Hi @user-f4f9e1 , thanks. Since we already started communication in your Support Ticket in 🛟 troubleshooting , is it alright if we continue there for now?

Hi @user-20283e. University WiFi networks often don't work. You're best bet is to use a dedicated local router. See this message for reference: https://discord.com/channels/285728493612957698/1047111711230009405/1431416744962031627 🙂

Hi @user-8fe6ae! Have you already share your connection from the PC to the phone? If on Windows (Internet Connection Sharing): - Open Control Panel > Network and Internet > Network Connections. - Locate your active internet source (e.g., "Wi-Fi"). - Right-click it and select Properties. - Go to the Sharing tab. - Check the box: "Allow other network users to connect through this computer's Internet connection". - In the dropdown menu below that, select the Ethernet adapter that your phone is plugged into. - Click OK. Your PC will now act as a router for your phone.

Note that exact instructions could vary across versions of Windows.

On Your Phone: - Turn off Wi-Fi and Mobile Data to verify the connection works. - You should see an Ethernet icon ( looks like <...>) in the status bar. The phone will treat this connection just like high-speed Wi-Fi. - Enter the Companion App and navigate to the network tab top-right, and check whether there is an IP shown there.

You can use that IP to connect directly https://pupil-labs.github.io/pl-realtime-api/dev/methods/simple/connect-to-a-device/#connect-directly-via-ip-address

Hi @user-8fe6ae , just building on my colleague's post, you can find detailed steps for Windows 11 here. Please note that Windows 10 does not have the necessary features to support a direct connection, but it will work with a router.

Also, please note that Windows does not always maintain a stable connection when you plug the Ethernet cable directly from Neon into the PC. It sometimes randomly re-assigns the IP address to the phone, breaking the connection. Forcing Windows to give a static IP to the phone should resolve that.

Overall, it is easier and recommended to use a wired connection to a local router:

• Connect the PC via Ethernet to the router
• And connect the Neon via Ethernet to the same router

This will work for all versions of Windows.

You should not lose any appreciable transmission latency and auto-discovery will work. Windows users have reported success with this router out-of-the-box, for wired connections.

Hi @user-bd5142 , may I ask for more clarification about "observation range"?

Hi @user-bd5142 , the viewing distance is in principle unrelated to eye tracking quality. For Pupil Core, it will ultimately depend on the quality of calibration that is done for each participant. It is usually better to calibrate at the distance that they will be viewing, although it can extrapolate outside of that region reasonably well.

For Neon, you do not need to think about this, since it is deep-learning powered and calibration free. Please see this message for an explanation: https://discord.com/channels/285728493612957698/1047111711230009405/1371897435781337200

The tracking quality for our eyetrackers is only "loosely related" to scene camera resolution. For Pupil Core, the calibration targets are usually large with clear centers. Neon's algorithms have been optimized for the chosen scene camera resolution. For both, we have chosen cameras with wide FoVs and high spatial resolution.

To clarify, Neon's FoV covers the human visual field and its images are 1600 px wide by 1200 px tall. This pixel density provides high resolution to the underlying eyetracking models, supporting its out-of-the-box ~1.5-1.8 degrees of gaze accuracy. You can see live examples here. Similarly, the eye cameras each have a 192x192 pixel resolution, giving clear images of the eyes. You can also see examples of Neon's eye images at that page.

That said, Pupil Core can show some reduction in performance when used outside in sunlight. Neon is again much improved and robust against sunlight. It functions equally in pitch black and bright sunlight on the beach. Since Neon is calibration free, you similarly do not need to think about calibration target size, since it is not necessary.

Hi @user-d5a41b ! The Motorola Edge 40 Pro is compatible with any USB-C chargers, but depending on their characteristics it may charge slower.

You can find here the official recommendations from Motorola.

And here you can directly purchase it form Motorola if you prefer https://de.motorolachargers.com/product/turbopower-27w-wall-charger-usb-c/ (if available)

Hi @user-3c26e4 ! This applies regardless of whether the data comes from Neon or Invisible, it’s simply a new metric calculated on the AOIs.

