Hello @nmt,

When I am exporting the processed file from pupil player software, I am getting following two errors: 1. av.AVError: [Errno 12] Cannot allocate memory (libav.mpeg4: Failed to allocate packet of size 11170000) 2. BrokenPipeError: [Errno 32] Broken pipe

For your reference I have attached the screenshot also.... (I am using pupil core eye tracker) Please let me know the solution

Thanks @nmt, the issue is sorted.

Is there a new place to buy the frames if we're going the DIY route? Seems like the Shapeway store link is broken, and the previous two people to ask about this were not given answers.

hI all i am new to developmenet and having a query regarding open source software provided by pupils labs i want to utilise your code for doing some modification and the sequential flow of the code along with the in detail documentation i want to read does any help from community can be provided my main concern is to find the section of the code from which it is projecting the coordinates of eye gaze , that section i want to understand because there are 40+ files so how to dig the code

Hi, I am developing a custom script to use Pupil core in a psychophysic experiment

I am having trouble trying to run Pupil Capture on my Macbook pro M1 running MacOs - Sequoia, specifically

1. I can only receive camera feed through pupil core by running sudo /Applications/Pupil\ Capture.app/Contents/MacOS/pupil_capture as suggested by the official document

1. I can not receive images about each pupil

Helps are very appreciated, thanks in advance

Thank you for your help. PupilCore's world camera has such image quality. Is it broken?

@nmt @user-d407c1 Any info on how I can acquire the frames to go through the DIY route? This is time sensitive. I'm exploring different eye tracking companies and yours stood out because of the DIY route, but if this is not available anymore, I'll turn my attention elsewhere.

@user-d407c1 @nmt Thanks! I will try the script, I think I forgot to turn on eye 0/1 detection on the side menu

In addition, can we see the blink data on exported file using core? I don’t see a blink column in the export file (gaze_positions.csv, pupil_positions.csv). I am guessing that we can use confidence with a threshold to guess blinking, but just wondering if there is an official support of this data since there is a blink detector function in Pupil Capture

Also, I am having trouble understanding the timestamp variable in gaze_positions.csv, is it in seconds? it is usually in form like :13694.37392…etc, is it a offset from 0 seconds? so the above means around 1.3 seconds from beginning

Thanks ahead!

Hi, I wanted to know if the cables connecting the IR cameras to the usb-c hub are custom-made. I might have accidently damaged one of the connector that goes into the IR camera and am looking to replace it.

Hello, I would like to ask about the calibration process in Linux. I clicked the mouse five times as required but it always says Gaze data error

Wǒ fāxiàn wǒ fànle yīgè cuòwù,5 cì diǎnjī shì jìnxíng qǔxiāo, dàn wǒ zài shǐyòng ScreenMaker jìnxíng jiàozhǔn shí píngmù shì bái píng 47 / 5,000 I found that I made a mistake and 5 clicks were to cancel, but the screen was blank when I calibrated with ScreenMaker

Attached is a calibration video. In addition, I used WSL2 to run from source code

I want to integrate yolov5 object detection with the pupils labs eye gaze tracking software can anyone tell how to do it

Hello, I have a problem with installing the Pupil Capture plugin. I was trying to write a plugin to send the data I received from the AI back to Pupil Capture. My problem is that I can see the plugin is running, but when I check the logs, I encounter the following error. I tried running a simple contour edge detection code based on the example here: https://gist.github.com/papr/b938ddc6315525d0f03da3668568e75c. I would greatly appreciate your help!

Noob question: I want to have live data sent via UDP of where my participants are looking at at the screen. Also I want to have just one eye-cam. Are these markers what I need?

So far I only know that Core sends UDP messages with raw tracking data. And that I can do some kind of calibration.

(image just shows the markers I saw here in chat. So you know what I'm talking about.)

cool, good to know. thank you!

since I have the attention now (😅 ): Are there any trade fairs where Pupil Labs attends to? So one could see your stuff in person?

that's cool. Yea, just found the announcements. Sorry, kinda new to discord.

