
|-
| '''CPU''' || Audio decoding is CPU-intensive. Monitor system load after enabling.
|-
| '''Database''' || Modern installations (2020+) have required columns. Legacy systems may need schema updates (see below).
|}


;2. CPU Resources:
== Configuration ==
:Decoding and analyzing audio streams is a CPU-intensive process. Enabling these features will increase the overall CPU load on the sensor. Monitor your system's performance after enabling them.


;3. Database Schema:
=== Parameters ===
:All modern VoIPmonitor installations (approximately since 2020) include the necessary database columns by default. If you are running a current version, '''you do not need to perform any database modifications'''. The sensor will log an error on startup if any required columns are missing, which should only happen on very old legacy systems.


== Configuration ==
{| class="wikitable"
|-
! Parameter !! Default !! Description
|-
| <code>silencedetect</code> || no || Enable silence detection and MOS adjustment
|-
| <code>clippingdetect</code> || no || Enable audio clipping detection
|-
| <code>silencethreshold</code> || 512 || Audio level threshold for silence. Lower = stricter detection.
|}


=== Step 1: Enable in voipmonitor.conf ===
=== Enable in voipmonitor.conf ===
To activate the features, add or uncomment the following lines in <code>/etc/voipmonitor.conf</code>:


<syntaxhighlight lang="ini">
<syntaxhighlight lang="ini">
# Enables both silence and clipping detection logic
# /etc/voipmonitor.conf
silencedetect = yes
silencedetect = yes
# Enables the specific feature that detects clipped audio frames
clippingdetect = yes
clippingdetect = yes
# Defines the audio level threshold for silence. The default of 512 is
# tolerant to low-level background noise. Lower values are stricter.
silencethreshold = 512
silencethreshold = 512
</syntaxhighlight>
</syntaxhighlight>
After making changes, restart the voipmonitor service:


<syntaxhighlight lang="bash">
<syntaxhighlight lang="bash">
Line 71: Line 78:
</syntaxhighlight>
</syntaxhighlight>


=== Step 2: For Legacy Upgrades - Adding Database Columns ===
{{Note|1=For legacy installations (pre-2020) that report missing columns on startup, add them via SQL:
'''Note:''' This step is only required if you are upgrading from a very old VoIPmonitor installation and the sensor logs errors about missing columns on startup. For all recent installations, these columns already exist.
 
If required, you can add the columns to your <code>cdr</code> table with the following SQL command. Only add the columns that are reported as missing.
 
<syntaxhighlight lang="sql">
<syntaxhighlight lang="sql">
ALTER TABLE cdr
ALTER TABLE cdr
Line 89: Line 92:
   ADD COLUMN b_mos_silence_mult10 tinyint unsigned DEFAULT NULL;
   ADD COLUMN b_mos_silence_mult10 tinyint unsigned DEFAULT NULL;
</syntaxhighlight>
</syntaxhighlight>
}}


== Features & Metrics ==
== Features & Metrics ==
Once enabled, you will have access to several new metrics and visualizations throughout the GUI.


=== 1. Silence as Packet Loss & MOS Adjustment ===
=== Silence as Packet Loss (MOS Adjustment) ===
The system identifies frames of absolute silence (e.g., from a muted mic) and treats them as a form of packet loss. This allows VoIPmonitor to correctly calculate a low MOS score for calls that were silent, even if no actual packets were lost on the network.


This is visualized in the CDR Graph Detail, where a new MOS score labeled '''"Sil"''' (Silence MOS) will appear.
Silence frames (e.g., muted mic) are treated as packet loss, allowing accurate [[Glossary|MOS]] calculation for silent calls. In CDR Graph Detail, a new '''"Sil"''' (Silence MOS) score appears.


[[File:cdr_spectral_silence_example.png]]
[[File:cdr_spectral_silence_example.png]]
[[File:cdr_graph_silence_lowmos_example.png]]
=== Overall Silence Percentage ===


[[File:cdr_graph_silence_lowmos_example.png]]
CDR Detail shows total silence percentage for caller and callee streams.


=== 2. Overall Silence Percentage ===
{{Tip|1=>95% silence in one direction strongly indicates one-way audio.}}
The CDR Detail view will show the total percentage of silence for both the caller and callee audio streams. A call with over 95% silence in one direction is a strong indicator of a one-way audio problem.


