[Bug]: OneEuroFilter applied to DFR output

[RamesTheGeneric]

Version Number

latest

Describe the bug

The the OneEuroFilter is being applied to the DFR OSC output stream according to the state of the filter in the settings.

Steps to reproduce

1. Enable the filter and observe the smoothed, less responsive output from the dfr osc datastream.

2. Disable the filter, restart the app, and observe much more responsive output from the dfr osc datastream.

Expected behavior

DFR OSC datastream should always contain the raw, unfiltered datastream for the lowest latency.

Operating System

Windows 11

Hardware setup

Bsb2 native ET

Log Files

N/A

Other Notes

N/A

[dfgHiatus]

Is there an advantage of using raw data over filtered data for DFR? If we are concerned about latency, I can add some benchmarks to to the OneEuroFilter.

For Pimax, and DFR in general, the more data we send over the better (I presume), but what perf. goals do we have? Like sending data at 120hz, 90hz, etc.

[RamesTheGeneric]

Yes, there is very much a noticable difference in latency. Not from the actual computational time it takes to run the filter but rather the "smoothing lag" it adds. As for dfr in general I'm not sure if sending more often would help in this case. I know sending as soon as we have a prediction would be optimal though. Speaking of which, I remember seeing a fork a while back regarding a memory mapped file. What is our stance on using that as the means to transfer gaze data to steamvr's eye gaze endpoints?

…

On Tue, Mar 3, 2026, 8:56 AM dfgHiatus ***@***.***> wrote: *dfgHiatus* left a comment (Project-Babble/Baballonia#232) <#232 (comment)> Is there an advantage of using raw data over filtered data for DFR? If we are concerned about latency, I can add some benchmarks to to the OneEuroFilter. For Pimax, and DFR in general, the more data we send over the better (I presume), but what perf. goals do we have? Like sending data at 120hz, 90hz, etc. — Reply to this email directly, view it on GitHub <#232 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AMVTM2EO3CVG7VBDWZINDR34O3P7LAVCNFSM6AAAAACWDOLGHGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTSOJRGI2TCMRXHE> . You are receiving this because you authored the thread.Message ID: ***@***.***>