05
참고
2026-09-11 조사. 제품 가격·버전은 빠르게 바뀌므로 인용 전 재확인. 일부 논문(ACM DL)은 봇 차단으로 초록 원문 대신 메타데이터로 확인했다.
먼저 읽을 것 10선
- Weiser & Brown, The Coming Age of Calm Technology (1996) — 주변과 중심을 오가는 기술. https://calmtech.com/papers/coming-age-calm-technology
- Amershi et al., Guidelines for Human-AI Interaction (CHI 2019) — 18개 가이드라인. https://dl.acm.org/doi/10.1145/3290605.3300233
- Horvitz, Principles of Mixed-Initiative User Interfaces (CHI 1999) — 언제 끼어들 것인가의 수학. https://dl.acm.org/doi/10.1145/302979.303030
- Porcheron et al., Voice Interfaces in Everyday Life (CHI 2018) — "대화"가 아니라 요청-응답 설계. https://dl.acm.org/doi/10.1145/3173574.3174214
- Beneteau et al., Communication Breakdowns Between Families and Alexa (CHI 2019). https://dl.acm.org/doi/10.1145/3290605.3300473
- Ammari et al., Music, Search, and IoT (TOCHI 2019) — 25만 건 로그. https://dl.acm.org/doi/10.1145/3311956
- Cha et al., Hello There! Is Now a Good Time to Talk? (IMWUT 2020, KAIST 이의진 랩). https://dl.acm.org/doi/abs/10.1145/3411810 · PDF https://kimauk.github.io/file/paper/IMWUT20_smartspeaker.pdf
- Lau, Zimmerman & Schaub, Alexa, Are You Listening? (CSCW 2018). https://dl.acm.org/doi/10.1145/3274371
- Défossez et al., Moshi (2024) https://arxiv.org/abs/2410.00037 + Skantze 턴테이킹 리뷰 (2021) https://www.sciencedirect.com/science/article/pii/S088523082030111X
- Nielsen, AI: First New UI Paradigm in 60 Years (2023) https://www.nngroup.com/articles/ai-paradigm/ + Sesame, Crossing the Uncanny Valley of Conversational Voice (2025) https://www.sesame.com/research/crossing_the_uncanny_valley_of_voice
오픈 음성 스택 · 프로토콜
- Home Assistant Voice PE — https://www.home-assistant.io/voice-pe/
- Voice Chapter 10 (KiCad 공개, 2025-06-25) — https://www.home-assistant.io/blog/2025/06/25/voice-chapter-10/
- Voice Chapter 11 (TTS 스트리밍, 위성당 2 파이프라인) — https://www.home-assistant.io/blog/2025/10/22/voice-chapter-11/
- HA LLM API → MCP 노출 — https://developers.home-assistant.io/docs/core/llm/
- HA MCP Server — https://www.home-assistant.io/integrations/mcp_server/
- xiaozhi-esp32 — https://github.com/78/xiaozhi-esp32
- ElatoAI — https://github.com/akdeb/ElatoAI
- linux-voice-assistant — https://github.com/OHF-Voice/linux-voice-assistant
- voicepe-realtime (OpenAI Realtime + OpenClaw 위임/보고) — https://github.com/TristanBrotherton/voicepe-realtime
- ESPHome voice_assistant — https://esphome.io/components/voice_assistant/
- Wyoming — https://github.com/rhasspy/wyoming
- openWakeWord (모델 CC BY-NC) — https://github.com/dscripka/openWakeWord
- Piper (GPLv3) — https://github.com/OHF-Voice/piper1-gpl · 한국어 요청 스레드 https://github.com/rhasspy/piper/discussions/680
- Speech-to-Phrase (한국어 없음) — https://github.com/OHF-Voice/speech-to-phrase
- Going Direct: ESP32 voice → OpenClaw (HA 사칭 브리지) — https://blog.darrenjrobinson.com/going-direct-esp32-voice-for-openclaw/
- HA 60초 타임아웃 사례 — https://blog.darrenjrobinson.com/hardware-voice-assistant-for-openclaw/
- Pipecat — https://github.com/pipecat-ai/pipecat · pipecat-esp32 https://github.com/pipecat-ai/pipecat-esp32