[[File:cdr_detail_silence_table.png]]
[[File:cdr_detail_silence_table.png]]


=== 3. Silence at the End of a Call ===
=== Silence at End of Call ===
This metric measures the duration of silence leading up to the end of the call. A high value may indicate that a user hung up out of frustration because they couldn't hear the other party. This value is shown directly in the main CDR listing.
 
Measures silence duration before hangup. High values may indicate user frustration (couldn't hear other party). Shown in main CDR listing.


[[File:cdr_summary_silence_example.png]]
[[File:cdr_summary_silence_example.png]]


=== 4. Audio Clipping Detection ===
=== Audio Clipping Detection ===
The system counts the number of "clipped" audio frames, which occur when the audio signal is too loud and hits the maximum amplitude, causing distortion. This metric is shown in the CDR detail and can be used for filtering and creating charts to find problematic endpoints.
 
Counts clipped frames (signal too loud, hits max amplitude = distortion). Shown in CDR detail; usable for filtering and charts to find problematic endpoints.


[[File:charts_silence_values.png]]
[[File:charts_silence_values.png]]


=== 5. MOS Scores During Silence Periods ===
=== MOS During Silence Periods ===
When silence detection is enabled, VoIPmonitor calculates additional MOS values that appear after the standard MOS scores in the CDR table. These values represent the voice quality specifically during periods of silence in the call.


{| class="wikitable"
{| class="wikitable"
Line 124: Line 129:
! Column !! Description
! Column !! Description
|-
|-
| <code>a_mos_silence_mult10</code> / <code>b_mos_silence_mult10</code> || '''Average''' MOS score during silence periods (caller/callee)
| <code>a_mos_silence_mult10</code> / <code>b_mos_silence_mult10</code> || Average MOS during silence (caller/callee)
|-
|-
| <code>a_mos_silence_min_mult10</code> / <code>b_mos_silence_min_mult10</code> || '''Minimum''' MOS score during silence periods (caller/callee)
| <code>a_mos_silence_min_mult10</code> / <code>b_mos_silence_min_mult10</code> || Minimum MOS during silence (caller/callee)
|}
|}


'''Important Notes:'''
{{Note|These metrics only populate when <code>silencedetect=yes</code>. Otherwise, regular MOS is shown.}}
* These metrics are '''only active when silence detection is enabled''' (<code>silencedetect = yes</code>)
* If silence detection is disabled, the system shows the full/regular MOS score instead
* These values help identify calls where audio quality deteriorates during silent periods, which may indicate endpoint problems
 
=== 6. Detecting Calls with Background Noise ===


You can use the silence detection feature to identify calls occurring in noisy environments (e.g., call centers with background chatter, factories with machinery, or offices with environmental noise). This works because '''calls with background noise are not registered as silent''' - the continuous audio above the threshold prevents silence detection from triggering.
== Detecting Calls with Background Noise ==


'''How it works:'''
Calls with background noise show '''low silence percentage''' because continuous audio prevents silence detection.


{| class="wikitable"
{| class="wikitable"
|-
|-
! Audio Environment !! Silence Detection Behavior !! Typical Silence % Range
! Environment !! Silence % Range
|-
|-
| Clean / quiet office | Normal silence during pauses | 40-70%
| Clean/quiet office || 40-70%
|-
|-
| Environmental noise (air conditioning, distant voices) | Partial detection | 10-39%
| Environmental noise (A/C, distant voices) || 10-39%
|-
|-
| High noise (factory floor, crowded call center) | Almost no silence detected | 0-9%
| High noise (factory, crowded call center) || 0-9%
|}
|}


'''Workflow to identify noisy environments:'''
'''Workflow:'''
# Enable <code>silencedetect=yes</code> with appropriate <code>silencethreshold</code>
# Filter CDR for <code>Caller Silence < 20%</code> OR <code>Called Silence < 20%</code>
# Review flagged calls' audio recordings to confirm noise
# Create saved filters or [[Alerts|alerts]] for ongoing monitoring


1. Configure <code>silencedetect = yes</code> and <code>silencethreshold = 512</code> (or appropriate level for your environment)
{{Tip|Use spectral analysis in CDR Graph Detail to distinguish white/brown noise (continuous spectrum) from sporadic noise (intermittent) and genuine speech (natural patterns with pauses).}}


