λ₯νμ΄λ§μ λ¬΄μ¨ μΌμ νλ κ³³μΈκ°μ?
What does DeepHearing do?
λ₯νμ΄λ§μ λ€λ
κ°μ μΈκ° μ²κ°μ°κ΅¬μ μΈκ³΅μ§λ₯κΈ°μ μ°κ΅¬λ₯Ό λ°νμΌλ‘ ν κΈ°μ λ ₯μΌλ‘,νμ μ μΈ μ¨λΌμΈ μμ¬μν΅ λ°©μμ μ§μμ μΌλ‘ κ³ λ―Όνκ³ λ°μ μν€κ³ μμ΅λλ€.
λ₯νμ΄λ§μ μΈμ μ΄λμλ , λꡬλ κΉ¨λνκ³ λͺ
λ£νκ² μν΅ν μ μλ μΈμμ μ§ν₯ν©λλ€.
Deep Hearing is a technology based on many years of research on human hearing and artificial intelligence technology, and is continuously thinking and developing innovative online communication methods.
Deep Hearing aims for a world where anyone can communicate clearly and clearly anytime, anywhere.
Β
μλμ λ€μν λ²μ μ μ€λμ€ μνλ€μ 체νν΄λ³΄μΈμ!
Try the various versions of the audio samples below!
μΌμμ μΈ νκ²½(everyday environment) ποΈΒ
μΌλ°μ μΌλ‘ λλ μ μλ νκ²½μμμ λ
Έμ΄μ¦ μ κ±° λ°λͺ¨ μ€λμ€μ
λλ€.
Denoise demo audio in an environment that can be felt in general.
Cafe
Before
After
Metro
Before
After
μλλ¬μ΄ νκ²½(noisy environment) π§Β
μΌμμ μΈ νκ²½λ³΄λ€ νΉμν νκ²½μμμ λ
Έμ΄μ¦ μ κ±° λ°λͺ¨ μ€λμ€μ
λλ€.
This is a denoising demo audio in a special environment rather than a normal one.
Drill
Before
After
Drone
Before
After
DATASET LICENSES
Howling(νμΈλ§) π
π‘Howling(νμΈλ§)μ΄λ?
β μ΄λ€ μ₯μΉμ μΆλ ₯μ΄ μ£Όλ³ μ
λ ₯ μ₯μΉλ‘ λ€μ΄κ°μ μ¦νλμ΄ λ€μ μΆλ ₯λλ μΌμ΄ λ°λ³΅λλ νμ. μ νκΈ°μμλ μ΄λ‘ μΈνμ¬ μμμ΄ μ§μλμ΄ μ¬λλ€μ κ²½νμ μ νμν¬ μ μμ΅λλ€.
β A phenomenon in which the output of a certain device enters a peripheral input device and is amplified and output again is repeated. On the phone, this can cause persistent noise and degrade people's experience.
βΒ μλ¦¬κ° ν¬λ μμ²μ μ£Όμ λ°λλλ€. (Please be careful when watching as the sound is loud.)β

Acoustic Echo(μμ½)
π‘Acoustic Echo(μμ½)λ?
β μ΄λ€ μ₯μΉμ μΆλ ₯μ΄ λ€μ μ
λ ₯ μ₯μΉλ‘ λ€μ΄κ°μ μΆλ ₯μ΄ λ°λ³΅λλ νμ.
β A phenomenon in which the output of a device goes back to the input device and the output is repeated.
μμ½ λ° λ Έμ΄μ¦ μ κ±° (Echo and Denoise)
Echo+Speech
Speaker
Mic
Enhanced
Noise+Speech
Mic
Enhanced
Echo+Noise+Speech
Speaker
Mic
Enhanced
Β
DEVICE
PIXEL 4
DATASET LICENSES
λ°©ν₯μ± μ리 μ ν(Directional sound selection) π£
λμΌ λ§μ΄ν¬λ₯Ό μ΄μ©νμ¬ μνλ λμμ λͺ©μ리λ₯Ό μ¬λ¬ νμ μμμ λΆλ¦¬ν©λλ€.
It uses dual microphones to separate the voice of the desired target from among multiple speakers.

μ£Όλ³μ리 λ°μ λΆλ₯¨ μ μ΄(Ambient sound response volume control) ποΈ
μλλ¬μ΄ νκ²½μμ μμ±μ μΈμνμ¬ μμ± ν¬κΈ°μ λ°λΌ λΆλ₯¨μ μ μ΄ν©λλ€.
Recognizes your voice in a noisy environment and controls the volume according to the volume of your voice.

Β
μ€λ¬Έ(Survey)
Β