@user-15edb3 , I'm replying to your message (https://discord.com/channels/285728493612957698/285728493612957698/1191684449432244305) here. Can you please elaborate on your research question and planned analysis?

Thanks....hi..So i have this wearable glasses and i have been collecting data earlier by uploading the data in cloud and downloading.In that the gaze positions were not normalised and was in absolute values so i get gaze values like (700,400),(600,500) etc. But recently i found that in offline mode data collection the frame number also is getting recorded. However the gaze data ie, norm_x and norm_y are normalised values. However for my algorithm i need absolute values. Basically this new data is just scaled and reference is changed .But from where do i find that scaling factor.

that seem to work partially but still i doubt if it is accurately matching with the same values as it was giving previously unscaled.or am i doing something wrong. I multiplied norm_x with 1600 and multiplied norm_y with 1200 and substacted from 1200(siince reference is tope left for unscaled and ref is bottom left for scaled ones). i feel 1600 is not the exact scalar value to be multiplied.this seemed neglible discripency but my aoi detection algorithm doest seem to work now.kindly help

that worked. i hd some other issue in code. thanks for ur response

Hello, I am working on pupillometry with the NEON device, and in the raw data, I do not see any distinction made between the right and left eye. Similarly, in the online documentation, there is no clarification. Could you tell me if this is a randomly chosen eye or an average of the diameters of both pupils ? Thank you !

Thank you for the response, however, the diameters of the pupils are not the same between right and left to within 1 mm, which is huge when the variations related to cognition do not exceed 0.5 or 0.6 mm... The relative variations in pupil diameter are the same, however, it's true. It would be interesting in the future to actually have the data for each eye separated. Have a good day

Hello,

I am trying to connect our neon to LSL. I have installed the lsl-relay but it cannot find any devices. I was able to get the stream to my browser (even though the URL did not work, I used the IP) so I am in the same network. But I am also not the biggest python pro so might have made a mistake there. I have installed lsl-relay package in pycharm and then used the the pupil_invisble_lsl_relay.exe.

Alright I kind of fixed it by connecting directly using the IP.

@user-cdcab0 Thank you. Now, in a test run the gaze tracking worked. Is there a way to get the pupil size data directly or do I have to first upload the video to the cloud and then download the pupil size?

Are there plans to get it in real-time or at least locally on the phone?

Hello, When I am trying to switch workspace or add a wearer. It's showing like this. What should I do ?

Hi @user-b6f43d ! could you please try logging out of the app and back in?

Hi again @user-b327ef👋🏽 ! Regarding your question here: https://discord.com/channels/285728493612957698/285728493612957698/1194271538153783337, Neon's pupil-size measurements are provided as a single value, reported in millimetres, quantifying the diameter of the 3D opening in the centre of the iris. Currently, our approach assumes that the left and right pupil diameters are the same. According to physiological data, this is a good approximation for about 90% of wearers.

Note, however, that we do not assume that pupil images necessarily appear the same in the left and right eye images, as apparent 2D pupil size strongly depends on gaze angle. We are working on an approach that will robustly provide independent pupil sizes in the future. Unfortunately, we can not offer a concrete timeline for completion.

Hi there Pupil labs, we received our 2 neon ready set go trackers just this week, and have been testing it. However, 1 of the 2 trackers is not working properly. It will not always connect to the companion properly, it stays stuck with the white led on. At other times it will connect (and the app keeps asking the permissions each time), but then, recordings will just stop randomly. I'll send a photo of the state when not connecting to the phone.

We've tested both companions with both neon devices, and found that the problem is not related to the companion phones. These both work fine with the other neon tracker.

What could the issue of the neon tracker be? should we send it off for a replacement perhaps? thanks for your support!

