Call Progress Detection

CPD (Call Progress Detection) is a call analysis technology that monitors the audio signals during and after call connection to detect call progress events — ringing, busy signals, answering machines, live voice, fax tones, and SIT (Special Information Tones) for disconnected numbers. CPD is broader than AMD, encompassing not just answering machine detection but the full range of call outcome identification.

How It Works in VICIdial

In VICIdial, CPD is available as an alternative AMD method through the amd_type setting. When CPD is selected, VICIdial uses a more comprehensive audio analysis engine that detects multiple call progress states beyond the basic HUMAN/MACHINE classification. CPD can identify: live voice, answering machine, fax machine, SIT tones (indicating disconnected/invalid numbers), busy signals that bypass carrier signaling, and silence.

CPD’s answering machine detection works similarly to AMD — analyzing initial audio duration, pauses, and patterns — but typically uses different algorithms and sensitivity tuning. Some CPD implementations use frequency analysis to detect the distinct audio characteristics of answering machine beeps, providing an additional detection method beyond speech pattern analysis.

The CPD results are mapped to AMDSTATUS values for compatibility with VICIdial’s routing logic. The AMD agent route options apply identically regardless of whether AMD or CPD performed the detection.

Why It Matters

CPD provides value beyond answering machine detection. Its ability to identify fax machines, SIT tones, and other non-voice outcomes helps clean lead data automatically — numbers that consistently return fax tones or SIT tones can be flagged for removal, improving list quality and reducing wasted dial attempts.

For operations where standard AMD accuracy is insufficient, CPD may provide better results depending on the implementation and call profile. Testing both AMD and CPD on your specific traffic and comparing accuracy metrics is the best way to determine which performs better for your campaigns. See the AMD optimization feature page for comparison testing methodologies.