hi, I have been working on DIY projects recently. now I can see my eyes on the screen with eye-camera by Pupil Capture, however, the both welded IR-led （4050-z）seem DO NOT work well on Microsoft-HD6000 I have not detected any 850nm IR light when the camera is working , which really confused me because i have followed all the procedures from buying products to welding IR-led.

I'm not sure if this is the right place to ask, but in the cloud-service, can I move recordings between work spaces? if I were to create a new one, could I move an existing recording from one workspace to another?

thank you very much for the quick info! I have to say, you're one of the most helpful companies out there, truly.

Hello, we are using the invisible glasses for recordings. But we are having problems downloading the videos in the program. We can download videos but then you can't see the markers of fixations anymore. Is there a special setting we have to think of? Thank you in advance!

Hello, is it possible to trim videos directly in Pupil Cloud in order to create an "enrichment" using only certain portions? I noticed that there is a video duration limit (3 minutes) for processing.. thanks!

I mean downloading the videos from Pupil Cloud:).

Thank you, I can use the video renderer but can't save the video in the next step. I can't click on it (it's grey).

Hi, I am using the Neon glasses and real-time API to try and create an android app to visualize/record coordinate and IMU data. I and trying to deploy python script directly on the Kotlin app, but there is a version conflict with pydantic and chaquopy—chaquopy doesn't support pydantic v2.x / pydantic core. But it can support pydantic v1.x, so I am wondering if there is a version of the real-time API that is compatible with Neon and pydantic v1.x

@user-d801e5 here stepping in for @user-f43a29

The use of pydantic on the API is minimal, mainly for the templates feature, so if you want to keep using chaquopy you can fork it and modify the models.py

Thanks for the response, I will follow up if I continue to run into problems!

I'll also take a suggestion here: For now, I want to make a real-time data viz mobile application to be run on the companion device that displays gaze and IMU data as the Neon is being used. Right now, I have a setup that relies on a LAN and python client on my desktop. But I want to create a version that's independent and can run with just the companion device. I'm leaning to forking the SDK since it might be valuable to use python if the project becomes more complicated.

Yes, I invited you in our workspace. Thank you!

Hello,

Currently, I’m attempting to implement a method of streaming the video feed from the Neon glasses over long distances between two devices. Currently, the application I’m working on has support for local communication already using the Neon API.

I was able to use ZeroTier, a service that allows long-distance communication between devices, to assign a new IP address to the phone using the Neon companion app, as well as the laptop running the app I’m working on, and they are able to communicate the streaming data between each other when they are both on the same network using the local HTTP port 8080 with these new IP addresses. This falls apart, however, when the devices are on two separate networks, where they are unable to connect with each other. My concern is that the Neon companion app is broadcasting the connection request on the specific local port, and the laptop that is attempting to connect on a different network is unable to access this local port, which is required when using the Neon API to manually set up a device without using the discover_devices function.

Does anyone know if it is possible to allow this local port to be accessed remotely, and if not, is there any known alternative way to go about this that is supported by Pupil Labs?

Hi @user-c29368 ! May I ask is Zero Tier something a like a VPN?

Hi guys, i am using (out of an emergency) a script with different Eye Trackers. One is the Neon Tracker and one the Invisible. I am tracking eye movements of two people playing the piano.

Device Serial for Neon: 321970 (Phone Name: Neon Companion - Companion Device: rtwo_g)

Device Serial for Invisible: qc49m (One Plus 8T)

My python script connects via port 8080 (for both) and different IP adresses to the respective eye tracker. Clicking on space i start recording and clicking space again i stop the recording. In some cases i have no problems. In other cases i get the issue that Neon does not stop and i get the error: AttributeError: 'tuple' object has no attribute 'tb_frame' regarding „device2.recording_stop_and_save()“. I need to stop the script manually then by going to the phone and pressing stop. This is not only inconvenient but problematic since i disrupt the experimental processes.

Another issue that happens very often is the invisible simply does not record. I sometimes get a „400 already recording“ in python. Thus the script is not throwing an error. I opened the real time API with pi and neon (8080) port to monitor. If i manually click on record for invisible it repeats "already running!" but on the phone there is no sign to indicate recording. I have to restart the program then. For every run i restart the kernel to be sure. It worked for example well yesterday. But todays recording were working absoluetly horrendous. I am not sure how much its my scripts fault and how much its a network issue or a network card issue from my laptop running the script. I'll send you my script attached.

Thank you! I will try it out

Oddly this issue occured again. Less often than before but i got the "already running"

for the invisible only however