@user-df9629 your question from core ported to this channel:
thank you for clarifying the functionality. I thought the "R" was for record. My bad. One more thing, I noticed the reset button on the companion app but couldn't figure out how to add the offset. Any advice on that?
Offset correction is an feature that we're currently testing (not a finalized feature). In the current iteration you can use it by doing the following: 1. Make a short recording with the Wearer looking at known landmarks in the scene. Save the recording. 2. Open the recording from the recording view. Pause the video when wearer is looking at the landmark. Touch & Drag anywhere on screen to move the crosshair to the "corrected" position. Click save offset button. 3. Any gaze data now produced with this Wearer profile will now have the offset applied. Gaze circle color changes from red to green to indicate offset applied by the user. 4. You can reset the user defined offset in the wearer view.
Thank you @wrp . This will help a lot!
We acquired pupil invisible few days ago and we noticed an error of gaze estimation
is there anyway to fix that by calibration ?
@user-367c1f , @wrp talked about this in the link above, I hope this is helpful. Let me know if you need more input.
@user-367c1f if you have any questions/feedback after trying the above steps, please let us know.
Today the phone is not saving the data to the cloud. It is stuck at 0 percent, with the option of pausing. Any idea?
@user-30a16e I believe this was responded to already in another channel of communication.
@wrp your previous link was helpful indeed we could check it quickly
I would like to know if it is possible to change gaze offset calibration after a recording and post fix the gaze data
so far we have been doing it manually
@user-367c1f post-hoc offset is a feature that we are considering adding to Pupil Cloud. Please also note that the current offset feature is very much a beta feature that is being actively iterated on/developed.
Hi, I am testing the raw IMU data and the accelerometer data makes sense when read as float32 big endian instead of float32 little endian (as mentioned in the developer docs). Can somebody please look into this?
I have an issue with one PI recording when exporting to pupil player
I'm getting " Oops! There was error updating the recording" message and no gaze data is apparently generated by P player
Lookup tables for Eye0 and Eye1 are missing so I'm wondering what would be the issue on player
The recording data (the one with update problem )seems similar to other ones
Hi @user-367c1f thanks for reporting this issue! Can you send us the player.log file? If you are running Pupil from bundle, you can find it in your user folder under pupil_player_settings. It is overwritten every time you open Pupil Player again, so please make sure to reproduce the issue and then send it.
This is the player Log after error
Hi, I acquired pupil invisible two days ago, and I noticed that the pupil invisible recording folder does not have any audio files. If I want to record audio using pupil invisible, what I need to do to enable audio recording?
Hi @user-367c1f
thanks for sending the files. Can you also share the info.invisible.json
file that you find in the recording folder?
Hello! We purchased the invisible a while ago. However, the videos are not saving to the cloud (it remains at 0%). Has this question been resolved? I saw that this question had been asked in this discussion, and it was mentioned that this issue was discussed in another channel of communication. However, I cannot find where this issue was discussed in another channel of communication.
@user-c5fb8b
Hi @user-87c1b3 what version of Pupil Invisible Companion app are you using? The issue had to do with the app version pointing towards an staging cloud env.
Hello, i want to use the Invisible Glasses in a sports setting. I have a question about the analysis.
Sorry for the split post.
Is there already a software which contains standard analyis like saccade time, fixation time etc.? Or do i have to program somthing?
Hi, is there a way to do the data collection with PI offline and afterwards get the calibration? We like the automatic calibration but we don't have internet connection in outdoor settings. Also, do I need to do something to enable audio recording using PI?
@user-367c1f Thanks you, we are looking into the issue and will report back to you soon!
Hi @wrp thank you for the message! I am using App version 0.6.19-beta-prod. Is this the app that had this issue?
Thank you for updating the recording format's documentation 🙂
@user-df9629 sure, sorry I did not get back to you on that, but we acted on your comment and fixed the docs.
@mpk , no problem. Thank you so much for looking into it!
@user-fd5a69 Responses to your points: - You don't need internet connection to use Invisible Companion App. You only need internet connection once for sign up/log in. If you want to use Pupil Cloud features, you will need an internet connection to sync wearers and templates with the DB. You can record offline and upload recordings later to cloud when your device is back online. - What version of the app are you using? - Audio - we will enable audio recording in the near future. All Pupil Invisible devices have a microphone, but we just have not yet enabled recording in the app.
@user-87c1b3 you should be able to upload to cloud.pupil-labs.com with this app version. The issue you are experiencing is not the same as the other user. I have a feeling this is related to your app settings. Can you please go to the app settings menu and check that Cloud upload toggle is on
?
@user-7df12d Pupil Invisible would be a great fit for sports research. You will be able to classify fixations in Pupil Player (our open source desktop app). In the future we will likely be adding more enrichments features to Pupil Cloud as well.
Hi @user-367c1f with the new release of Pupil v1.17 we might have fixed the issue of opening your Invisible recording. Please give it a go and report back to us if you are still having problems!
@user-c5fb8b thank you very much. We could get eye tracking data with the new release
Hi, I noticed a delay in scene video from the time I start recording on companion till it actually starts recording. Is this an intentional delay? If yes, can somebody tell me the exact delay in seconds/milliseconds ? The delay is shown in the playback on the app as well. Its almost 2 seconds from my observation but I can't measure it precisely.
@wrp I had updated the app through Google Play, and I was able to upload the videos. Thank you again so much for your help!
@user-87c1b3 thanks for the update. Pleased that this was resolved.
@user-df9629 I don't have the numbers in front of me right now, but it does take a little bit of time before frames are captured from the scene camera.
@wrp , no worries. Thank you for confirming the lag.
Hi, I know offset feature is still not a finalized one. May I suggest adding/ saving the offset as a separate file or in the info.json so the raw gaze data is untouched, and the offset can be applied to every individual recording as opposed to just the wearer.
@user-df9629 were working on a number of changes to the situational offset - thanks for the feedback
@wrp , thank you, and you're welcome!
@wrp Is there already a possibility to get information of saccades or quiet eye?
Thank you for the first informations above
@user-7df12d you can open Pupil Invisible recordings in Pupil Player and use the fixation detector. You can change the parameters to detect very long fixations. Would that help?
@mpk That helps but is there something to detect or analyse the saccades or gaps between fixations?
Does invisible has IR LED?
@user-94ac2a yes it does, but its only required when there is no environment IR present.
@mpk thx. And Invisible is running the same core algorithm?
Pupil Invisible uses machine learning for gaze estimation. Pupil Core uses a 'traditional' computer vision pipeline.
So invisible is better?
@user-94ac2a Invisible is better for work 'in the wild'. Pupil Core is better for the 'lab'.
Ok. Will core move to machine learning based algorithm in future?
Hi, when will support for working with imotions be available?
I am working with Invisible but I also need to process data with Imotions 😦
thanks
Hello, has anyone been able to run a setup for the pupil invisible for at least 4 hours? maybe by using an battery phone case or a usb c OTG splitter?