Fix the order of vqf updates to the recommended order. #9

New Issue

MrUnknownDE · 2026-04-05T17:51:36+02:00

MrUnknownDE commented

2026-04-05 17:51:36 +02:00

Originally created by @ksy2645 on 2/20/2026

The text may be awkward due to the use of a translator.

Issue: If the order is wrong, this will cause an issue where it updateAcc with the gyrQuat of the previous sample.

Reference:

[VQF DOC - VQF class]

~~# Test Setup~~
Device: Chest tracker (LSM6DSR + MMC5603NJ)
~~Magnetometer: OFF~~

~~Gyro bias calibration logs:~~

[INFO ] [LSM6DSR:0] Calibrated gyro bias at 22.921875C: 10.919210 -13.017056 2.500000
[INFO ] [LSM6DSR:0] Calibrated gyro bias at 36.132812C: 11.463371 -14.227714 5.920565

~~# Test Method~~
~~1. Prepare a chest tracker with 2-point temperature gyro bias calibration (22.9°C and 36.13°C as shown above).~~
~~2. Power on the tracker and leave it resting flat for 30 seconds~~
~~ - Purpose: allow “rest” handling and any remaining gyro bias calibration to settle.~~
~~3. Wear it directly on bare skin and stay active for 30 minutes~~
~~ - Purpose: stabilize strap tension and bring the tracker to a realistic operating temperature.~~
~~4. Perform Full Reset → Body Reset.~~
~~5. Record and observe drift for 30 minutes on the Body Proportions screen.~~

~~# Result~~
~~Compared to the previous behavior, drift was reduced by approximately 1–2 degrees.~~

~~This isn’t a perfectly controlled test (human-body bias, strap tension, environment, etc.), but even considering those factors, the improvement looked significant in my use case.~~

~~I’m sharing this because I think it could also be beneficial to apply the same fix to the SlimeVR Tracker firmware.~~

*Originally created by @ksy2645 on 2/20/2026* The text may be awkward due to the use of a translator. Issue: If the order is wrong, this will cause an issue where it updateAcc with the gyrQuat of the previous sample. Reference: - [[VQF DOC - VQF class]](https://vqf.readthedocs.io/en/stable/ref_cpp.html#) ~~# Test Setup~~ ~~**Device:** Chest tracker (LSM6DSR + MMC5603NJ)~~ ~~**Magnetometer: OFF**~~ ~~Gyro bias calibration logs:~~ ``` [INFO ] [LSM6DSR:0] Calibrated gyro bias at 22.921875C: 10.919210 -13.017056 2.500000 [INFO ] [LSM6DSR:0] Calibrated gyro bias at 36.132812C: 11.463371 -14.227714 5.920565 ``` ~~# Test Method~~ ~~1. Prepare a chest tracker with 2-point temperature gyro bias calibration (22.9°C and 36.13°C as shown above).~~ ~~2. Power on the tracker and leave it resting flat for 30 seconds~~ ~~ - Purpose: allow “rest” handling and any remaining gyro bias calibration to settle.~~ ~~3. Wear it directly on bare skin and stay active for 30 minutes~~ ~~ - Purpose: stabilize strap tension and bring the tracker to a realistic operating temperature.~~ ~~4. Perform Full Reset → Body Reset.~~ ~~5. Record and observe drift for 30 minutes on the Body Proportions screen.~~ ~~# Result~~ ~~Compared to the previous behavior, drift was reduced by approximately 1–2 degrees.~~ ~~This isn’t a perfectly controlled test (human-body bias, strap tension, environment, etc.), but even considering those factors, the improvement looked significant in my use case.~~ ~~I’m sharing this because I think it could also be beneficial to apply the same fix to the SlimeVR Tracker firmware.~~ <img width="1446" height="722" alt="BEFORE FIX" src="https://github.com/user-attachments/assets/eca7cc99-74ed-406c-98c9-041d3bb2f766" /> <img width="1440" height="738" alt="AFTER FIX" src="https://github.com/user-attachments/assets/deb449bd-31e9-4488-84cf-4e09c538e607" />

MrUnknownDE closed this issue

2026-04-05 17:51:36 +02:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github/SlimeVR-Tracker-ESP#9