中国电子技术网

设为首页 网站地图 加入收藏

?

[原创] TI TIDA-01589高保真近场双向音频参考设计

关键词:消费类电子 yabo88wap下载亚博体育--任意三数字加yabo.com直达官网家居 IP网络照相机 TIDA-01589

时间:2019-09-06 11:33:50?????? 来源:中电网

TI公司的TIDA-01589是具有降噪和回声消除的高保真近场双向音频参考设计.该参考设计演示了双麦克风通过立体声ADC音频输入,低功耗DSP进行降噪,回声消除和其它音质增强的算法.该设计还具有TIyabo88wap下载亚博体育--任意三数字加yabo.com直达官网放大器(Smart Amp)技术以提高质量和得到微型扬声器的高SPL的音频输出.音频作为人机接口(HMI)在许多建筑物自动化设备中如IP网络相机,Wi-Fi安全照相机,视频门铃,温度计,烟雾检测器,对讲机面板,yabo88wap下载亚博体育--任意三数字加yabo.com直达官网扬声器,音频和视频会议系统等已是新兴的趋势.在整个yabo88wap下载亚博体育--任意三数字加yabo.com直达官网家居设备中音频可提供全双工免提通信,使得用户体验更便捷更舒适和更可靠.从嘈杂环境中提取出清晰语音或音频的系统能力对使用语音激活和语音识别是十分重要的.特别是,免提系统实现语音激活和语音识别同时还具有音频播放通路.这又是语音识别系统的另一中挑战,因为它增加了其它的噪音元件.所以,免提通信系统有以下几种挑战:1)降噪,2)回声消除,3)音频去混响,4)双向通话中保证全双工性能,5)处理音频数据和运行几种算法提高系统性能的资源有限,6)音频数据实时处理能力,7)采用微型扬声器和扬声器保护所带来的更高声压电平(SPL).本参考设计TIDA-01589提出了以上挑战的解决方案.比如采用双向音频模块平台,TLV320ADC3101立体声ADC,TMS320C5517 DSP,yabo88wap下载亚博体育--任意三数字加yabo.com直达官网放大器TAS2557以及基于AM335x BeagleBone? Black Wireless支持可选择云连接等.本文介绍了参考设计TIDA-01589主要特性和主要指标,框图以及参考设计TIDA-01589 DSP板和MIC板的电路图,材料清单和PCB设计图.

Man machine interaction requires an acoustic interfacefor providing full duplex hands-free communication. Inhands-free mode, part of the far-end or near-end audiosignal from the speaker is coupled to the microphones.

Furthermore, in noisy environments the microphonesalso capture ambient noise in addition to the usefulnear-end audio signal. The captured multi-microphoneaudio signals are corrupted by acoustic backgroundnoise as well as by echoic signals that significantly degrade the intelligibility of the desired signal, andrestrict the performance of subsequent audioprocessing systems. This reference designdemonstrates dual microphones for audio input viastereo ADC, low-power DSP performing noisereduction, acoustic echo cancellation, and other audioquality enhancement algorithms. This design alsofeatures TI’s Smart Amp technology for high-qualityand high-SPL audio output from a micro-speaker.

Audio as Human Machine Interface (HMI) is an emerging trend in many building automation endequipmentsuch as IP network camera, Wi-Fi security camera, video doorbell, thermostat, smoke detector,intercom panel, smart speaker, audio and video conferencing systems, and others. The whole idea ofhaving audio in smart home devices is to provide full duplex hands-free communication making the userexperience more simple, comfortable, and reliable. The ability of the system to extract clear speech or audio from a noisy environment is important to applications that use voice activation and speechrecognition. Particularly, hands-free systems implementing voice-activation and speech recognition alsoincorporate an audio playback path. This represents another challenge to voice recognition systems, as itadds other noise components. The realization of a hands-free communication system requires solutions toseveral fundamental challenges:

? Noise reduction: Handling background noise to extract clear audio
? Echo cancellation: Ensuring that there is no echo under any usage condition
? De-reverberation: Ensuring consistent performance even during changes to echo path
? Ensuring a full-duplex performance during double talk
? Limited on-chip resources to process audio data and running several algorithms to improve systemperformance
? Real-time processing of captured audio data along with additional features like voice activation andspeech recognition
? Higher sound pressure level (SPL) using micro-speaker and speaker protection

There are various ways that audio artifacts can be introduced to the system degrading the quality ofcaptured audio. Ambient noise in the environment picked up by the microphones can create problems thatcorrupt the audio signal being transmitted and heard on the other end. Signal processing techniques suchas adaptive spectral noise reduction (ASNR) and automatic gain control (AGC) are used to improveoverall system performance ensuring high fidelity communication. ASNR helps to reduce background and broadband noise with a minimal reduction in signal quality and automatic gain control (AGC) helps toboost lower level speech signals in hands-free environments. Spectral noise reduction uses the frequencydomain to effectively remove ambient and unwanted noise components from an audio signal, making iteasier to hear the desired signal. Using adaptive filtering techniques creates an estimate of the signal’snoise component. This noise component, when subtracted from the composite signal, reveals the best possible representation of the actual speech signal.

