Poor audio quality: Choppy, silent, or robotic calls
Discover the causes of poor audio quality in calls and tips to fix them, thus improving your communication experience.
Symptom / Need
The client reports that the agents showed a degradation in voice quality during the interaction, exhibiting the following behaviors:
• Choppy calls: Intermittent interruptions in the audio during the conversation.
• Robotic audio: Voice with distortion, metallic effect or with perceptible delay.
• Silent calls: Total loss of audio in one or both communication channels (One-way / No-way audio).
• Other symptoms: Any additional behavior that affects audio quality.
Impact : This situation directly affects the customer experience and operational management indicators, such as quality of service, management time, and user satisfaction.
Context / Scenarios
In VoIP environments, audio quality depends on network stability, hardware performance, and the availability of infrastructure resources. Problems can arise from:
• Customer infrastructure: Unstable LAN network, use of Wi-Fi connections, low Internet link capacity, or low-performance computing equipment.
• Server or Platform: Resource saturation (CPU or memory), excessive concurrent calls, or lack of updates.
• Carrier or Telephone Provider: Instability or degradation in SIP trunks.
Timely identification of these causes allows for the restoration of call quality and ensures continuity of operation.
Answer / Solution
1. Minimum Information for Triage
Before making any technical adjustments, gather the following data to document the case:
• Identification Data:
or Name of the affected operation.
or Affected Server.
• Technical Details:
or DID used (Inbound).
or Prefix used (Outbound).
or Signaling and RTP trace.
or Caller IDs.
or Exact date and time of the event (HH:MM:SS format).
or Classification of the affected calls (listen to at least 10 recordings).
or Percentage of affected.
• Work Environment:
or Validate whether the Agents are on site (Inhouse) or working remotely (Home Office).
• Ownership of the trunk line:
or Confirm if the telephone is owned by wolkvox or the customer.
• Recording Evidence:
Listen to the recordings on the platform and validate:
If the recording sounds correct , but the agent reports failures, the problem may be related to the agent's headset, equipment, or local connection.
If the recording is of poor quality , the cause is likely related to the network or the telephone provider.
o Includes the detailed record of calls in which the novelty is evident.
2. Initial Validations
Perform the following validations in logical order:
• Server updates and resources:
or Check if recent updates were performed on the server.
or Validates the status of server resources (Outbound or Inbound).
or Check CPU and RAM usage.
If consumption exceeds 90% or there are more than 500 simultaneous calls, the server may be overloaded, so the support group should be alerted.
Or check if the same behavior occurs with other operations on the same server, and if the issue is with a specific operator.
or Check the quality and stability of the SIP trunks, regardless of whether they are provided by Wolkvox or by the customer.
• Call testing: Perform call tests to and from Wolkvox equipment to validate audio behavior.
• Working Tool: Validate if the problem persists when changing the operating environment:
or Desktop application.
or Web Manager.
If the failure occurs in the desktop environment, perform tests from the web environment, staying in the more stable environment.
• Network Monitoring:
Validate the customer's network service status using the following indicators:
or Internet Speed: Check the upload and download speeds.
Latency: Must be <500ms. It is recommended to generate latency reports and store the logs for at least 30 days.
or Jitter : Must be < 30ms.
or Packet loss: Must be 0%.
Customer LAN Network: Confirm that the network settings are adequate, including:
or VLAN
or QoS (voice prioritization)
or properly configured switches.
• Workstation Configuration:
Equipment : Verify that the workstations meet the minimum requirements:
or Equipment with good performance.
or Certified headbands.
or Network connection preferably wired, avoiding the use of Wi-Fi.
System exceptions: Verify that the following exceptions exist in your antivirus or firewall:
- Files : AGENTBOX.EXE, CONFIGURATIONBOX.EXE.
- Folder : C:\ipdialbox.
- System Registry: Add the exception to the path where wolkvox writes in REGEDIT. Computer\HKEY_CURRENT_USER\Software\AgentBox.
- Firewall : Validates configurations related to: SIP ALG and SIP Helper.
These functions must be disabled to avoid interference with SIP signaling.
3. Escalation Protocol
If the fault persists after initial validations:
Escalation Level: Report the case to Level 2 for authorization to escalate to the CLOUD team.
• Solution Manager: Infrastructure / Cloud Team.
• Escalation Criterion: Escalate when:
or The customer's local network has been ruled out.
or There is a suspicion of server overload or problems in the RTP flow.
Possible Causes
• Server Saturation: High CPU/RAM consumption, or excessive concurrent calls.
• Local Network Deficiencies: Lack of QoS (Voice Prioritization), Internet service congestion, or Wi-Fi usage.
• Firewall Restrictions: Blocking UDP ports used by RTP or active SIP ALG / SIP Helper configurations.
• Hardware problems: Use of non-certified headsets or equipment that does not meet minimum requirements.
Considerations
• Operating margin: In VoIP (Voice over Internet Protocol) environments, an impact of <5% of calls can be considered within the occasional operating margin.
• General behavior: If the same behavior occurs in other operations within the same server, the cause may be related to the shared infrastructure.
Recommendations
• Temporary Solution:
or Temporarily migrate the affected agents from the desktop application to the Web Manager, in order to rule out problems associated with the local software.
• Prevention:
or Maintain latency logs for at least 30 days.
or Ensure that agents use a wired connection.
or Use certified professional headbands.
• Good practices:
or Implement the recommended configurations on workstations and network devices to prevent recurring voice quality incidents.