My colleague also just informed me that someone from our institute (Max Planck institute for Psycholinguistics) will be travelling to Berlin tomorrow, so perhaps the tracker can be dropped off at Pupil Labs for investigation/repair?

hello so i have been using neon player to download the data offline.It does take long time for my large data.However am only interested in some specific excel files alone instead of the entire thing.Is there any chance to customise and speed up my process? Precisely i am in need of gaze position csv file which i get after drag drop,then click download symbol which creates folder like neonplayer/export/000/gaze_position.

Hi all. Trying to Define AOIs and calculate gaze metrics using the guide on Alpha Lab

When running #normalize patch values, I get the error "NameError: name 'aoi_values' is not defined"

This is the guide I'm using: https://docs.pupil-labs.com/alpha-lab/gaze-metrics-in-aois/#define-aois-and-calculate-gaze-metrics

Hi, I noticed that the video from the world camera has a lot of motion blur (good illuminated interior). How can I reduce such blur? Is it possible to set the exposure manually?

I suspect that's where I'm going wrong (I'm brand new to Python!)

Learning as I go along

hello there! does someone have a pdf or any other document in which there is written how the detection and computation of blinks and pupil size work with the pupil neon? I looked on the homepage and couldn´t find any. Wish you guys a great weekend 🙂

@user-c1bd23 We should probably have added typing annotations 😅 .

If you are new to Python, my recommendation is that you use the vanilla script I mentioned before. Simply save the script, and if you have the dependencies installed in your environment, it should be pretty straightforward to run it. https://discord.com/channels/285728493612957698/1047111711230009405/1195008075036381276

It is also somehow commented.

That said, in the tutorial (which is due to be updated soon 😉 ), the aoi_ids are used as aoi_values to show the AOIs selected first.

aoi_ids = pd.Series(np.arange(len(aois)))
plot_color_patches(reference_image, aois, aoi_ids, plt.gca())

Where would this be fed into in order to generate an overlay of AOIs on the reference image?

No. I only got a prompt to select the AOIs on the reference image

Once complete, it exited python and generated the aoi_ids file

Hi Pupil Labs team, Despite meticulous calibration efforts, we've noticed a systematic shift that requires us to individually adjust the areas of interest for each subject (The image shows the plotted gaze data using an R script). We're reaching out to see if there are any algorithms or tips available to retrospectively correct these shifts, eliminating the need for manual adjustments based on visual judgment. Any guidance you can provide on this matter would be immensely appreciated. Thank you in advance!

Hey the NEON was never sampled at 250Hz right ? Always 200Hz after uploading it to the cloud ?

Thanks much

Offset correction

Hello Pupil Labs team, I recently made three recordings that resulted in three different errors:

Recording 1: December 20th, 2023 The recording was supposed to last about 40 minutes. Afterwards it turned out that the recording stopped after about 15 minutes. The following error message was displayed: "Calibration not synced! Tap for more instructions." However, I couldn't press it because an FPGA update was supposed to be performed. However, the update hung up because the connection to the eye tracker was interrupted. After several attempts, I was finally able to carry out the update. The aforementioned error message has not occurred since then, but the recording is now no longer usable for me.

Recording 2: January 10th, 2024 The same companion device with different neon glasses was used for this recording. This time, no error message was displayed, and the recording duration is also stated as 43 minutes on the companion device. However, when I start the recording, the video is only 8.5 minutes long. In the Pupil Cloud, on the other hand, the recording is 43 minutes long, but from minute 8.5 only a gray screen can be seen (without scene camera, without sound, without fixations).

Recording 3: Here the original neon glasses were used again with the same companion device. The recording lasted approx. 1 hour and 29 minutes. The duration is also indicated as such on the companion device. However, when I start the recording, I can see that the video only lasts 4 minutes and a few seconds. In addition, the frame rate is greatly reduced. The video can be found in the Pupil Cloud with a duration of 1 hour and 29 minutes. However, the frame rate is also greatly reduced here and only the first 4 minutes have sound. This is followed by a gray screen for about 14 seconds, after which the recording of the scene camera can be seen again, but the sound is missing here. Again, there was no error message after the recording.

