Project Aria Gen 2 Profiles

Project Aria Gen 2 glasses offer a range of profiles, allowing users to select the optimal configuration of sensors and settings for their research or application needs. Each profile defines a specific combination of sensor streams, data rates, resolutions, and encoding formats. These profiles are designed to balance data quality, device performance, and operational stability. Profiles are divided into two categories:

Pre-defined profiles: Thoroughly tested, stable, and fully supported. Recommended for most users.
Custom profiles: User-defined configurations for advanced or experimental use. These may not be stable and are not recommended for general use. A custom profile example is provided below.

The table below details the specifications of the currently supported pre-defined profiles. New profiles may be added as the platform evolves.

Feature / Sensor	profile8	profile9	profile10	mp_streaming_demo
Description	General purpose recording profile	General purpose streaming profile	Recording profile with RGB 3MP 30Hz	Machine perception streaming demo (streaming only)
IMUs (Hz)	800	800	800	800
Magnetometer (Hz)	100	100	100	❌
Barometer (Hz)	50	50	50	❌
Audio Sample Rate (Hz)	16K	16K	16K	16K
Audio Encoding Format	OPUS	OPUS	OPUS	OPUS
Audio Bitrate	256000	256000	256000	256000
GPS (Hz)	1	❌	1	❌
BLE Period (ms)	30000	❌	30000	❌
BLE Duration (ms)	2000	❌	2000	❌
WiFi Period (ms)	30000	❌	30000	❌
WiFi Scan Mode	ACTIVE	❌	ACTIVE	❌
CV Cameras (Hz)	30	10	30	10
CV Resolution	512x512	512x512	512x512	512x512
CV Encoding	H265 (CQP QP=22)	H265 (CBR 900kbps)	H265 (CQP QP=22)	H265 (CQP QP=30)
RGB Camera (Hz)	10	5	30	10
RGB Resolution	2560x1920	2560x1920	2016x1512	2016x1512
RGB Encoding	H265 (CQP QP=22)	H265 (CBR 8Mbps), blur filter	H265 (CQP QP=22)	H265 (CBR 1.2Mbps), blur filter
ET Cameras (Hz)	5	❌	5	5
ET Resolution	200x200	❌	200x200	200x200
ET Encoding	H265 (CQP QP=22)	❌	H265 (CQP QP=22)	H265 (CQP QP=30)
ET (Hz)	30	30	30	30
HT (Hz)	30	30	30	10
VIO (Hz)	10	10	10	10
Device Info Period (ms)	30000	10000	30000	10000
PPG (Hz)	128	128	128	128
Temperature (Hz)	1	1	1	1
ALS (Hz)	9.434	9.434	10	❌
ALS Exposure (us)	3200	3200	3200	❌

Profile Usage Recommendations

When to Use Pre-defined Profiles

Pre-defined profiles (profile8, profile9, profile10, mp_streaming_demo) are:

✅ Thoroughly tested and validated
✅ Optimized for device stability and thermal performance
✅ Fully supported by all downstream tools
✅ Guaranteed to work reliably
✅ Recommended for production use and research studies

Recommended for:

Production deployments
Large-scale data collection
Research studies requiring reliable data
Users who want maximum compatibility
First-time Aria users

When to Consider Custom Profiles

Custom profiles may be appropriate for:

Advanced research with specific sensor requirements
Prototyping new applications
Experimental configurations
Specialized use cases not covered by pre-defined profiles

⚠️ Important: Custom profiles should only be used by advanced users who understand the risks and limitations.

Custom Profiles

Overview

Custom profiles allow advanced users to define their own sensor configurations for specialized research needs. However, custom profiles come with significant risks and limitations and should be used with caution.

