v2.3.0

2.3.0

• CHANGED: Updated Cyberon Wake Word engine.
• CHANGED: Updated XTC Tools to 15.3.0.
• CHANGED: Updated submodule fwk_io to version v3.6.0 from v3.3.0.
• CHANGED: Updated submodule fwk_core to version v1.1.0 from v1.0.2.
• CHANGED: Updated submodule fwk_voice to version v0.8.0 from v0.7.0.
• CHANGED: Updated submodule fwk_rtos to version 3.2.0 from 3.0.5.
• CHANGED: Updated submodule lib_src to version 2.7.0 from 2.4.0.
• CHANGED: Updated submodule xscope_fileio to version 1.2.0 from 1.1.2.
• ADDED: lib_sw_pll submodule v2.3.1.
• ADDED: xmos-ai-tools v1.3.1 Python requirement.
• REMOVED: Deleted inferencing submodule.
• ADDED: FFVA INT example with Cyberon speech recognition engine and model
  (DSpotter v2.2.18.0).
• CHANGED: Moved files in folders device_memory, gpio_ctrl, intent_engine and
  intent_handler from examples/ffd/src to folder modules/asr.
• CHANGED: Remove need to use external MCLK in FFVA INT examples
• ADDED: Support for DFU over I2C for FFVA INT example.
• ADDED: FFD example with I2S audio input to Cyberon speech recognition
  engine and model.
• REMOVED: flash settings in .xn files, as they are not required by XMOS
  Tools 15.3.0.
• ADDED: Support for reading registers over I2C slave in FFD examples.
• ADDED: Note in ASRC demo documentation about large latency in ASRC
  processing. References to alternative application notes have been provided.

v2.2.0

v2.2.0

CHANGED: Updated submodule fwk_io to version 3.3.0 from 3.1.0.
CHANGED: Updated submodule fwk_core to version 1.0.2 from 1.0.0.
CHANGED: Updated submodule fwk_rtos to version 3.0.5 from 3.0.3.
CHANGED: Updated submodule fwk_voice to version 0.7.0 from 0.6.0.
CHANGED: Updated submodule lib_qspi_fast_read to version 1.0.2 from 1.0.1.
CHANGED: Updated submodule lib_src to version 2.4.0 from 2.2.0.
CHANGED: Updated submodule xscope_fileio to version 1.1.2 from 1.1.1.
ADDED: Support for Microsoft OS 2.0 Descriptors for FFVA and ASRC examples so the DFU interface automatically shows up as a WinUSB interface on Windows.
ADDED: FFD example with Cyberon speech recognition engine and model (DSpotter v2.2.18.0).

v2.1.0 (unreleased)

ADDED: Mic aggregator app that bridges between 16 mics and TDM16 slave or USB Audio
ADDED: Asynchronous Sampling Rate Converter (ASRC) example application
ADDED: lib_xua 3.5.1 as submodule to support Mic aggregator app
CHANGED: Updated submodule fwk_io on to version 3.1.0 from version 3.0.1 to add support for TDM16 slave tx and 16ch mic_array

v2.0.0

2.0.0

ADDED: All speech recognition example deigns now use Sensory TrulyHandsfree speech recognition library.
ADDED: Low-power far-field dictionary (low-power FFD) example design using wakeword to exit standby mode.
ADDED: Mandarin model to far-field dictionary (low-power FFD) example design demonstration.
ADDED: Support for fast flash library which increases flash reading throughput by as much as 70%.
MOVED: Audio pipelines relocated to modules to allow easier re-use.
FIXED: Example designs now indicate when evaluation period has expired.
FIXED: Numerous minor bug fixes.

v1.0.0

1.0.0

ADDED: Improved documentation
ADDED: Speech recognition porting example design
FIXED: USB PIDs changed to 0x4000 for FFVA and 0x4001 for FFD
FIXED: Speech recognition model no longer in the filesystem
FIXED: Several Linux USB audio bugs
REMOVED: audiomux example design

v0.21.0-beta.0

0.21.0-beta.0

• ADDED: Improved Interference Cancellation using Voice-to-Noise Ratio estimator.
• ADDED: Low-power mode to far-field dictionary example design
• ADDED: USB DFU to far-field voice assistant example design
• ADDED: Optional AEC fixed-delay to far-field voice assistant example design

v0.20.0

0.20.0

• ADDED: Audio response playback
• ADDED: Wanson 20220923 model update
• ADDED: Documentation updates
• FIXED: USB audio device is now recognized on Windows
• FIXED: Applications moved to examples folder
• FIXED: Sample rate conversion from 48k fixed

v0.12.0

v0.12.0

• ADDED: Audio playback to FFD application
• ADDED: First draft of documentation
• KNOWN ISSUE: USB audio device is not recognized on Windows. This Does not effect FFD.

v0.10.0

v0.10.0

FFD demo using OLED display