- Smart Turn v3 — https://huggingface.co/pipecat-ai/smart-turn-v3
- LiveKit ESP32 SDK — https://livekit.com/blog/livekit-sdk-for-esp32-bringing-voice-ai-to-embedded-devices
- Speaches — https://github.com/speaches-ai/speaches
- Kyutai Unmute — https://kyutai.org/unmute/
- MCP 2026-07-28 스펙 — https://blog.modelcontextprotocol.io/posts/2026-07-28/ · 로드맵 https://blog.modelcontextprotocol.io/posts/mcp-roadmap/
- A2A — https://a2a-protocol.org/latest/
- OpenClaw — https://openclaw.ai/ · 연혁 https://en.wikipedia.org/wiki/OpenClaw
- Hermes Agent 음성 모드 — https://hermes-agent.nousresearch.com/docs/user-guide/features/voice-mode
- hermes-ha-integration — https://github.com/WolframRavenwolf/hermes-ha-integration
- 2026 음성 에이전트 지연·비용 정리 — https://www.reactify-solutions.com/articles/voice-ai-agents-production-2026
- 로컬 음성 비서 지연 예산 — https://botmonster.com/smart-home/build-private-local-ai-voice-assistant-2026/
- Alexa HypRank (스킬 라우팅) — https://www.amazon.science/blog/hyprank-how-alexa-determines-what-skill-can-best-meet-a-customers-need
한국어 음성
- 리턴제로 한국어 STT 벤치마크 — https://blog.rtzr.ai/korean-speechai-benchmark/
- 2026 한국어 STT/TTS 비교 — https://claw-ops.com/blog/stt-tts-korean-comparison
- Supertonic v3 — https://www.marktechpost.com/2026/05/15/supertone-releases-supertonic-v3-on-device-text-to-speech-model-with-31-language-support-fewer-reading-failures-and-expression-tags/
- Porcupine (한국어 공식) — https://github.com/Picovoice/porcupine
- HiKE 한영 코드스위칭 벤치마크 — https://arxiv.org/abs/2509.24613
하드웨어
- XMOS XVF3800 — https://www.xmos.com/xvf3800
- ReSpeaker XVF3800 + XIAO ESP32S3 — https://www.seeedstudio.com/ReSpeaker-XVF3800-4-Mic-Array-With-XIAO-ESP32S3-p-6489.html
- ReSpeaker XVF3800 USB — https://www.seeedstudio.com/ReSpeaker-XVF3800-USB-Mic-Array-p-6488.html
- ESP32-S3-BOX-3 — https://www.adafruit.com/product/5835
- IM69D130 데이터시트 — https://www.infineon.com/assets/row/public/documents/24/49/infineon-im69d130-datasheet-en.pdf
- 빔포밍 기초 — https://sistc.com/blog-microphone-beamforming-fundamentals/
- 스마트 스피커 음향 설계 규칙 — https://audioxpress.com/article/smart-speakers-audio-design-rules-for-voice-enabled-devices
- Echo Dot 5세대 티어다운 — https://www.briandorey.com/post/echo-dot-5th-gen-smart-speaker-teardown
- HomePod mini 티어다운 — https://www.ifixit.com/News/49645/homepod-mini-round-full-of-sound-and-now-torn-down
- Voice PE 티어다운 — https://karlquinsland.com/home-assistant-voice-pe-teardown/
- TI TAS2563 — https://www.ti.com/product/TAS2563 · TAS5825M https://www.ti.com/product/TAS5825M
- Dayton ND65-4 — https://www.daytonaudio.com/product/63/nd65-4-2-1-2-aluminum-cone-full-range-driver-4-ohm
- Espressif 커스텀 웨이크워드 — https://docs.espressif.com/projects/esp-sr/en/latest/esp32s3/wake_word_engine/ESP_Wake_Words_Customization.html
- microWakeWord — https://www.kevinahrendt.com/micro-wake-word
- Satellite1 — https://www.cnx-software.com/2025/05/16/satellite1-dev-kit-is-an-home-assistant-compatible-diy-voice-assistant-with-esp32-s3-module-xmos-xu316-audio-processor/
- 2026 HA 위성 비교 — https://www.smarthomeexplorer.com/guides/best-home-assistant-voice-satellite-2026