I mostly stay in germany. But if someone tells me that THE eye tracking event is somewhere else, I would consider it.

Hi Team, i have recorded a few studies and recently I am encountering the error below when I try to replay the files. I have already tried to reinstall the software but the same thing occurs. I hope there is no data loss as we are currently conducting studies and cant afford to lose these recordings: Update: I tried to replay the files on another laptop and it works so it seems to be an issue with some settings on my laptop. Is there anything that i should try to fix? Thank you.

player - [INFO] video_export.plugins.world_video_exporter: World Video Exporter has been launched. player - [INFO] video_export.plugins.eye_video_exporter: Eye Video Exporter has been launched. player - [ERROR] launchables.player: Process Player crashed with trace: Traceback (most recent call last): File "launchables\player.py", line 661, in player File "src\pyglui\ui.pyx", line 274, in pyglui.ui.UI.configuration.set File "src\pyglui\ui.pyx", line 266, in pyglui.ui.UI.set_submenu_config File "src\pyglui\menus.pxi", line 774, in pyglui.ui.Scrolling_Menu.configuration.set File "src\pyglui\menus.pxi", line 91, in pyglui.ui.Base_Menu.set_submenu_config File "src\pyglui\menus.pxi", line 552, in pyglui.ui.Growing_Menu.configuration.set File "src\pyglui\menus.pxi", line 91, in pyglui.ui.Base_Menu.set_submenu_config IndexError: pop from empty list

player - [INFO] launchables.player: Process shutting down.

Hello, I currently have a problem. I want to draw something on the world interface. How should I do it to complete this?

Hi @user-becdcb! I'm just trying to better understand your goals here. What do you mean by 'debug the plugin'?

I am simply trying to perform edge detection using contours. In debugging, if I want to print something like "no edges found," how can I do that? I tried using logger = logging.getLogger(name) for outputs, but I see no output in the def detect(self, frame, **kwargs): method. Does this mean the plugin is unable to run this method if nothing appears there?

AI Pupil Detector

Hi everyone, I need your help with my first project using Pupil Core. I am using the eye tracker to ensure that my participants fixate on the center of the monitor. I perform a standard calibration and validation in the Pupil Capture app, and the app shows a confidence level for both eyes between 0.9 and 1. However, when I record the eye movements and analyze the data in Pupil Player, I notice that the fixations are consistently about 10 cm shifted from the part of the display I was actually looking at. Is there anything I can do to improve the accuracy and quality of my data?

PupilDetectorPlugins can also

Hi @user-177371 👋 ! Good to hear that you have great confidence levels for both eyes. Your value indicates good detection of the pupils in each eye image. This is an important pre-requisite for good eyetracking performance with Pupil Core.

Then, to understand how well the device is estimating gaze, you want to take note of the Accuracy value that is shown in Pupil Capture, after you have finished a calibration. If calibration accuracy could be improved, then that could potentially explain the results you are seeing.

Make sure to also fit a good 3D model of the eye before proceeding to calibration.

Otherwise, if the headset has slipped significantly during the recording, then this could also explain the result.

Ultimately, if you could share a screenshot or perhaps even the recording with us, then we can provide better feedback. You can do this privately by sending an email to [email removed] if you wish.

~~Greetings! I am new here to the Pupil products and I currently have the core. I am working on getting real time data using the network API, but am running into an issue. I am using the filter_messages.py template and it does not appear to be working for me. It gets hung up at requesting the sub port whether it’s local or remote. I am not really sure precisely on how to debug this, so I wanted to see if anybody had a similar issue or any advice. Thanks everybody!~~

Got it working!

