Gradient lag on 1050?

"Updated grade algorithm to be much more responsive." in the new 23.09 beta for all x40 models. should be much quicker now , i will test today, yesterday i didnt check it. the 1030plus... a different story, that line with other x30 models are "dead", i doubt that will get this revised algorhythm via a new firmware. obviously the 1050 has this new method, so should act same way as the x40s with the beta. 

If you manage to perform this test with the beta version 23.09, I would be very grateful! Let us know, thank you very much.

GPLama as the Simsons, You already did it!

i tested it with my 840 and 23.09b and the gradient value is perfect, honestly "too perfect", it is MAGIC!!! (maybe some fused things with dual-band gps and dem file data)

E1040 firmware 23.09B - the terrain slope works perfectly, the display of slope values ​​is fast and the realization is fast. It's surprising that Garmin has been saying for many months that it has to be like this and that nothing can be done, and yet it turns out that it can be done. Let everyone answer the question whether this is a professional approach to using one of the most expensive devices on the market.

I absolutely agree with you, and Garmin needs to resolve the same issue on the Edge 1030 Plus since the devices are similar and probably contain the same localization chip. Will the same update for the 1040 ever come to the 1030 Plus? Thanks

i tested it more, in certain situations it is not so perfect, in dense forest or when there are more roads next to it but with different terrain (here at us usually the main road is flatter and the bike path near to it has much steeper ascent and descent or vice versa but sometimes i feel that the displayed number is not for the bike path rather for the main road or something else. (or the higher priority road, so my theory is that it is some kind of fusion with DEM and/or ClimbPro file and GPS altitude data rather than only the barometer measuerement ). i hope with the lots of multi gnss - multi band datas from Edge and other Garmin devices the Garmin maps can improve by itself (as the ClimbPro datafiles also), i think it is not an impossible task in 2024 with AI with big data.  any remark or comment about my idea?  aweatherall

If you have a backdoor to people listening at Garmin, please pass this also.

I for ages complained about how bad the grade calculation was on the 1040, even worse than the 1000, which seemed to have a lag of some 6seconds.

Likely the data used was very noisy and they had to low pass it (rolling average, IIR, etc) enough to make it stable. They probably overdid it, most likely their programmers aren't bikers.

In the 1000, I liked that at least on grades below 10% it would display one decimal digit, and fairly stable and accurate over a long steady slight climb or descent.  In the 1040 it is just the integer 0% 1% 2% 3% etc.

On both you would be getting kind of an average of the past 6 to 8s.  So if you are climbing 5 miles of steady 8% it would work fine and display accurate grade.  If you had a smaller climb with a max grade of just a few seconds, you would never know it was 22% on that curve, as the devices would only reach like 18% smoothing the data.  And in that the 1030 seemed worse than the 1000.

Now the good news - the 23.09 improves that a lot, the grade changes are faster, and after the top of the hill or bridge etc you won't be descending while the gps still says 8% 6% 4% 3% 2% 0% -1% -4%. It is like one second "lag" at most.

Still I lust for the 1000 decimal point, but can't say that in the new algorithm that would be noisy as ***, and I can live without knowing i am "climbing' 1.7% or 1.9%.

One thing I did notice, and it is annoying, is that very often the readings, on a gradually steeper climb, will go 0% 0% 1% 3% 5% 4% 6% 8% 9% 11% etc. - always a hiccup, an overshoot in the grade and later stable.

I would risk an opinion that they are likely instead of a 8s lowpass filter they changed to a shorter average, with edge enhancement to 'predict' grades that has this odd temporal response.  They could likely gather data from actual runs, or look at simulated data input and easily see that, but chose sharper response over accuracy.

It is like oversharpened pictures - you get those edge artifacts.

Hopefully they tweak it a bit to a situation without the fake lower grades at every climb.