You can have multiple recordings from the same wearer, so both metrics are valid depending on what you want to measure:

If you are interested on how many wearers looked at that area of interest, that would be Wearer reach. If you are interested on how many recordings they looked at that AOI, independently if the same person was recorded multiple times or not, that would be Recording Reach.

What's the added value? E.g. You can see 100% of your participants looked at that area of interest, but looking at the recordings reach, it may drop if they did not look at that AOI on every recording.

Hi @user-0b1519 👋! Thanks for reaching out and for outlining your questions, happy to clarify them.

1) Software Licenses Fees: If you plan to work offline using Neon Player or pl-neon-recording both are free and open-source - there is no fee to use these. Everything needed to record and access your data is included with the hardware, including the Companion app and streaming libraries if needed. Have a look at our ecosystem here.

We do offer additional paid plans for Pupil Cloud, but these are optional.

We can certainly explore what you need over a video call if you prefer.

2) Software updates: Updates for the Companion App, Neon Player and other libraries are included at no extra cost.

3) Training / onboarding: A 30-minute onboarding call is included with every purchase. If you need more in-depth or ongoing training, we can arrange that separately.

4) Ongoing support: General support via email and Discord is included. More advanced / dedicated support packages are also available through consultancy services if needed.

5) Regional restrictions: There are no usage restrictions specific to Germany or the EU. Everything functions the same across regions.

If you'd like a demo and Q&A session, you can book one directly through our website: https://pupil-labs.com/

Hi @user-b5a8d2 👋. Responses below: 1. Yes, that's correct. All of your existing recordings will be available to work with after you've activated an Unlimited Analysis Plan. Note that even now, you can still download your recordings and work with them offline. See this message for reference: https://discord.com/channels/285728493612957698/1047111711230009405/1431119082001793095 2. You just need the activation key to register the Cloud plan. Once purchased, they will receive a confirmation email with the key. You can find more instructions in this section of the docs.

Hi @user-0b1519 - I saw you also reached out via email - I reply here as well, but you'll find more details in the email.

• Yes, Neon has a CE Certificate. I shared it via email.

• All data collected with Neon can be exported as CSV files (see here).

• All AOI-based metrics can be exported when using our analysis platform, Pupil Cloud in CSV files. Pupil Cloud offers powerful gaze mapping tools that allow you to map gaze data from the eye tracking recordings onto screens or areas of interest in the real world environment. You can then extract detailed fixation metrics (e.g., dwell time, number of fixations), and create scanpaths. You can read more about the analysis and visualization tools here.

• Neon always allows manual raw data export via USB. If you plan to use Pupil Cloud, note that with every Neon purchase, you get access to all the features of Pupil Cloud and up to 2 hours of recording storage, such that you can try it out. To unlock unlimited features on Pupil Cloud, we recommend the Unlimited Analysis Plan (see here). As said before, Cloud-based analysis is optional. The raw data are always accessible locally by transferring recordings from the phone to your computer and run your own analysis pipeline with no dependency on Pupil Cloud.

• Neon comes with a 1-year hardware warranty covering manufacturing defects.

• In terms of support, every Neon bundle includes a 30-minute onboarding workshop to help you set up and guide you through our software, but you can always reach us via email or on our Discord server. For more tailored support (e.g., training seminars, custom analysis pipelines), we offer consultancy packages.

Update: I was able to do this by manually setting the playhead and then using "Move event to playhead"

Neon Player will do what you need, and it's meant to work just how you're doing it, but it sounds like there is a problem. Are you able to collect and share the Neon Player log file and/or the recording you're attempting to load?

Hi @user-1baa3e , thanks for the clarification. I moved your question to the 👓 neon channel.

Do I understand correctly that those 18 boxes are the only AOIs presented during the experiment?

Hi @user-7c5b51 , first, please note that we don't recommend University or work networks for connecting to Neon. Even when IT makes Neon accessible that way, the large volume of traffic from other people on the network can interfere with Neon's signal. We typically recommend a local, dedicated WiFi router.

If you want to try it anyway, then you need these ports open to access Neon:

• UDP 5353 (for mDNS)
• TCP 8080 (for HTTP)
• TCP 8086 (for RTSP)
• TCP 8686 (for WebSockets)

Also, note that if your PC and Neon are on different sub-networks (e.g., this is usually the case when the PC is connected via Ethernet to the Uni network and Neon is connected via WiFi), then you might not be able to connect to Neon, even with those ports open. Your IT team might be able to find a solution for that, though.

