Question
Answer (43)
Ticket Type
🐛 Bug Report
Robot Type
Lite
Operating System
macOS
Browsers Used
- [ ] Chrome
- [ ] Firefox
- [ ] Edge
- [ ] Safari
- [ ] Opera
- [ ] Other
Browser version(s)
No response
Troubleshooting
- [x] I have read https://huggingface.co/docs/reachy_mini/troubleshooting
Update Status
- [x] My robot is up-to-date via the dashboard.
- [x] I'm running the latest Reachy Mini release from this repo.
Description
I have noticed that the gstreamer backend yields really dark images on macOS. It appears that auto-exposure may not be enabled and by default a really low exposure time is used. The really dark images make it really hard for detectors to work and even for humans to see what is in an image.
Context & Reproduction
- Start reachy mini
- run a script which pulls camera images off the reachy
- display camera images
Relevant logs or stack trace
Additional Info / Workarounds
Can open FaceTime after starting the reachy. This seems to force the system to adjust exposure time.
Answer:
It seems that the gstreamer video stream is excessively dark when using macOS. This issue might be due to the absence of auto-exposure or a low default exposure time. The dark images make it difficult for detectors to function and for humans to interpret the content.
Steps to reproduce:
- Start the Reachy Mini.
- Run a script that fetches camera images from the Reachy Mini.
- Display the camera images.
Possible workaround: You can try opening FaceTime after starting the Reachy Mini. This might force the system to adjust the exposure time and improve the image brightness.
If the issue persists, you may want to explore the following options:
- Manually adjust exposure settings in your script or application.
- Check for any macOS-specific configuration issues or updates.
- Switch to a different video streaming backend such as OpenCV or another supported library.
If none of the above solutions work, please submit a new ticket with more details and relevant logs for further investigation.