As the TWS (True wireless stereo) headset market continues to grow, users' demand for product experience has also upgraded from simple quick connections to higher standards. For example, as of this year, a large number of TWS headsets with clear calls as their selling point have emerged on the market.
In order to enable clear voice communication in a very noisy environment, is it possible to generate a solution that combines the signals from the inner ear and the external microphone to achieve an intelligent, environment-adaptive sub-band mixer technology? In fact, some domestic and foreign algorithm companies are committed to this and have achieved certain results.
Of course, many solution companies now have call noise reduction solutions that specifically emphasize edge AI (this is one), but in fact they are more about optimizing existing call noise reduction solutions, so let’s remove this part and first look at some basic introductions, that is, what call noise reduction can specifically do.
In general, call noise reduction relies on the synchronization of Uplink and Downlink. It is roughly Microphone Array/AEC/NS/EQ/AGC/DRC, and the logical relationship is as follows:
ADM (Adaptive Directional Microphone Array) Adaptive omnidirectional microphone array
This is a digital signal processing technology that creates a directional or noise canceling microphone using only two omnidirectional microphones. ADM automatically changes its directional characteristics to provide optimal noise attenuation in a variety of environments while fully maintaining signal quality. The adaptive process is fast and frequency selective, and can cancel multiple interferences simultaneously.
In addition to its good directional characteristics, ADM is less affected by wind noise than traditional acoustic directional microphones. ADM technology allows two types of microphone configurations: "endfire" and "broadside".
In an endfire configuration, the signal source (the user's mouth) is on-axis (the line connecting the two microphones). In a broadside configuration, it's aimed at a line on the horizontal axis.
In the end-fire configuration, the ADM has two modes of operation; "far-talk" and "close-talk". In far-talk mode, the ADM acts as an optimal directional microphone, attenuating signals from the back and sides while retaining the signal in front. In close-talk mode, the ADM acts as the best noise-canceling microphone, effectively eliminating distant sounds. The relatively free acoustic design makes the ADM ideal for mobile phones, as it can "softly" switch between far-end speakers and near-end speakers. However, when this type of design is used on headphones, especially TWS headphones, it is more restricted by whether the user wears it correctly. Similar to airpods, the author has observed many people wearing them in "weird" ways on the subway. Some are due to the shape of the user's ears, and some are due to wearing habits, which results in the algorithm not necessarily working under ideal conditions.
Acoustic Echo Canceller (AEC)
When part of a signal in a duplex (simultaneous two-way) communication returns to the source signal, it is called "echo." In long-distance analog and almost all digital communication systems, even a small echo signal can cause interference due to the significant round-trip delay.
In voice communication terminals, acoustic echo is generated due to the acoustic coupling between the loudspeaker and the microphone. Due to the nonlinear processing applied in the communication channel, such as lossy vocoders and transcoding, the acoustic echo must be processed (cancelled) locally inside the device.
Noise Suppressor (NS)
Noise suppression technology reduces stationary and transient noise in single-channel speech signals, improving the signal-to-noise ratio, enhancing speech clarity, and reducing listening fatigue.
Of course, there are many specific methods for this part, such as BF (Beamforming) or PF (Post filter) adjustment methods. Generally speaking, AEC, NS, BF, and PF are the core parts of call noise reduction. Of course, the advantages and disadvantages of each algorithm solution provider are different.
Automatic Gain Control (AGC) and Dynamic Range Compressor (DRC)
In a typical voice communication system, the level of the voice signal may vary over a wide range due to the different distances between the user and the microphone and due to the characteristics of the communication channel.
Dynamic range compression (DRC) is the simplest way to equalize signal levels. Compression reduces the dynamic range of a signal by lowering (compressing) loud speech segments while leaving soft speech segments intact. As a result, the entire signal can be additionally amplified so that soft signals can be heard better.
AGC technology digitally increases signal gain (amplifies) when the speech signal is weak, and compresses it when the speech signal is strong. In noisy places, people tend to speak loudly, and this automatically sets the microphone channel gain to a small value, thereby reducing surrounding noise while maintaining the voice of interest at an optimal level. Alternatively, in a quiet environment, people speak relatively quietly, so their voices will be amplified by the algorithm without too much noise.
Previous article:The Importance of Ultra-High Signal-to-Noise for MEMS Microphones
Next article:Concept and classification of speech synthesis (TTS)
- Popular Resources
- Popular amplifiers
- Red Hat announces definitive agreement to acquire Neural Magic
- 5G network speed is faster than 4G, but the perception is poor! Wu Hequan: 6G standard formulation should focus on user needs
- SEMI report: Global silicon wafer shipments increased by 6% in the third quarter of 2024
- OpenAI calls for a "North American Artificial Intelligence Alliance" to compete with China
- OpenAI is rumored to be launching a new intelligent body that can automatically perform tasks for users
- Arm: Focusing on efficient computing platforms, we work together to build a sustainable future
- AMD to cut 4% of its workforce to gain a stronger position in artificial intelligence chips
- NEC receives new supercomputer orders: Intel CPU + AMD accelerator + Nvidia switch
- RW61X: Wi-Fi 6 tri-band device in a secure i.MX RT MCU
Professor at Beihang University, dedicated to promoting microcontrollers and embedded systems for over 20 years.
- LED chemical incompatibility test to see which chemicals LEDs can be used with
- Application of ARM9 hardware coprocessor on WinCE embedded motherboard
- What are the key points for selecting rotor flowmeter?
- LM317 high power charger circuit
- A brief analysis of Embest's application and development of embedded medical devices
- Single-phase RC protection circuit
- stm32 PVD programmable voltage monitor
- Introduction and measurement of edge trigger and level trigger of 51 single chip microcomputer
- Improved design of Linux system software shell protection technology
- What to do if the ABB robot protection device stops
- CGD and Qorvo to jointly revolutionize motor control solutions
- CGD and Qorvo to jointly revolutionize motor control solutions
- Keysight Technologies FieldFox handheld analyzer with VDI spread spectrum module to achieve millimeter wave analysis function
- Infineon's PASCO2V15 XENSIV PAS CO2 5V Sensor Now Available at Mouser for Accurate CO2 Level Measurement
- Advanced gameplay, Harting takes your PCB board connection to a new level!
- Advanced gameplay, Harting takes your PCB board connection to a new level!
- A new chapter in Great Wall Motors R&D: solid-state battery technology leads the future
- Naxin Micro provides full-scenario GaN driver IC solutions
- Interpreting Huawei’s new solid-state battery patent, will it challenge CATL in 2030?
- Are pure electric/plug-in hybrid vehicles going crazy? A Chinese company has launched the world's first -40℃ dischargeable hybrid battery that is not afraid of cold
- [Runhe Neptune Review] Five PWM
- How to use the wireless transceiver module
- Understanding of MSP430G2553
- A lightweight and precise positioning algorithm based on multiple inertial measurement units (LSM6DSOX)
- PT1000 circuit help
- Brief analysis of the working principle of DSP
- Share post: Just completed FPGA interpolation filter design
- Introduction to Common PCB Microvia Technology
- Analysis of the design principle of single chip microcomputer music program
- Beetle_ESP32-C3 Arduino routine learning