Hi @user-7c5b51 , do you mean you want to stream the Neon scene camera with gaze overlay to a monitor or TV via a service like Chromecast?

Hi @user-fcdbb6 👋 ! Really cool frame there! Have you already seen the Ready Set Go! frame? Seems similar to what you are trying to achieve.

That said, yes we can share those! We will update the repository next week to include new files 😉

@user-39ee45 Sorry for the delay getting back to you. Could you send us an email to info@pupil-labs.com about this, so we can send you the files?

Hi @user-0cdc2d , may I first ask if you are using Neon or Pupil Invisible?

Hi @user-53b146 , may I ask some clarifying questions first:

• Do you mean you have extracted code from the real-time-screen-gaze package and mixed that with the Surface Tracker plugin of Neon Player?
• The default assumption of the real-time-screen-gaze package is that you are mapping gaze to the surface of the screen (i.e., the monitor boundaries), so it is expected that the coordinates obtained are interpreted relative to said boundaries. What is the ultimate goal you are trying to achieve?

Hi @user-ffef53 , yes, recording is not a pre-requisite for streaming. With the Real-time API, you can

• Start/stop a recording without streaming (& send Events)
• Streaming without recording
• Record and stream simultaneously

including more.

You can even do all of this while using the Lab Streaming Layer integration at the same time, and multiple computers can connect simultaneously to the same device to access all of this functionality.

You can also run Neon Monitor at the same time as recording & streaming, similarly with PsychoPy and other programmatic integrations.

With respect to storage location, you mean you want the data not saved immediately to the phone's harddrive (or not saved at all to the phone?), but immediately saved to the harddrive of some other computer?

Hi @user-910ceb , I moved your posts to the 👓 neon channel. Great to hear it worked out. Yes, you just pull/push the two tabs at the back to adjust the strap. Let us know how it goes and if you have any other questions.

Hi @user-ffef53 , it is not necessary to have cell service to create a hotspot. Many phones create hotspots by re-purposing the WiFi chip in the phone. This approach of running the hotspot on the Companion device has been tried and is used by some, I believe.

Hi @user-a55486 , clock offset is independent of transmission latency, but you are correct that it is the offset between the two clocks. Also, roundtrip time is the amount of time it takes for the Time Echo data to reach Neon and be returned to the client; not necessarily the amount of time it takes for an Event packet to reach Neon. The offset and roundtrip time provided by that function measure two different things and you do not want to subtract or add them to each other. Please see this post and let us know if that clears things up: https://discord.com/channels/285728493612957698/1047111711230009405/1438132613775622244

Regarding the timing discrepancy that you see, it will be helpful to see the snippet of code that manages the offset measurement and Event sending.

Thanks. Trying to time the send_event to happen in correspondence with the flip like that is not guaranteed by PsychoPy, unless you use their callOnFlip method. Could you try wrapping that send_event into a simple function that is passed to callOnFlip as such?

def send_fixation_event_psychopy():
    device.send_event("fixation onset", int(time.time_ns() - offset))

win.callOnFlip(send_fixation_event_psychopy)

...

fix_cross.draw()
win.flip()

I want to know the coordinates on the monitor. The red line in the image was drawn by me to indicate that the monitor position is displayed incorrectly when the code I created is executed.

Hi @user-98b2a9 , yes, this is possible. Please check our MATLAB wrapper for Neon's Real-time API and let us know if you have any questions.

Hi @user-2e5a7e 👋 ! If you do not mind using Python and the Native recording format, you can simply use our library pl-neon-recording .

Simply install it pip install pupil-labs-neon-recording and:

import pupil_labs.neon_recording as nr
recording = nr.open("YOUR_PATH_TO_NATIVE_REC")
print(recording.eye.time[0])

would get you the first eye camera timestamp.

PS: If you have astral's uv , you can open a python shell with the library by using

uv run -p 3.12 --with pupil-labs-neon-recording python and copy the above.

Hi @user-5b6ea3 g 👋
Just to make sure I understand correctly, what you report is that their hands are falling outside the scene camera’s vertical field of view (~80°)?

