J10 Hearing Aid chip and algorithms
J10 Hearing Aid chip and algorithms
Ai based hearing aid solution in single RSL10 chip, lower power consumption,great voice quality.

J10 Hearing Aid Chip and Algorithms

Supporting customers in China and the Asia Pacific region, JHEAR supports customers focusing on healthcare applications such as Hearing Health and Biometric Monitoring. For hearing health, a unique system development platform is provided based on an artificial intelligence (AI) platform, enabling them to provide customized features, tailor-made system designs, smartphone apps and hearing aid fitting.

On the basis of single RSL chip,we exploited hearing aid algorithm.On the premise of maintain the quality of voice,we input noise equally that less than 5dB,adding EQ,WDRC,frequency shift,AEC,INR and other algorithm.Which limit the occurence of howlround to some extent.Our output curve is as follows(the smooth of frequency response is able to indicate the probability of howlround,THD+N aslo reflects the quality of voice) Based on customer feedback on hearing aid algorithms.Which own some distinct advantage.For instance, far field voice pickup.Furthurmore,it has clear low frequency , legible voice as well as background noise and so on.

Meanwhile, we also offer supporting web fitting and miniprogram fitting tool,these tools is provided as open source ,so customers are able to implement the fitting application by themselves.

What we can provide:
mature j10 hearing aid chip modules  and remote microphone PCBA
provide customized features, tailor-made system designs, smartphone apps and hearing aid fitting.

J10 Software

  • TX SRC Sample Rate Convert 输入信号的重采样,32k->16k
  • AFC Feedback Control 声反馈处理,线性部分对回波进行处理,非线性部分交给AI。
  • ADC Gain 输入信号增益
  • Analysis Filterbank 子带分析滤波器组,将信号分解成N个Band

  • Synthesis Filterbank 子带合成,将信号还原成时域信号。
  • WDRC 宽带动态反馈控制

  • Auto Attack/Release time fitting 自动 Attack/Release 时间调整

  • StarrySoundAI Hearing Sense Enhancement 
  • Starry AI声信号前端处理算法,不仅针对语音,对听感Hearing Sense整体提升。

  • Starry AI对比传统算法,在处理非稳噪声、非线性问题都有极大优势,在冲击Impulse,风噪Wind Noise,啸叫Nolinear部分的处理上比较明显。
  • Starry AI生成声音存在(Sound Existence)标志,取代传统VAD标志。

  • Starry AI对音乐场景同样能保持极高的还原保真度。
  • Band Hearing Equalizer N-Band均衡滤波器组,针对助听场景生成特定的滤波器组。

  • AGC-O 防破音自动增益控制,尽可能地放大声音
  • RX SRC Sample Rate Convert 输出信号的重采样,16k->32k

  • Pure-Tone Audiomery(PTA) 纯音测听功能,生成任意频率单频音
  • DTMF Generate  任意双频点按键音的生成

  • Frequency Shift  频移算法,<5Hz的频移防止自激振荡
  • Howling Detect  啸叫检测,检测到啸叫通过改变助听器频响尽快降低影响。