The FM1093 is an integrated DSP based voice processing platform. With bundled V33 echo cancellation/noise suppression software, FM1093(V33) is ideal for the following three typical applications

1. VoIP phone/video conferencing phone: FM1093(V33) supports 16kHz echo cancellation with wideband audio capability. It delivers 16kHz input sampling rate with support to either 8kHz or 16kHz output (7kHz wideband audio)
2. Portable speakerphone (microphone and speaker integrated in the same enclosure): Fortemedia’s patent pending HyperSampling technology delivers exceptional full duplex performance with 16kHz microphone input sample rate.
3. Detached automotive HFCK: With small array microphone, FM1093(V33) provides beam forming effect to suppress non-stationary noise and preserve voice quality in the noisy environment

FM1093(V33) supports array microphone with two configurations

1. Omni and Uni directional microphone configuration: This configuration provides beam forming to suppress non-stationary noise up to 25dB. For the details of the beam-forming effect please refer to Fortemedia’s white paper.
2. Omni and Omni directional microphone configuration: This configuration provides a conferencing mode which does not have a beam-forming effect.

Both configurations perform up to 65dB acoustic echo cancellation and 13-18dB stationary noise suppression and handles non-linear echo. With its patent pending HyperSampling technology, the full duplex performance for the portable conferencing is second to none particularly in the low Signal-to-Echo ratio environment

FM1093 has integrated DSP, CODECs, RAM, digital interface (I2C, UART, Serial Port, interface to external flash, and I/Os), analog interface (2 microphone inputs, line input, line output, and speaker output). Since there is no embedded software in the chip, external flash is required for bundled V33 software.

V33 Software Features

• Suppress up to 18dB stationary noise and up to 25dB non-stationary noise to provide good voice quality and increase voice recognition rate
• Support two array microphone configurations. Omni-uni directional microphones for beam forming (100 degree angle), and omni-omni directional microphones for canceling non-linear echo and enhancing full duplex
• Patent pending HyperSampling technology offers superior full duplex
• Differential input/output to minimize the RF interference and white noise
• AS3NLF (Adaptive Step-Sized Sub-band NLMS with Non-Linear Filter) delivers excellent acoustic echo cancellation performance: 52dB to 65dB
• Dedicated non-linear echo cancellation process to suppress non-linear echo
• Side tone cancellation (25dB to 35dB)
• Programmability through an optional SHI (Serial Host Interface) or UART for customization
• Support A-Law(8bit), µ-Law(8bit), 13 bit sign extension, 13 bit zero padding, and 16 bit linear PCM mode on Serial Port
• Noise suppression convergence time: 1-2 seconds
• Echo convergence time: <45ms.
• Process delay time from microphone to line output (line_out): ~56ms
• Time from reset to signal out: ~164ms
• 5-band equalizer for line_out with adjustable gain
• Dynamic range control at speaker output (speaker_out) for better linearity

Hardware Features

• High integration (DSP, 3 channels of A/Ds, and 2 channels of D/As, and RAM) and small footprint
• Differential inputs and outputs for microphone input, line input, speaker output, and line output.
• 3 channels of ADCs, 16kHz sampling rate, 78dB SNR.
• 2 channels of DACs, 8kHz or 16kHz sampling rate, 78dB SNR.
• Programmable gain controls for microphone input (mic_in), line input (line_in), line_out, and speaker_out.
• Interface to download data from serial Flash (SST and ST)
• UART or SHI interface to communicate with external host
• Serial Port for digital voice data interface or Bluetooth controller
• 48-pin LQFP package

IC Block Diagram

HyperSampling Technology

• Echo Separation Technology Significantly Enhance Full Duplex
• Microphone Input (16kHz) With Higher Sample Rate Than Line Input (8kHz)

Beam-forming Polar Pattern