Product
Datasheet

A single hardware variant is available. USB audio output is provided only as a 12.1-channel stream: 12 microphone channels plus one dedicated low-frequency channel.

When SYNC is active, the last channel contains a one-sample negative pulse at each SYNC rising edge. MASTER emits a 248 ms code cycle: 20 ms × 10, then 24 ms × 2 between pulse starts. The SDK detects the pulse from its one-sample drop-and-recovery shape; the threshold is scaled from the volume reported by the audio endpoint.

Highspeed Mode: The 12.1-channel USB audio stream remains available, but onboard SSL results are not provided.

The array uses the same 12-microphone hardware configuration in both suffix variants. The I/N suffix identifies infrasound support.

Coordinates in millimeters; all Z coordinates are 0.

Direction estimation: All microphone elements are located at Z = 0, so the built-in SSL cannot distinguish directions mirrored across the microphone plane. Full-sphere direction discrimination is not possible with one MIC1223. It requires multiple MIC1223 arrays synchronized through SYNC and processed with SDK SSL.

8 ring LEDs (7 red + 1 yellow at 0° reference) visualize the built-in hemispherical SSL output in real time. When multiple sources are detected, multiple LEDs illuminate simultaneously. Elevation magnitude is encoded in LED brightness. Full-sphere direction discrimination requires multiple arrays connected through SYNC and SDK SSL. Each LED can be manually controlled (brightness 0–7) via l command on the CDC COM port; setting brightness to 8 restores firmware SSL output.

Green and blue status LEDs are positioned at the top of the device to indicate current power and streaming status. The ring LEDs are user-controllable via CDC COM port — brightness and pattern can be overridden independently of SSL output.

A single MIC1223 provides hemispherical direction estimation only. Full-sphere direction discrimination requires multiple MIC1223 arrays synchronized through SYNC and processed with SDK SSL.

The on-board SSL engine runs within the device MCU's real-time budget, which bounds algorithm complexity. Angular error, multi-source separation, and noise robustness are all constrained by available on-chip processing capacity. The result is a practical estimate suitable for general use, not a high-precision measurement.

For higher-accuracy localization, users can capture the raw multichannel audio and run their own SSL algorithm on the host. Multiple arrays can be hardware time-aligned via SYNC to further improve accuracy. See Section 6-3 for host-side SSL setup and recommended multi-array configuration.

To obtain accurate SSL results, the direct audio propagation path between the sound source and the array must be free of obstacles. Physical obstruction — mounting brackets, enclosure walls, or nearby structures — attenuates or reflects sound before it reaches the array and degrades SSL accuracy.

Real-time tracking data is provided via USB CDC as a virtual COM port. Since this is a virtual COM port, the baud rate setting is ignored — the host may set any baud rate. The port uses 8 data bits with parity (8N1 or as configured by the host). Up to 10 sources can be tracked simultaneously (3 sources are practical for standard operation). If multiple sound sources are detected in a single frame, the device transmits multiple lines (one string per source) sequentially.

Status output reports device runtime information via the CDC COM port when enabled. The temperature measurement point is the MCU internal temperature.

GPIO samples are transmitted on the 8 Hz CDC output cycle. When the GPIO update rate is set above 8 Hz (e.g. 100 Hz), multiple samples accumulate per cycle and are sent as a burst — for example, at 100 Hz up to approximately 13 samples may be batched in a single transmission.

The host PC may send configuration commands to the device via the CDC COM port. Commands take effect immediately in RAM. To persist settings across power cycles, send s (save). Send r to restore factory defaults.

Multi-microphone ID: When multiple MIC1223 units are used together, each unit can be distinguished by its device ID. Set the ID through the CDC i command; the configured ID is included in the device output and may also appear in the USB device name, for example 12CH MICARRAY ID0. Some operating systems cache USB device names and may not refresh the displayed name immediately after the ID is changed. If the displayed USB device name does not change after setting the ID, reset or clear the host USB device configuration/cache and re-enumerate the device.

Persistence: Commands take effect immediately but are not written to flash until s is sent. On power-up, the last saved flash state is restored. r resets to factory defaults and saves automatically.

The SYNC pin (Pin 1 of the 2.54 mm header) normally carries the inter-microphone synchronization pulse. When SYNC synchronization is not required, the same pin can be reconfigured as a general-purpose GPIO. GPIO mode and SYNC mode are mutually exclusive: activating GPIO mode disables SYNC output regardless of the CDC f-command setting (treated as f0).

Each MIC1223 provides a dedicated SYNC I/O pin header (2.54 mm, 2-pin). Before recording, configure exactly one connected unit as MASTER (f2) and every other unit as FOLLOWER (f1). The firmware does not perform automatic master election. When active, the last channel of each recording stream contains a one-sample negative pulse at each SYNC rising edge; full 16-bit audio resolution is maintained at all other times.

The 248 ms SYNC code cycle uses one-sample negative-pulse markers at each rising edge and provides a common reference point across all units. The unique 20 ms -> special marker -> 24 ms cadence is used only for initial FIFO alignment, independently for every MASTER–FOLLOWER pair. After lock, the SDK updates each pair from its next common marker without distinguishing special and ordinary markers; a missing marker leaves that pair's offset unchanged, and only that pair returns to initial-lock search after three seconds without an update. MASTER does not insert a marker merely because firmware writes the output: its driven PA11 rising edge is observed through the same EXTI interrupt path used by FOLLOWER units. The SDK recognizes only a one-sample drop followed by recovery, using a per-endpoint threshold scaled from the current audio volume. Marker-corrupted samples can be preserved or interpolated by the host SDK output options.

Connecting MIC1223 arrays via SYNC hardware-aligns their audio streams to a single time reference. The same SYNC connector is shared by CAM0222GT thermal cameras and CAM0222GS stereo cameras, enabling heterogeneous synchronization of microphone audio and camera frames. Each stream retains its native sample or frame rate; timestamps associate audio samples and video frames with the common time reference. Host-side SSL can then treat microphones across units as one large-aperture array. This approach is not limited by on-chip processing capacity — any algorithm can be applied on the host.

Multiple-Array SDK SSL:
Use multiple MIC1223 arrays connected through SYNC; select the 12I or 12N suffix for infrasound support.
Each unit has the same 12-microphone hardware configuration. Connect multiple units through SYNC and use SDK SSL for spherical direction estimation.

A browser-based test dashboard is provided for MIC1223 operation with the CAM0412 and CAM0510 fisheye-lens camera options. The dashboard overlays the MIC1223 SSL direction on the fisheye camera view and provides Web Serial / Web Audio / WebRTC controls for evaluation and application prototyping.

All MIC1223 units are individually calibrated before shipment to ensure consistent performance across the microphone array. The calibration data is stored on-device and applied automatically during operation.

For sustained SSL accuracy in long-term deployments, recalibration is recommended after the 2-year warranty period. The manufacturer provides a recalibration service to restore factory-level performance.