Step counter on Vertical

sartoric

@dreamer_ said in Step counter on Vertical:

SV and S9PP and they are using the bad algorythms just to avoid those royalties, right?

No
SV and S9 should afair still using FB

Edit
Very probably step counter was always Suunto’s own algo as FB doesn’t seem to provide that metric

https://www.firstbeatanalytics.com/en/features/

EzioAuditore

@dreamer_ I wouldn’t classify bad and good. It is something that is delivered to Suunto as a black box to use. Now this black box (Firstbeat) was purchased by Garmin and on its devices it seems to function better. But who knows if the ‘boxes’ delivered to Suunto have the latest features, bug fixes or even full compatibility with the watch sensors. So going to use own developed might be good might be bad. Good because Suunto can easily ‘fix’ and update staff, bad because the user base of Firstbeat and their algos is much bigger (irrelevant now since it is controlled by Garmin, same as ANT+. This is what conglomerates do, buy companies that are used by many and enjoy the benefits afterwards

dreamer_

@sartoric I see. Sometimes my “engrish” is not very good . In any case it’s just a matter of using the same algorythms as the ones used in the Race if possible. And that means, the solution and the fix should be already developed.

dreamer_

@EzioAuditore Thanks for the explanation. I didn’t know that either. The Suunto Race uses Suunto’s algorythms, right?
I understand the hadware is different but if Suunto has already developed the “good” logic for the Race, that’s pretty interesting.

EzioAuditore

@dreamer_ I believe that yes, the Race is using Suunto’s own implementation of fitness data (steps, sleep, etc.). For example I am not getting only different steps and calories but also sleep (the only similar thing is the HRV which I guess is directly influenced by the OHR). Plus the Race gives you REM stages of your sleep. I would love to see this ported to the SV as well but somewhere it was mentioned that it too needs enabling/paying (OHR or no I am not sure). Although, I do not really need the watch to tell me if I slept good or bad I was ‘observing’ myself and found that the more REM sleep I have the better at performing tasks at work I am and the more deep sleep the better at performing sports (this of course is my own observation not scientifically proven - maybe there is scientifically one but I do not have the time to invest and research it)

sartoric

@dreamer_
Some notes about Race’s algos

https://forum.suunto.com/post/131343

isazi

@dreamer_ a little background story you can also find in the forum if you look around. At some point the S3 algorithm, the one that is much less sensitive to movement, was used, and many users complained that it did not give them enough steps and they definitely walked more, and Suunto decided to use the more sensitive algorithm on the S9 series

dreamer_

@sartoric Dimitrios’ post is very claryfying
And this:
“Bottomline, things will get aligned , but needs time.” , totally consistent.

@isazi I don’t know about the algorythm used in the S3 but have a look at what @pilleus posted here: https://forum.suunto.com/post/144035

Those seem pretty interesting tests and what is for sure is that the algorythm that is now used in the SV and the S9P is horrible, the worst Suunto has used yet (as my understanding)

isazi

@dreamer_ @pilleus was using the S3

dreamer_

@isazi yes sorry. I was also thinking in the Race since we were quoting Dimitrios’ post.

But even the algorythm used in the S3 (we don’t know if it’s the one used in the Race) seems a lot more accurate of what we actually have in the SV. In fact It does pretty good compared with what the SV does. The SV counts steps all time (at this moment).

What it seems pretty consistent is that last Dimitrios’ post (and totally understandable in fact)

isazi

@dreamer_ said in Step counter on Vertical:

But even the algorythm used in the S3 (we don’t know if it’s the one used in the Race) seems a lot more accurate of what we actually have in the SV.

Yes, the algorithm of the S3 is widely considered the least prone to over counting steps.

Greennorth

@pilleus nice comparisons . My Suunto 3 was also very accurate for step counting. My S9PP counts almost twice as many steps, which was a surprise when I switched watches.

dreamer_

@pilleus it would be very interesting if you could recheck the step counter with similar tests you made before.

For me it seems that is not working after the fix. I still need to check at the end of the day but I woke up at 6:10AM and at 7:45AM the watch says that I have more than 1400 steps, which is totally impossible.

dreamer_

@pilleus . The counter seems fixed for me, which is good news. It is not counting steps all time like before. So good news.
My guess is that they are using now the other algorythm. I don’t know, of course. But It makes sense. Very well Suunto!

pilleus

@dreamer_ said in Step counter on Vertical:

@pilleus . The counter seems fixed for me, which is good news. It is not counting steps all time like before. So good news.
My guess is that they are using now the other algorythm. I don’t know, of course. But It makes sense. Very well Suunto!

Step counter is working better since the update, but the result is far away from the Suunto 3 and the real steps. I will check it the next days and do some tests with 9PP and Vertical.

dreamer_

@pilleus my impression is also that it is not counting all time like It was before but It is still counting more steps than it should. In fact, I published a post in other thread but I reseted the watch and I need more data to be sure.

dreamer_

This is definitively working much better than with the previous firmware. At least like 30% less steps are measured, which is fine.

Luis Andés Olmedo

@dreamer_ said in Step counter on Vertical:

This is definitively working much better than with the previous firmware. At least like 30% less steps are measured, which is fine.

I agree.

pilleus

@dreamer_

https://forum.suunto.com/post/148099

dreamer_

@pilleus thanks for checking it. So It has been improved but is still counting more than it should and more than the S3 (which was perfect), as we were talking about