Hello, I would like to ask what is the abnormality when I calibrate the pupil core?[03:09:50] WARNING world - uvc: Turbojpeg jpeg2yuv: uvc_backend.py:917 b'Premature end of JPEG file'
WARNING world - uvc: Turbojpeg jpeg2yuv: uvc_backend.py:917 b'Premature end of JPEG file'
^C[03:10:00] WARNING MainProcess - root: Launcher shutting down with main.py:342 active children: [<Process name='eye0' pid=147
parent=104 started>, <Process name='world'
pid=115 parent=104 started>]

hi ! Im trying to use pupil-capture to track my eye using a webcam. For some reason i cant get the stream from my webcam into the app. Im fairly sure that my webcam is working properly, as i can see footage from it in camera app. Can anyone please help? Adding some logs --

till now i've tried 1. falling back to v 3.4 2.running the program with admin permissions

Hello, I’m using your core product well. I have a question: In the Gazer3D class, the eye0_hardcoded_translation and eye1_hardcoded_translation are set to [20, 15, -20] and [-40, 15, -20] respectively, and ref_depth_hardcoded = 500. Are these values in millimeters (mm) or pixels (px)?

Also, I’d like to know the unit for self.last_gaze_distance. Thank you

The software also doesn't work with my laptop's inbuilt webcam so thought I'd say that too

Is there any off-the-shelf camera module for the eye-camera including IR emitters that you would recommend? E.g. ones you used on the core during the years? Or were these custom designs all the time?

I know you have DIY instructions, but these modules are so clunky. And not having to remove any filters would be kinda nice, too.

We have this module, but the emitters are to close to the lense: https://de.aliexpress.com/item/1005004639261673.html

Hi, does anyone has tips for massive export of recording data? Currently, I am manually dragging recording files into Pupil Player and hitting the 'e' key for exporting. I tried looking for API but have no luck. I am wondering if there is an open API to automate the process and to have the gaze_position.csv and pupil_position.csv loaded into appointed file

• I am using pupil core

Thanks !

Hello guys, I hope you are doing well. Some questions regarding eye tracking Core: 1) I wrote a python code for my study where I included the recording of calibration and the presentation of stimuli. I saw that using validation after calibration is regarded as a best practice however I haven't found anywhere the code to run the validation using python. Is it required to use validation after a good calibration? For your information I use 2D method for calibration and I have 6 small experiments (4-7 minutes) where participants are static (in sitting position). 2) Also, sometimes even when the calibration is good according to the precision and accuracy numbers (e.g., both below 1) I still get an error ''20% of pupil data was dismissed due to low confidence'' I am aware that this indicates that the pupil detection can be improved. I wonder what are the implications of using such data. 3) Which is the rule of thumb for the distance between the participant and the screen? I use a 27.5 inches screen and I noticed that being too close (e.g., 60cm) often results in calibration failure 4) Also, if we increase the time of calibration, does this increase the chances of having better results?

Hello, I have one of the first version of the core mount. Some part of the eye tracker are 3d printed. Unfortunately, the case attached to the mount seems to be broken inside. The USBC port is loose and cannot be connected anymore. I would like to open the case and 3d print it again. Are those STL files available somewhere?

Hi @user-412dbc! Thanks for your questions. Here are the responses: 1. Validation essentially verifies the accuracy of a calibration. It's not necessary, but it can provide important information. For further reference, see this message: https://discord.com/channels/285728493612957698/285728493612957698/838705386382032957 2. The implications are that you are losing 20% of the data collected during the calibration. It's difficult to provide further feedback without seeing a recording. Could you post one here for feedback? Or share it with [email removed] We would need the full recording folder that includes all stages of setup and calibration. 3. Generally, the goal is to calibrate at a distance close to the viewing distance you'll use in your experiment. For practical reasons, this isn't always possible. However, note that the 3D pipeline extrapolates outside the calibration area/distance quite well. 4. Potentially, but it's a trade-off. After a while, your participant might become fatigued and glance away from the calibration target, introducing errors into the calibration. If they can maintain full concentration and focus on the target for the duration, a longer calibration that covers more of the visual field can technically yield better results.

Is there any 3D print support for the pupil core glasses frame?

Hello @nmt I facing this error. Please let me know what is the issue?

