NotebookLM can now talk 👀 Updates from Google, Bandwidth, SoundHound, Hume and much more 🔥

Google NotebookLM can now explain complex topics out loud.

Top Updates 💪

• Text-to-speech market expected to reach $17B by 2029

• Bandwidth unveils the largest BYOC ecosystem for enterprise solutions

  

• Kanari AI and Kata AI form a strategic partnership to elevate conversational AI

• AI-Media and Speechmatics announce strategic partnership

• Parloa unveils its AI Agent lifecycle platform

  

• Fish Audio unveils open-source TTS with multilingual support and voice cloning

• Prodege launches new conversational AI feature

• Mihup.Ai unveils advanced GenAI suite to boost Contact Centers performance

• LLaMA-Omni: Open-source AI that’s giving Siri and Alexa a run for their money

• Gemini’s voice mode is out now for free on Android

• SoundHound AI acquires Amelia

• Hume unveils EVI 2, its new voice-to-voice foundation model

• Exotel unveils new local cloud and AI solutions for enhanced customer experience

• Enterprise-focused AI translation platform Smartcat raises $43M in funding

Noteworthy 💪

• CX AI startups: Blending innovation with tradition

• Would you let an AI robot handle 90% of your meetings?

• AI-powered meeting tools - Krisp offers transcription and noise cancellation

  

• Jabra unveils its AI-powered speech enhancement solution for contact centers

• Exploring AI ambient clinical voice and AI medical scribes

• Phantom voices: Defend against voice cloning attacks

• The sound of your voice may reveal that you have diabetes

• DeFake tool protects voice recordings from cybercriminals

• Altered AI review: Using AI for real-time voice morphing

---

Science and Demo Corner 😎

• Integrating real-time speech translation API with Python

• Amazon Transcribe: Transforming audio to text

• Retrieval-augmented correction of named entity speech recognition errors

• Apple proposes an algorithm to optimize byte-level STT and compare with UTF-8

• Multi-speaker STT with multi-stage encoders and attention-weighted embedding

• Refining synthesized speech using speaker information and phone masking

• Textless unit-to-unit training for many-to-many multilingual speech-to-speech

• Priority-encoder ensemble for speech recognition

• ABR demonstrates the world's first single chip solution for full vocabulary STT

• Whispered STT based on audio data augmentation and inverse filtering