Acoustic echo occurs when the audio being played back is picked up directly or indirectly by themicrophones and fed back in to the system. It can be exacerbated when the audio playback volume ishigh or the speaker and microphones are in close proximity. While some of the echo can be mitigated bykeeping distance between the microphone and speaker however, this might not be possible in small spaceconstrained systems. A more robust answer is to use acoustic echo cancellation (AEC) algorithms. Inbrief, an AEC samples audio from the far side as it comes in and compares it to audio from captured by microphones. If any of the audio matches, the AEC removes it by creating adaptive filters to prevent effectof echo in real time. A critical attribute of the acoustic echo cancellation is its ability to quickly detect thetime variances and the reflected audio signals and adapt to a changing environment while minimizing anyaudio distortion. In real-life, this could be introduced by a participant who is moving while speaking or is ina dynamic environment that might have changing attributes like a door opening and closing.

参考设计TIDA-01589主要特性:

? Two-way audio modular platform supporting analogor digital microphone for audio capture andplayback on micro-speaker using TI Smart Amptechnology
? TLV320ADC3101 Stereo ADC with 92-dBA SNR,0.003% THD integrated PGA, AGC, and mini-DSPfor implementing decimation filtering
? Powerful TMS320C5517 DSP for implementingvarious audio-processing algorithms like: AdaptiveSpectral Noise Reduction (ASNR) and AcousticEcho Cancellation (AEC) handling echo tails up-to128 ms
? Smart Amp TAS2557 with integrated ultra-lownoiseaudio DAC, class-D power amplifier andbuilt-in speaker sensing and protection to deliverhigh SPL using micro-speaker
? Attaches to AM335x-based BeagleBone? BlackWireless to support optional Cloud connectivity

参考设计TIDA-01589应用:

? Video Doorbell
? Thermostat
? IP Network Camera
? Wi-Fi? Camera
? Smart Home Alarm Systems

图1. 参考设计TIDA-01589外形图

图2. 参考设计TIDA-01589框图

参考设计TIDA-01589主要指标:


图3.TIDA-01589硬件的HD AEC信号流程图

图4. TIDA-01589电源树图

图5. TIDA-01589演示建立图

图6. TIDA-01589测试建立图

DSP板

图7. TIDA-01589 DSP板电路图(1)

图8. TIDA-01589 DSP板电路图(2)

图9. TIDA-01589 DSP板电路图(3)

图10. TIDA-01589 DSP板电路图(4)

图11. TIDA-01589 DSP板电路图(5)

图12. TIDA-01589 DSP板电路图(6)

图13. TIDA-01589 DSP板电路图(7)

图14. TIDA-01589 DSP板电路图(8)

图15. TIDA-01589 DSP板电路图(9)

图16. TIDA-01589 DSP板电路图(10)

图17. TIDA-01589 DSP板电路图(11)

图18. TIDA-01589 DSP板电路图(12)
TIDA-01589 DSP板材料清单:





图19. TIDA-01589 DSP板PCB设计图(1)

图20. TIDA-01589 DSP板PCB设计图(2)

图21. TIDA-01589 DSP板PCB设计图(3)

图22. TIDA-01589 DSP板PCB设计图(4)

图23. TIDA-01589 DSP板PCB设计图(5)

图24. TIDA-01589 DSP板PCB设计图(6)

图25. TIDA-01589 DSP板PCB设计图(7)

图26. TIDA-01589 DSP板PCB设计图(8)

MIC板

图27. TIDA-01589MIC板电路(1)

图28. TIDA-01589 MIC板电路(2)

图29. TIDA-01589 MIC板电路(3)

TIDA-01589 MIC板材料清单:


图30. TIDA-01589 MIC板PCB设计路(1)

图31. TIDA-01589 MIC板PCB设计路(2)

图32. TIDA-01589 MIC板PCB设计路(3)

图33. TIDA-01589 MIC板PCB设计路(4)

图34. TIDA-01589 MIC板PCB设计路(5)

图35. TIDA-01589 MIC板PCB设计路(6)

图36. TIDA-01589 MIC板PCB设计路(7)

图37. TIDA-01589 MIC板PCB设计路(8)
详情请见:
http://www.ti.com/lit/ug/tidue77/tidue77.pdf
tidrxb5.zip
tidrxb4.zip
tidrxb7.zip
tidue77.pdf

  • 分享到:

?

猜你喜欢

  • 新品
  • 新闻
  • 方案