AFAIK, currently no. "Shapeways" had a webshop where you could buy the 3d models for the frame. But the shop is closed down. You can still find the link in the DIY section: https://docs.pupil-labs.com/core/diy/#getting-all-the-parts Not sure if the models will be made available again. Maybe the pupil team has some insights.

Hi, we are trying to get pupil capture working on Linux. Installed the application. Icons are available but on clicking the application does not start. Could someone help?

Hi @user-3bcb3f 👋 ! Could you share some details about which distro, computer architecture and version of Pupil Capture are you trying to use? Does it work if you try starting from the terminal ?

Computer details are in the screenshot and Pupil Capture version: pupil_capture_linux_os_x64_v3.5-8-g0c019f6. When I start from terminal, it says cysignals failed to execute...and terminal closes..so I am not able to get the whole error.

May I ask if, besides the integrated GPU, you also have a dedicated GPU? If so, could you try right-clicking the application and selecting "Launch with dedicated GPU"?

Additionally, my colleague @user-f43a29 mentioned that our software supports only X11 and not Wayland. He has more expertise in Linux and might be able to provide further insights if needed.

Hi @user-3bcb3f , I am briefly stepping in for @user-d407c1 . Do I understand correctly that the system does not have an Nvidia or AMD GPU? In other words, was the system previously used as a server?

Running from terminal will typically only work if you instruct the system to use your dedicated GPU when launching from the command line.

Before that, could you briefly run the following command in the terminal and send the output?

lspci | grep -i vga

Thanks. That chip is certainly powerful enough, but we have not extensively tested using integrated GPUs with Pupil Capture on Linux, so I cannot make any guarantees.

Having said that, may I ask, are you trying to run Pupil Capture from source?

Ok. My experience has been that the cysignals error, when seen on Ubuntu, is usually resolved by:

1. Switch to X11
2. Use the GPU drivers supplied by the manufacturer, not the open-source drivers
3. Starting Pupil Capture by right-clicking the icon in the Application Launcher and selecting: Launch with dedicated GPU

With respect to point 2, you might want to check if Intel provides dedicated Linux drivers for that chip, rather than using Ubuntu's open-source default of MESA.

Since you have only an integrated GPU, that option will not be present. It probably will also not be present after installing Intel's drivers, if they provide them. I just list it for completeness, as the usual process for resolving the issue.

As mentioned, I cannot make any guarantees that it will work with an integrated GPU on Linux.

In the worst case, you could swap in a GPU from some other computer. The dedicated GPU does not need to be incredibly new or particularly powerful to run the Pupil Capture GUI.

Hi, we see that pupil capture is not working on Ubuntu 22.04 with just an integrated GPU. We haven't tested yet with dedicated GPU. Has anyone reported this before? With Ubuntu 20.04 however, with integrated as well as dedicated GPU pupil capture works fine.

Hi @user-3bcb3f , looking back, it seems there was some success with that issue on Ubuntu 22.04 when running Pupil Capture from source with the develop branch, potentially with cysginals v1.11.4, although it's probably worth a test first with the latest version of cysignals.

Is it possible to copy-paste the error at the end of the document into the chat here? Or upload it into Discord as a text file with .txt file extension?

UnboundLocalError: cannot access local variable 'g_pool' where it is not associated with a value [16:02:46] ERROR world - launchables.world: Process Capture world.py:837 crashed with trace:
Traceback (most recent call last):
File
"/home/profnivethida/pupil/pupil_src/launchable
s/world.py", line 124, in world
from gl_utils import GLFWErrorReporting
File
"/home/profnivethida/pupil/pupil_src/shared_mod
ules/gl_utils/init.py", line 11, in
<module>
from .draw import
draw_circle_filled_func_builder
File
"/home/profnivethida/pupil/pupil_src/shared_mod
ules/gl_utils/draw.py", line 16, in <module>
from pyglui.cygl.utils import RGBA
ModuleNotFoundError: No module named 'pyglui'

Hello, After we download the data, Ιs there any way to inspect the quality of the calibration conducted and the data collected so we can decide if we will use them for further analyses?