I now have a total of three recordings that can no longer be used. This is very annoying because these recordings should be used for a thesis. Is there a way to avoid these mistakes in the future or even to restore the recordings?

Thank you in advance Jonas

Hello pupil team. I am working on ROS2 integration using the Pupil Neon, but I'm experiencing significant artefacts when connecting to it. The image looks very broken after moving and takes 1-2 seconds to go back to normal. Is this purely caused by network quality? Will a wired connection solve this issue?

Will switching to the async API improve performance?

After some experimentation I found it probably was my current desktop. Switching to my laptop with a stronger Wi-Fi card solved the issue

Hi all, was the working range of the Neon tested? If yes, how much is it? Is it equal vertically and horizontally? Thank you!

Hi @user-97997c ! Do you mean whether accuracy has been measured across the field of view? If so, have you seen our whitepaper measuring Neon's accuracy?

Hi - can somone help me rescue a corrupt file?

Hi, we’re trying to record 30-minute sessions of data across 4 devices and are finding that the companion app/device will crash about 16% of the time. Can you provide any troubleshooting steps? We’re using LSL Rely. Could that be a factor? Relatedly, is there a desktop app, like Pupil Capture that is compatible with Neon?

Hello, I would like to know if the data exported using Neon Player matches the data downloaded from Cloud? thanks

Thanks

Hi all

I'm new in Neon. I have registered a video and I want to see on the desktop. I installed the Neon Player v4.0.0 but I receveid this messagge: session setting are from a different version of this app. I will not use those

can you help me?

Hi - I have a logistic questions/ask for advice: when I record an experiment with different participants, would you have one wearer per participant, each with either one or multiple files (depending on whether the same participant is doing >1 trials)? Is there another/better way to organise my recordings or is that what you would do/did?

Hi - question on fixation: I have tried to compare two different trials using events. In one, I moved my eyes restlessly, very quickly all around the room without stopping on anything in particular. In the other trial, I fixated one point for a few seconds. I then plotted the fixation durations and it shows a much higher value for the former trial than for the latter. I was expecting the exact opposite. I used this tutorial: https://docs.pupil-labs.com/neon/real-time-api/track-your-experiment-progress-using-events/

Can you explain why? Also, in the code on the tutorial there seems to be some confusion between fixation numbers (which I would expect to be higher in the first trial) and fixation duration mean (which I would expect to be higher in the second trial). Thanks for the help!

Hi, one subject's eye video is frozen after 15 seconds, but the gaze point and video are good. Do you have any idea about how to fix the eye video?

Hi @user-b6f43d 👋🏽 ! The processing time for your enrichment can vary depending on several factors. The duration of your recordings and the current workload on our servers are two major variables.

We highly recommend to slice the recording into pieces and apply the enrichments only in the periods of interest, using Events.

See also this relevant message for more details: https://discord.com/channels/285728493612957698/1047111711230009405/1186329479753256991

After the analysis iam not getting anything in the heat map, what might be problem ?

its not showing anything

Hi

I have a question regarding gaze_visualization when downloaded from cloud. When I see gaze data in Neon device, gaze seems much noiser that when I look at the gaze_visualization video from the cloud. Actually, when I manually plot gaze data over the pupilabs video, raw data seems noiser than the red circle from the gaze_visualization video.

Hope I make sense. Do you apply any filter to the gaze data in the video from pupilabs? If htis is the case, it would be great to knwo which filter you are applying.

Thanks!

Hello, I have been trying to use the newly developed Neon Player software instead of the Pupil Cloud because of online data processing restrictions with my University. However, the software doesn't seem to have any enrichments/pugins for detecting faces in the visual environment (i.e., an equivalent of the Face Mapper in Pupil Cloud). Is this in fact the case? And, if so, are there any future plans to add a face mapper enrichment to the Neon Player as well? Thank you!