2. Identify sample calls from different environments:
== See Also ==
  * Calls from "clean" offices (e.g., administrative staff, IT department)
  * Calls from "noisy" locations (e.g., sales floor with headset chatter, manufacturing site)


3. Compare the <code>caller_silence</code> and <code>called_silence</code> percentages in the CDR:
* [[Glossary]] - MOS, jitter, packet loss definitions
  * Use the CDR filter to show <code>Caller Silence &lt 20%</code> OR <code>Called Silence &lt 20%</code>
* [[Comprehensive_Guide_to_VoIP_Voice_Quality]] - Voice quality factors
  * This will display calls with continuous audio (potential background noise)
* [[Call_Detail_Record_-_CDR]] - CDR view and filtering
* [[Charts]] - Creating charts from silence/clipping data
* [[Alerts]] - Setting up alerts for quality thresholds


4. Analyze the audio recordings of these low-silence calls to confirm the presence of noise
== AI Summary for RAG ==


5. Create saved filters or alerts for systematic monitoring
'''Summary:''' VoIPmonitor's Silence and Clipping Detection analyzes G.711 (a-law/u-law) audio to diagnose issues invisible to network metrics. Key features: treats silence as packet loss for accurate MOS calculation, calculates silence percentage (>95% indicates one-way audio), measures end-of-call silence (user frustration indicator), counts clipped frames (distortion), and detects calls with background noise (low silence %). Configuration requires <code>silencedetect=yes</code> and <code>clippingdetect=yes</code> in voipmonitor.conf. Modern installations have required database columns by default.
 
'''Tip:''' Use the spectral analysis view in CDR Graph Detail to visually distinguish between:
* White/brown noise (continuous frequency spectrum - machinery, air conditioning)
* Sporadic noise (intermittent - voices, doors slamming, equipment beeps)
* Genuine speech (natural frequency patterns with pauses for silence)
 
== AI Summary for RAG ==
'''Summary:''' VoIPmonitor's Silence and Clipping Detection analyzes G.711 (a-law/u-law) audio to diagnose issues invisible to network metrics. Key features: treats silence as packet loss for accurate MOS, calculates silence percentage (>95% indicates one-way audio), measures end-of-call silence, counts clipped frames, and '''detects calls with background noise''' (low silence %). Configuration requires <code>silencedetect=yes</code> and <code>clippingdetect=yes</code> in voipmonitor.conf. Modern installations have required database columns by default.


'''Keywords:''' silence detection, clipping, one-way audio, dead air, MOS, G.711, silencedetect, clippingdetect, silencethreshold, mos_silence, background noise, noisy environments
'''Keywords:''' silence detection, clipping, one-way audio, dead air, MOS, G.711, silencedetect, clippingdetect, silencethreshold, mos_silence, background noise, noisy environments, audio analysis, endpoint issues


'''Key Questions:'''
'''Key Questions:'''
Line 181: Line 177:
* Why is a call's MOS score high even though the user reported hearing nothing?
* Why is a call's MOS score high even though the user reported hearing nothing?
* How does VoIPmonitor detect silence in a call?
* How does VoIPmonitor detect silence in a call?
* What does the "Silence MOS" score mean in the CDR graph?
* What does the "Sil" (Silence MOS) score mean in the CDR graph?
* How can I find calls with audio clipping?
* How can I find calls with audio clipping?
* How can I filter calls based on background noise and differentiate between caller and called party noise?
* How can I filter calls based on background noise?
* What silence percentage indicates a noisy environment?

Latest revision as of 16:47, 8 January 2026


This guide explains VoIPmonitor's audio analysis features for G.711 codecs, including silence and clipping detection. These tools diagnose audio quality issues not caused by network problems, such as one-way audio, muted microphones, or faulty hardware.

Overview

Standard network metrics (packet loss, jitter) only identify transport-layer problems. Many audio issues originate at endpoints. The Silence & Clipping Detection engine analyzes decoded G.711 audio to identify these problems.

Use cases:

  • "Dead air" complaints when network statistics look perfect
  • One-way audio (one party not transmitting)
  • Audio clipping (clicks or distortion)
  • Accurate MOS calculation by penalizing silence periods
  • Detecting calls in noisy environments (see below)

Prerequisites

Requirement Details
Codec G.711 a-law/u-law only. Other codecs are not analyzed.
CPU Audio decoding is CPU-intensive. Monitor system load after enabling.
Database Modern installations (2020+) have required columns. Legacy systems may need schema updates (see below).