⚠️ Important Considerations and Risks

Before creating a custom profile, understand these critical limitations:

1. Device Stability and Data Quality

Thermal and Power Issues:

Custom profiles that push sensor settings (e.g., high frame rates, high resolutions, enabling many sensors simultaneously) can cause the device to overheat or drain the battery rapidly
This can lead to unexpected shutdowns or reduced recording/streaming times
Some combinations may trigger warnings in the Aria Companion App, but not all edge cases are covered automatically
Using extreme values or not basing your custom profile on a tested template significantly increases these risks

Data Quality Risks:

Unusual sensor configurations may result in:
- Poor data quality
- Synchronization issues between sensors
- Missing data streams
- Inconsistent sensor performance
These issues may not be immediately apparent and could compromise your research data

Fragmented Datasets:

Custom profiles can lead to fragmented datasets, making it harder to:
- Compare data across different studies
- Aggregate data from multiple users
- Reproduce research results
This reduces the overall utility of the data for broader research purposes

2. Compatibility and Downstream Tooling

Tooling Incompatibility:

Downstream tools (e.g., Project Aria Tools, MPS, analytics pipelines) may not be fully compatible with all custom profile configurations
Data recorded with custom profiles may:
- Not be processed correctly by standard tools
- Require additional engineering work to support new formats
- Be incompatible with existing analysis pipelines

Lack of Standardization:

Custom profiles are not as thoroughly tested or supported as pre-defined profiles
This can lead to:
- Unexpected issues when using standard tools
- Difficulties sharing data with collaborators
- Limited community support

3. Known Bugs and Open Issues

Sensor Data Retrieval:

There have been reports of issues retrieving certain sensor data (e.g., GPS) when using custom profiles
These issues may not be present with recommended pre-defined profiles

Parameter Guardrails:

While the Companion App includes some guardrails, not all problematic combinations are blocked
Users may still encounter hardware or data quality issues when defining profiles via the SDK

⚠️ Use at Your Own Risk: Custom profiles are provided as-is for advanced users. Meta does not guarantee stability, data quality, or tool compatibility for custom profile configurations.

Custom Profile Template

Below is a template for creating custom profiles. It is strongly recommended to base your custom profile on one of the pre-defined profiles (e.g., profile8) rather than starting from scratch.

This example is based on profile8 with the following modifications:

RGB frame rate: Reduced from 10Hz to 5Hz
SLAM camera frame rate: Reduced from 30Hz to 10Hz
Hand tracking: Reduced from 30Hz to 10Hz

{
   "name": "custom_profile",
   "description": "Custom profile based on profile8: Reduced RGB frame rate from 10Hz to 5Hz. Reduced SLAM frame rate from 30Hz to 10Hz. Reduced HT from 30Hz to 10Hz",
   "imus": {
    "rate_hz": 800
   },
   "magnetometer": {
    "rate_hz": 100
   },
   "barometer": {
    "rate_hz": 50
   },
   "audio": {
    "sample_rate": "R_16KHZ",
    "frame_period": "P_20MS",
    "sample_format": "S16",
    "encoding_format": "OPUS",
    "opus_config": {
     "complexity": 10,
     "bitrate": 256000
    }
   },
   "gps": {
    "rate_hz": 1
   },
   "ble": {
    "period_ms": 30000,
    "duration_ms": 2000
   },
   "wifi": {
    "period_ms": 30000,
    "scan_mode": "ACTIVE"
   },
   "slam_cameras": {
    "rate_hz": 10,
    "auto_exposure": {},
    "encoding": "H265",
    "video_config": {
     "cqp": {
      "qp": 22
     }
    }
   },
   "et_cameras": {
    "rate_hz": 5,
    "auto_exposure": {},
    "encoding": "H265",
    "video_config": {
     "cqp": {
      "qp": 22
     }
    },
    "resolution": "R_200x200",
    "ir_led": "TRUE"
   },
   "rgb_camera": {
    "rate_hz": 5,
    "auto_exposure": {},
    "encoding": "H265",
    "video_config": {
     "cqp": {
      "qp": 22
     }
    },
    "resolution": "R_4032x3024",
    "width": 2560,
    "height": 1920
   },
   "et": {
    "rate_hz": 30
   },
   "ht": {
    "rate_hz": 10
   },
   "vio": {
    "rate_hz": 10
   },
   "device_info": {
    "period_ms": 30000
   },
   "ppg": {
    "rate_hz": 128
   },
   "temperature": {
    "rate_hz": 1
   },
   "als": {
    "rate_hz": 9.434,
    "exposure_time_us": 3200
   },
   "vio_high_frequency_pose": {
    "rate_hz": 800
   },
   "tags": [
    ""
   ]
  }

