Zubin Pratap

Ouvrir l’original · Paragraphe 4 Source du texte original

Extrait original

It is tempting to choose the models based on benchmarks that measure them in laboratory conditions. But real-life conversations are messy, and that makes enterprise-grade voice agents a hard engineering problem. We cannot solve these hard problems by treating voice models as interchangeable commodities. Instead, they must be picked, analysed and measured against their performance for their intended use case .

model evaluation30 juin 2026

TTFS and TTCT as voice latency metrics

The author considers TTFS (Time to Final Segment), also called TTCT (Time to Complete Transcript), the metric that matters for voice agents.

Éléments favorables

Zubin Pratap·30 juin 2026

Ouvrir l’original · Paragraphe 14 Source du texte original

Extrait original

For voice agents, we believe the metric that matters is TTFS - Time to Final Segment (aka the Time to Complete Transcript (TTCT)).

model evaluation30 juin 2026

Turn detection is critical and distinct from ASR accuracy

Turn detection, also called end pointing, detects when a user has finished their turn. The author describes it as extremely hard without visual or other human conversational cues, and says poor turn detection causes long pauses and high latency.

Éléments favorables

Zubin Pratap·30 juin 2026

Ouvrir l’original · Paragraphe 18 Source du texte original

Extrait original

Turn-detection is also known as end pointing – because it refers to detecting when the user has finished their turn. It is extremely hard to do accurately as Speech to Text models do not have visual or other cues like humans do in conversation. That makes high-quality turn detection incredibly important for conversational AI - poor turn detection results in very long awkward pauses and high latency in the voice agent’s conversation.

model evaluation30 juin 2026

WER is misleading without category-level analysis

Word Error Rate can mislead when the error classes are irrelevant or when a non-streaming ASR model is compared with a streaming model. The author says many benchmarks fail to distinguish streaming and non-streaming use cases.

Éléments favorables

Zubin Pratap·30 juin 2026

Ouvrir l’original · Paragraphe 25 Source du texte original

Extrait original

Word Error Rate (WER) for STT/ASR is a misleading metric if you’re looking at the wrong or irrelevant classes of errors. Or if you’re comparing a non-streaming ASR model (which is inherently easier to be accurate on) with a streaming ASR model. Unfortunately many benchmarks do not distinguish between ASR streaming vs non-streaming use cases.