Watch unusable in wind

Brad_Olwin

@cosme-costa The Suunto graph:

https://i.postimg.cc/qMhcJDDr/Screenshot-20250128-114024.png

The spikes are clearly the issue. I do believe you and it is apparent but…. I rarely, rarely see this with any of my watches. Can you contact me by PM? I am going to going to ask the test group. I think this issue is not common, at least for me. I may ask to collect some files.

Stefano M64

@Brad_Olwin said in Watch unusable in wind:

The spikes are clearly the issue.

the algorithm that calculates all the related quantities (altitude gain, loss, etc …) should include a running average to smooth out the altitude data, to eliminate the spikes if any.

stromdiddily

@Stefano-M64 said in Watch unusable in wind:

@Brad_Olwin said in Watch unusable in wind:

The spikes are clearly the issue.

the algorithm that calculates all the related quantities (altitude gain, loss, etc …) should include a running average to smooth out the altitude data, to eliminate the spikes if any.

if you use the search feature here, you’ll find that the folks on the forum actually helped with the algo ended up in use for the S9B when it first came out. There is some pretty open dialogue about what the underlying metric actually looks like

isazi

@Stefano-M64 interestingly enough, the more you smooth and correct, the more you end up losing real-time metrics like pace or vertical speed

isazi

@maszop said in Watch unusable in wind:

I think that further discussion is pointless because we have been writing the same thing for years and years go by and nothing happens.

Actually, just at the end of last year me and others provided Suunto with activity data for the same use case you have, bad weather (and wind in particular) affecting the ascent/descent calculation.
I believe it has not been prioritized because it happens to only some, and not always, and there is always something more urgent to fix.

Stefano M64

@isazi said in Watch unusable in wind:

@Stefano-M64 interestingly enough, the more you smooth and correct, the more you end up losing real-time metrics like pace or vertical speed

a good algorithm should be able to process raw data by eliminating noise and spurious values. Of course, that depends also on the kind of activity. For “slow” activities as walking or hiking data smoothing over 1 minute should give good results. It happens relatively often that SA gives altitude related metrics that are overestimated, only by importing the gpx data into an app (I use OruxMaps) that allow to the filter the data I get values that agree with the expected ones.