a field guide to voice

everything voice, from air pressure to agents.

I’m building a practical A–Z of how machines listen, understand, speak, fail, and earn trust. It will grow slowly, in public.

26 letters6 layersone growing map

No content farm. No fake completeness. Just the concepts, experiments, failures, and diagrams I actually understand.

one voice, many systems

follow a sentence through the stack

Tap any stage. This is the path the atlas will slowly unpack—one experiment and field note at a time.

Capture

A microphone turns air pressure into samples.

the voice atlas

A to Z, without pretending it is simple.

Some letters have stories. Some are just a map for now. The empty space is intentional—I’ll fill it as I learn.

Barge-in

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Beamforming
  • Barge-in
  • Buffers

Conversation

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • CTC
  • Codecs
  • Conversational AI

Diarization

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Diarization
  • Denoising
  • DSP

Echo

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Echo cancellation
  • Endpoints
  • Embeddings

Forced alignment

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Forced alignment
  • Features
  • Front-end processing

Graphemes

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • G2P
  • Gain control
  • Griffin–Lim

Hotwords

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Hotwords
  • Hallucinations
  • HMMs

Interruption

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Intent
  • Inference
  • Interruption

Jitter

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Jitter
  • Jitter buffers
  • Jitter measurement

Keyword spotting

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Keyword spotting
  • KWS
  • Keyphrase biasing

Latency

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Latency
  • Language models
  • Loudness

Microphones

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Microphones
  • Multilingual speech
  • MOS

Noise

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Noise suppression
  • Neural vocoders
  • Normalization

Overlap

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Overlapping speech
  • Opus
  • Out-of-domain audio

Prosody

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Prosody
  • Phonemes
  • Pronunciation

Quality

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Quality evaluation
  • Quantization
  • Question detection

Real time

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Real-time factor
  • Resampling
  • RNN-T

Streaming

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Streaming
  • Speaker ID
  • Speech synthesis

Turn-taking

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Turn-taking
  • TTS
  • Transducers

Utterances

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Utterances
  • Upsampling
  • Underflow

Voice activity

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • VAD
  • Vocoders
  • Voice cloning

Word error rate

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • WER
  • Wake words
  • Waveforms

X-vectors

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • X-vectors
  • Cross-talk
  • Cross-lingual speech

YIN

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • YIN pitch
  • Yield
  • Your voice data

Zero-shot voice

draft

Not written yet. The topic map is ready; the explanation will be Anshul’s.

  • Zero-shot synthesis
  • Zero crossing
  • Zero-shot ASR

field notes

learned by building, not copied from a glossary

the promise

Every complicated voice idea should eventually have a page that makes it click.

Back to A–Z ↑