@user-068eb9, responding to your message (https://discord.com/channels/285728493612957698/1195352014486523925/1197911358050664570) here. Are you asking about correcting the gaze signal or correcting mapping after running an enrichment, like Marker Mapper or Reference Image Mapper?

hello..i have been facing some issues suddenly with log in issues in pupil cloud...is it only me or is there any maintanence going on ?..so it doesnt say incorrect credentials etc but doesnt let me in as well

Hello - I am trying to use marker mapper. This is what I see on my screen: it seems it doesn't allow to add markers on the enrichment before running it. Am I doing something wrong? [also, it would be super useful to have more details in the tutorials... many things like these are kind of given for granted 🙂 ]

Hi What is the width and height of gaze data in pupil neon? Are they 1600 width and 1200 height? In pupil invincible, it is 1088 width and 1080 height.

Hello, what do the red thick circes and the blue ones in the results of the scene camera mean? What is the difference?

thanks for helping out in advance

Hi Pupil Labs, I'm looking for a way to import events to my recordings on pupil cloud using an external dataframe (csv/docx or any other format). Is this possible right now? It'll be useful for ex. when we have multiple recordings which share the same events and the event timestamps were collected on a separate device as UNIX timestamps (on the same network clock).

Hi, also is it possible to import completely deleted but exported projects back again into the workspace?

I have the problem, that the face detector didn´t recognize every record of a face, since the stimuli (in a movie) were too bright or blinded the glasses. How can I add this info afterward?

thanks a lot in advance

Hello, i want to interaction Neon with my glasses. But i do not know where can i find the connect board PCB file?

Hi @user-a2d00f ! That would be the Bare Metal, if you already have Neon and only need the PCB, you can get it here.

Hi @user-e3da49 👋 ! The red circles are gaze, and the blue circles fixations. You can read more about them in the respective links!

When you say that you exported the projects, do you mean that you downloaded the data from Cloud? To be clear, data cannot be re-imported to Cloud.

Since faces were not always mapped in your recording, do you plan to manually label faces in your recordings, by hand? You can at least add events to the timeline that indicate when those unlabelled faces are present in the recording. Or have you found a different face mapper tool that you would like to try? If so, then other face mapping tools can only be run offline, not in Cloud.

Thanks a lot!

furthermore I have the urgent problem, that a started recording stopped unexpected and the following message appeared: "We have detected an error during recording!" - Now I am missing a recording of 2,5 hours and don´t know how to avoid such an incident in future. Is it possible for you guys to fish out the recording out of somewhere somehow? Please help..

Hello! Is anyone else having trouble downloading the 'timeseries csv and scene video' recording data? every time i try to download and open the zip, i get this error.

Can i get a tutorial video in which i learn about face mapper enrichment ?

You can also submit feature requests here: https://feedback.pupil-labs.com/

But i do not understand what meaning in your web of "we will open source ... and electrical interface. You can make your own interface PCB". If we have not the electrical schematic, how do we develop our own interface PCB.

Hi pupil labs. I tried your updated Areas of Interest (AOIs) and it's pretty cool!! I have generated the image in the video (https://www.youtube.com/watch?v=RUmcU8zxMF0 ), but in 7.0v it describes that it generates a CSV file, ‚ AOI Metrics After you have drawn AOIs you will automatically get CSV files of standard metrics of fixations on AOIs: total fixation duration, average fixation duration, time to first fixation, and reach. ‘ but I didn't find the standard metrics of fixations in the download. Comparing the old version, I only see a new variable "fixation detected in reference image".

Am I overlooking something here? Thank you!

Hi Pupil Labs Team, Currently, I am starting to set up my environment for my experiment (website user experience). I tried to set the monitor behind, in front and beside the window, but I found a little bit of a problem with the focus from the camera, it showed only a white screen and could not detect the website's component. I captured some examples under this message, can you give me information about my issue? Is the problem because of brightness, color contrast or something else? And do you have any tips about that? But the last capture on my laptop screen could work

Hello,

For my thesis, I am conducting an experiment/research using Pupil Labs Neon. It involves an attention study comparing the wearing of face masks. I will measure the difference in gaze behavior during real-life interaction with a person and interaction with a person on a projector. The research is framed as a study related to eating (to encourage participants to gaze more freely at others). Various factors will differ, such as the type of face mask, the gender of the research fellow, etc.