Template Notes

This example demonstrates a conservative approach to customization: making minimal changes to a tested profile (profile8) to reduce power consumption while maintaining core functionality.

Best Practices for Custom Profiles

If you must use a custom profile, follow these guidelines:

1. Start from a Pre-defined Profile

✅ Base your custom profile on profile8, profile9, or profile10
✅ Make incremental changes rather than dramatic modifications
❌ Avoid starting with a blank profile or extreme parameter values

2. Test Thoroughly

Test the profile with short recordings first
Monitor device temperature during use
Check battery consumption rates
Verify all expected sensor streams are present
Validate data quality before large-scale data collection

3. Monitor Device Health

Watch for thermal warnings in the Companion App
Note any unexpected shutdowns or errors
Track recording duration vs. battery life
Monitor data completeness and quality

4. Conservative Parameter Selection

Frame rates: Lower rates reduce power and thermal load
Resolution: Higher resolutions increase data size and processing
Encoding quality: Lower QP values (higher quality) increase bitrate
Sensor combinations: Enabling all sensors simultaneously increases load

5. Validate Data Compatibility

Test with Project Aria Tools to ensure data can be loaded
Verify compatibility with your analysis pipeline
Check that all required sensor streams are accessible
Consider data sharing and reproducibility needs

6. Document Your Configuration

Keep detailed records of custom profile parameters
Document the rationale for each modification
Note any issues or limitations encountered
Share configuration details with collaborators

Parameter Guidelines

When modifying parameters, consider these general guidelines:

Parameter Type	Considerations
Frame Rates	Higher rates = more data, higher power consumption, thermal load
Resolution	Higher resolution = larger files, more processing, better quality
Encoding Quality (QP)	Lower QP (e.g., 18) = higher quality, larger files; Higher QP (e.g., 30) = lower quality, smaller files
Sensor Combinations	More active sensors = higher power draw, more thermal stress
Bitrate	Fixed bitrate (CBR) for consistent file sizes; Variable rate (CQP) for consistent quality

Using Custom Profiles

Custom profiles can be loaded via the Client SDK. See the Client SDK documentation for detailed API usage and examples.

Support and Troubleshooting

For Pre-defined Profiles

Full support available through standard channels
Issues will be investigated and resolved
Documentation is comprehensive and maintained

For Custom Profiles

⚠️ Limited support available
Community forum may provide assistance
Users are responsible for validating their configurations
No guarantee of compatibility with tools or services

Getting Help

For questions about profiles:

Pre-defined profiles: Contact support via support channels
Custom profiles: Check the community forum and known issues

Profile Usage Recommendations​

When to Use Pre-defined Profiles​

When to Consider Custom Profiles​

Custom Profiles​

Overview​

⚠️ Important Considerations and Risks​

1. Device Stability and Data Quality​

2. Compatibility and Downstream Tooling​

3. Known Bugs and Open Issues​

Custom Profile Template​

Template Notes​

Best Practices for Custom Profiles​

1. Start from a Pre-defined Profile​

2. Test Thoroughly​

3. Monitor Device Health​

4. Conservative Parameter Selection​

5. Validate Data Compatibility​

6. Document Your Configuration​

Parameter Guidelines​

Using Custom Profiles​

Support and Troubleshooting​

For Pre-defined Profiles​

For Custom Profiles​

Getting Help​