bfranceschiello
commented
1 year ago I think you can perform one calibration at the beginning and drift check after, just make sure the trigger referring to the start of each task is well printed on the signal
Il giorno gio 7 set 2023 alle ore 11:57 esavary @.***> ha scritto:
While working on outputting calibration data, I found a potential alternative to recalibrating the ET before each task. Right now, we calibrate with 9 points at the start and then use 5 points for subsequent tasks. But I've seen a few places where they suggest that too much recalibration might not be the best idea because we could replace a good calibration model with an inferior one.
I wonder if it would not be better to conduct a thorough calibration at the start and subsequently perform "drift checks" for the following tasks rather than carrying out a full recalibration every time. WDYT @bfranceschiello https://github.com/bfranceschiello @oesteban https://github.com/oesteban ?
I suspect that, as the documentation seems to suggest, the ET retains calibration between tasks when it's not turned off. Can someone confirm if this is accurate?
For context, here are a few discussions highlighting this issue:
SR Research Discussion https://www.sr-research.com/support/thread-7741.html PsychoPy Forum Discussion https://discourse.psychopy.org/t/saving-eye-tracker-calibration-data/18236/3 Another SR Research Discussion https://www.sr-research.com/support/thread-7821.html
— Reply to this email directly, view it on GitHub https://github.com/TheAxonLab/hcph-sops/issues/174, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7BDAAFEN7QX3UMQHUMB33XZGK7PANCNFSM6AAAAAA4OUIJRY . You are receiving this because you were mentioned.Message ID: @.***>
--
Benedetta Franceschiello, Ph.D.s, (she/her/they/them)
Associate Professor,
School of Engineering, Institute of Systems Engineering,
HES-SO Valais-Wallis, Route de L’industrie 23, Sion, Switzerland
The Sense Innovation and Research Center,
Lausanne and Sion, Switzerland
Department of Radiology, Lausanne University Hospital and University
of Lausanne (CHUV-UNIL), Avenue de Provence 82 Linktree https://linktr.ee/bfranceschiello, Mat-Tech Lab https://sites.google.com/view/mat-tech-lab/home?authuser=1
While working on outputting calibration data, I found a potential alternative to recalibrating the ET before each task. Right now, we calibrate with 9 points at the start and then use 5 points for subsequent tasks. But I've seen a few places where they suggest that too much recalibration might not be the best idea because we could replace a good calibration model with an inferior one.
I wonder if it would not be better to conduct a thorough calibration at the start and subsequently perform "drift checks" for the following tasks rather than carrying out a full recalibration every time. WDYT @bfranceschiello @oesteban ?
I suspect that, as the documentation seems to suggest, the ET retains calibration between tasks when it's not turned off. Can someone confirm if this is accurate?
For context, here are a few discussions highlighting this issue:
SR Research Discussion PsychoPy Forum Discussion Another SR Research Discussion