Configuration

Parameters

Parameter Default Description
silencedetect no Enable silence detection and MOS adjustment
clippingdetect no Enable audio clipping detection
silencethreshold 512 Audio level threshold for silence. Lower = stricter detection.

Enable in voipmonitor.conf

# /etc/voipmonitor.conf
silencedetect = yes
clippingdetect = yes
silencethreshold = 512

systemctl restart voipmonitor

ℹ️ Note: For legacy installations (pre-2020) that report missing columns on startup, add them via SQL: 

ALTER TABLE cdr
  ADD COLUMN caller_silence tinyint unsigned default NULL,
  ADD COLUMN called_silence tinyint unsigned default NULL,
  ADD COLUMN caller_silence_end smallint default NULL,
  ADD COLUMN called_silence_end smallint default NULL,
  ADD COLUMN caller_clipping_div3 smallint unsigned default NULL,
  ADD COLUMN called_clipping_div3 smallint unsigned default NULL,
  ADD COLUMN a_mos_silence_min_mult10 tinyint unsigned DEFAULT NULL,
  ADD COLUMN b_mos_silence_min_mult10 tinyint unsigned DEFAULT NULL,
  ADD COLUMN a_mos_silence_mult10 tinyint unsigned DEFAULT NULL,
  ADD COLUMN b_mos_silence_mult10 tinyint unsigned DEFAULT NULL;

Features & Metrics

Silence as Packet Loss (MOS Adjustment)

Silence frames (e.g., muted mic) are treated as packet loss, allowing accurate MOS calculation for silent calls. In CDR Graph Detail, a new "Sil" (Silence MOS) score appears.

Overall Silence Percentage

CDR Detail shows total silence percentage for caller and callee streams.

💡 Tip: >95% silence in one direction strongly indicates one-way audio.

Silence at End of Call

Measures silence duration before hangup. High values may indicate user frustration (couldn't hear other party). Shown in main CDR listing.

Audio Clipping Detection

Counts clipped frames (signal too loud, hits max amplitude = distortion). Shown in CDR detail; usable for filtering and charts to find problematic endpoints.

MOS During Silence Periods

Column Description
a_mos_silence_mult10 / b_mos_silence_mult10 Average MOS during silence (caller/callee)
a_mos_silence_min_mult10 / b_mos_silence_min_mult10 Minimum MOS during silence (caller/callee)

ℹ️ Note:

Detecting Calls with Background Noise

Calls with background noise show low silence percentage because continuous audio prevents silence detection.

Environment Silence % Range
Clean/quiet office 40-70%
Environmental noise (A/C, distant voices) 10-39%
High noise (factory, crowded call center) 0-9%

Workflow:

  1. Enable silencedetect=yes with appropriate silencethreshold
  2. Filter CDR for Caller Silence < 20% OR Called Silence < 20%
  3. Review flagged calls' audio recordings to confirm noise
  4. Create saved filters or alerts for ongoing monitoring

💡 Tip: Use spectral analysis in CDR Graph Detail to distinguish white/brown noise (continuous spectrum) from sporadic noise (intermittent) and genuine speech (natural patterns with pauses).

See Also

AI Summary for RAG

Summary: VoIPmonitor's Silence and Clipping Detection analyzes G.711 (a-law/u-law) audio to diagnose issues invisible to network metrics. Key features: treats silence as packet loss for accurate MOS calculation, calculates silence percentage (>95% indicates one-way audio), measures end-of-call silence (user frustration indicator), counts clipped frames (distortion), and detects calls with background noise (low silence %). Configuration requires silencedetect=yes and clippingdetect=yes in voipmonitor.conf. Modern installations have required database columns by default.

Keywords: silence detection, clipping, one-way audio, dead air, MOS, G.711, silencedetect, clippingdetect, silencethreshold, mos_silence, background noise, noisy environments, audio analysis, endpoint issues

Key Questions:

  • How can I diagnose one-way audio problems?
  • Why is a call's MOS score high even though the user reported hearing nothing?
  • How does VoIPmonitor detect silence in a call?
  • What does the "Sil" (Silence MOS) score mean in the CDR graph?
  • How can I find calls with audio clipping?
  • How can I filter calls based on background noise?
  • What silence percentage indicates a noisy environment?
Retrieved from "https://www.voipmonitor.org/doc/index.php?title=Silence_detection&oldid=10036"