If that’s the case, one option to extend the FoV, would be to use an additional third-party camera with a wider field of view and map the gaze data, as shown here:
https://docs.pupil-labs.com/alpha-lab/egocentric-video-mapper/

Hi @user-e4d004 👋 ! If you have a Motion Capture system that uses passive reflective markers, you would probably want the frame Every Move You Make or I Can Track Clearly Now depending on whether you want to use prescription lenses or not.

Hi @user-9131cb 👋 ! Do you have your recordings in Cloud?

Okay thanks. I have send you an invite.

Hi @user-e0026b , do I understand correctly that you are using Neon Monitor?

May I ask:

• How long are those recordings?
• What version of the Neon Companion app are you running?

Regarding Pupil Capture, there is only experimental support for using Neon that way and it requires running from source on Linux or Mac. This will also mean that you use Pupil Core's gaze estimation pipeline, which requires calibration. Why would you like to use it this way?

Hi @user-5c56d0. I’m happy to clarify:

Q1: Regarding the "Just act natural" and "Nothing to see here" frames, yes, the only difference is cosmetic. Specifically, it affects how visible the Neon module is to others. For reference, the "Bare Metal" is not the Neon module itself, but the nest PCB, which can be purchased separately if you want to develop your own custom frame. The cosmetic difference between "Just act natural" and "Nothing to see here" can be useful in applications such as consumer research, where researchers want participants to wear a very inconspicuous eye tracker to ensure natural behavior, as if they were wearing normal glasses.

For "I can see clearly now", you are correct. This frame comes with a magnetically swappable lens kit ranging from -3 to +3 in 0.5 diopter steps. This allows researchers to fit the lenses needed for each participant so they can use the eye tracking glasses without wearing their own prescription glasses.

Q2: Yes, all Neon bundles come with an Android smartphone as the Neon Companion device. We are currently shipping the Samsung S25.

Q3: Every Neon bundle already includes the Neon module attached to the frame, meaning the system is ready to use. When purchasing a bundle, you do not need to buy anything extra, as it already includes the frame of your choice, the Neon module, and the Neon Companion device.

If you would like to swap frames depending on participant needs (for example, a "Just act natural" frame for participants who do not need glasses, and an "I can see clearly now" frame for those who do), you can purchase an additional accessory frame - not a full bundle - and simply transfer the module across frames. This is straightforward, as the module attaches to the frame with just two screws.

Hi @user-912183 👋 ! Eye lid aperture, also referred as eye openness in our documentation is part of the eye state data.

Your are pointing to the realtime api documentation, and yes, it is available in real time if you enable "Compute eye state" from versions +2.9.0 or above of the Companion App.

But it seems you want to access the data post recording instead, and using the native recording rather than the Timeseries, is that correct? If so, you would probably want to access the eye state data using pl-neon-recording as shown here https://github.com/pupil-labs/pl-neon-recording/blob/main/examples/csv_export.py#L186-L217

Hi @user-743e57 👋. Both Pupil Cloud and Neon Player allow you to process and analyse data recorded with our Neon eye tracking system. Both enable you to replay recordings and export data streams, such as fixations, saccades, and video overlays, to name a few.

For a full breakdown of analysis tools and data formats, please refer to our online documentation: - Pupil Cloud - Neon Player

That said, Pupil Cloud does offer advanced capabilities not available in Neon Player, including aggregate AOI metrics, workspace sharing for team collaboration, and powerful computer vision algorithms like the Reference Image Mapper. It also has optional anonymisation features, like face blurring.

Do you have a specific use case in mind or questions about any of these? We’d be happy to arrange a demo and Q&A session to walk you through both platforms and determine the best fit for your project. Just let me know!

Hi @user-b4808a 👋. If you want to work with those recordings in Pupil Cloud, you can either purchase an Unlimited Analysis Plan, or download and delete recordings from the free 2-hour quota to free up space. Recordings will become available in turn. If you choose to work with your recordings offline, we have free workflows. For example, you can use Neon Player, our desktop software. Just drag and drop them onto the software to playback and process them. We also have a Python recording library if you want to work with the recordings programmatically.