I need advice on data processing. The data crucial for me is "time to first fixation" and "fixation duration". Is there a way to automatically obtain this data with Neon without manual counting? I mean, through code or another method?

Thank you in advance! Hope to hear from you soon. If there are any questions or uncertainties, feel free to ask them.

For the marker mapper enrichment in Pupil Cloud, is there a reason that recording id is not included as a column in the surface_positions csv file? I see that it is included in the gaze csv file. Given that an enrichment automatically runs on all videos in that project, it makes it hard to parse out which segments of data in the surface_positions csv file comes from which video.

Also, I am noticing when using apriltags that when the user moves their head (e.g., tilts to the left or right), for about 5-15 frames, all or some of the apriltags show as not visible when they are still in view. When the user is constantly moving their head, these gaps in data become very frequent and the data quickly becomes noisy. Is this expected behavior or is there something specific about our setup that might be causing this? When the user's head is stationary, there is almost no data loss on detected marker IDs .

Hi Neon team, is it possible to set the parameters (e.g., wearer_id, wearier_ied, etc) of the wearer profile using the Real-Time API instead of using the companion app before the recording?

Hi Neon team,

having issues streaming the data to windows 10 laptop. any tips?

Hello team, I'm rohan. I was recently assigned a task to configure and setup the pupil labs plugin with the new neon configuration. I was closely following this pupil labs neon installation page ( https://www.psychopy.org/api//iohub/device/eyetracker_interface/PupilLabs_Neon_Implementation_Notes.html#pupil-labs-neon ) and setup and tried to install the pupil labs plugin through the plugin manager. But it fails repeatedly and shows the same error ( refer the image ). I tried re-installing psychopy entirely and it still gives the same error. I’m using a Macbook pro laptop with the latest m1 pro silicon chip. Any suggestions would be helpful. thank you

Can i map two surfaces simultaneously in a single frame ?

Hi! I am trying to use the iMotions Exporter within Neon Player 4.0.0, however I consistently get this error: Currently, the iMotions Exporter only supports 3d gaze data

gotcha, did not realize it was an outdated plugin. Will also update to 4.1 now! Thanks Miguel!

no worries! Kindly note that we have recently renamed the "Raw sensor data" to "Native Recording Format" and I am not sure if iMotions already reflects this change.

Regarding the improved thermal performance of Neon mentioned in the 'announce' channel; is that change due to hardware changes? firmware changes? Just the changes to the frames? Mostly curious to know if it was a firmware change that would improve the thermal performance of the neon modules we already have. Thanks!

Hi guys, there has been an update, I missed coming up and now I am wondering where I can find the "fixations_on_face" information since there is just "gaze_on_face" left apparently. Thanks for your help!

after I do the face mapping enrichment o.c. 🙂

and last one for now 🙂 what´s the advantage of the neon player in comparison to the Pupil Cloud? Also what is a "recording directory" I have to drag into the player?

Hi, anyone here have purchased Lenst Kit? Is this for NEON?

Hey Dom, sorry for the delayed response. its 2023.2.3.

Hi Pupil Labs, is there a way to check/set the parameters of audio-recording, gaze sample rate, etc. when remotely starting a recording with pupil-labs-realtime-api? I know we can set it with the companion device but it'll be much more dependable if we can ensure it before every recording without having to physically check the app settings.

Also, unfortunately, a recording of over 2 hours disappeared after a push-up notification on the screen regarding the low battery status of the phone appeared. This time I could find the Scene Videos and the Eye videos in the storage data of the phone - which I can´t trace back to its location. Please help me find it somewhere.. the wearer didn´t click on anything (neither save nor discard options were shown)

its with the installer from the website