- Pi 메모리 가격 인상 — https://www.raspberrypi.com/news/more-memory-driven-price-rises/
인증 · 제조 · 지원
- KC 적합성평가 개요 — https://elementkorea.kr/magazine-kc-%EC%A0%81%ED%95%A9%EC%84%B1%ED%8F%89%EA%B0%80-%EC%99%84%EC%A0%84%EC%A0%95%EB%B3%B5/
- 한국 RRA 요건 (모듈 인증 ≠ 완제품 면제) — https://ib-lenhardt.com/kb/rra-requirements
- 인증 비용 가이드 — https://k-rnd.com/electronics-certification-guide/
- 배터리 KC — https://www.ufinebattery.com/ko/blog/lithium-battery-kc-certification-guide/
- Matter 인증 비용 — https://matter-smarthome.de/en/development/how-the-matter-certification-works/
- 금형 비용 — https://www.haizol.com/blog/injection-molding-tooling-cost-china
- 시제품 공정 사다리 — https://creallo.com/ko/blog/posts/prototype-mockup-3dprinting
- EVT/DVT/PVT — https://www.kdproductdev.com/blog/evt-dvt-pvt-demystified-for-consumer-electronics-startups/
- 2026 국내 지원사업 — https://k-rnd.com/2026-startup-support-guide/
- 월패드 망분리 — https://www.genians.co.kr/blog/wallpad
시장 · 제품
- Google Home Speaker — https://blog.google/products-and-platforms/devices/google-nest/google-home-speaker-gemini-features/ · 리뷰 https://itdaily.com/reviews/workplace/google-home-speaker-review/
- Alexa+ — https://www.aboutamazon.com/news/devices/alexa-plus-available-free-prime-members-us
- OpenAI 기기 — https://builtin.com/articles/openai-device
- Humane·Rabbit 사후분석 — https://www.digitalapplied.com/blog/ai-product-failures-2026-sora-humane-rabbit-lessons
- Friend 리뷰 — https://fortune.com/2025/10/03/friend-ai-necklace-review-avi-schiffmann/
- 펜던트 비교 — https://www.layer3labs.io/comparisons/limitless-vs-bee-vs-omi
- 국내 AI 스피커 시장 2026-06 — https://www.ajunews.com/view/20260623113720979
- 클로바 스피커 판매 종료 — https://news.mtn.co.kr/news-detail/2026042010413118950
- ElevenLabs CEO 인터뷰 — https://techcrunch.com/2026/02/05/elevenlabs-ceo-voice-is-the-next-interface-for-ai
- a16z AI hardware — https://a16z.com/ai-hardware/
HAX 연구 (본문에서 인용한 나머지)
- Shneiderman, Human-Centered AI — https://www.tandfonline.com/doi/full/10.1080/10447318.2020.1741118
- Raskin, The Humane Interface — https://raskincenter.org/jef/humane-interface/
- Stivers 2009 — https://www.pnas.org/doi/10.1073/pnas.0903616106
- Roberts & Francis 2013 — https://web.ics.purdue.edu/~francisa/Articles/Roberts-Francis_JASAEL13.pdf
- Bentley 2018 — https://dl.acm.org/doi/10.1145/3264901
- Better to Ask Than Assume (CHI 2024) — https://dl.acm.org/doi/10.1145/3613904.3642193
- Meng 2023 (다중 사용자 프라이버시) — https://dl.acm.org/doi/10.1145/3544549.3585689
- Karpathy LLM OS — https://x.com/karpathy/status/1723140519554105733
- Amazon 다중 에이전트 상호운용 가이드 — https://developer.amazon.com/en-US/alexa/voice-interoperability/design-guide/baseline-guidance
- One-Breath Test — https://developer.amazon.com/en-US/blogs/alexa/post/531ffdd7-acf3-43ca-9831-9c375b08afe0/things-every-alexa-skill-should-do-pass-the-one-breath-tes
- Google PAIR 멘탈모델 — https://pair.withgoogle.com/guidebook-v2/chapters/mental-models/
- Pearl, Designing Voice User Interfaces — https://www.oreilly.com/library/view/designing-voice-user/9781491955406/