This forum uses cookies
This forum makes use of cookies to store your login information if you are registered, and your last visit if you are not. Cookies are small text documents stored on your computer; the cookies set by this forum can only be used on this website and pose no security risk. Cookies on this forum also track the specific topics you have read and when you last read them. Please confirm whether you accept or reject these cookies being set.

A cookie will be stored in your browser regardless of choice to prevent you being asked this question again. You will be able to change your cookie settings at any time using the link in the footer.

Differences in distance-out-of-GPS-data between Phyphox versions?
#1
Hello,
I use Phyphox for (slow) running only. The iphone tracks the GPS data, afterwards Excel converts the data.
Before the last update the calculated distance was about 30% less than in the actual version. The calculated speed seems to be unchanged.
I didn't find any relating information except:
"Changelog -> Änderungen für Android und iOS -> GPS-Experiment zeigt nun Entfernung vom Startpunkt"
Did the algorithm for extracting distances out of the GPS data change?
Thank you!
Cheers, Timo

edit:
If one uses the exact(?) formula from kompf.de
dist = 6378.388 * PI/180 * acos(sin(lat1) * sin(lat2) + cos(lat1) * cos(lat2) * cos(lon2 - lon1))
the distance seem to be comparable to the "old" calculation.


Attached Files Thumbnail(s)
   
Reply
#2
The "distance from start" is just the distance from the current position in a direct line to the start of the measurement. It is not exported and it should not affect your data.

The other change (which was not documented separately as it was supposed to be only a "cleanup" of the code) in the experiment configuration was that the individual steps of the calculation were replaced by the new formula-node, which is easier to read. However, it looks like I made a little mistake by mixing up a variable with the result that change in longitude is counted as if you were running on the equator instead of somewhere in Germany, which would fit the difference you are seeing.

Can you provide the exported raw data? This would make it easier to verify that this indeed is the error.
Reply
#3
(07-16-2019, 11:00 PM)Sebastian Staacks Wrote: The "distance from start" ...
Can you provide the exported raw data? ...

Hello,

see the excel files in the attachment, please.
If you prefer another format provided by Phyphox (which type of CSV?), please let me know.

20190717_Phyphox_name_mismatch.PNG
If I look for the older measurements, there seem to be some mismatches between the visible names in the list (here to be seen as "Lauf HZDR 19.07...") and the names in the dialogue (here "Lauf HZDR 28.08.18, 16:17").
So it's not easy to find the wanted measurement, one has to look into all files.
Thank you for help and for the nice program suite!

Cheers, Timo


Attached Files Thumbnail(s)
   

.xls   phyphox 2019-04-16 19-34-24.xls (Size: 386 KB / Downloads: 733)
.xls   phyphox 2019-05-07 16-44-12.xls (Size: 517 KB / Downloads: 653)
Reply
#4
Thanks, it will be fixed with the next update, hopefully within the next week.

The names of the saved states are shortened with trailing dots "..." because of the limited width of the column. It looks to me like you are using an increased font size? It looks like we could be a bit better at estimating how many letters would fit in there.
Reply
#5
Might it be that a large dataset in a saved state causes a loooong startup time for phyphox?
Reply
#6
Yes. I assume, you have seen this on iOS? In theory, this might happen on Android as well, but it should hardly be noticeable, but the problem is worse on iOS because of the way, the experiments are loaded on iOS. The problem is that the recorded data is saved in a single file along with the experiment description and therefore all the data from every experiment has to be parsed when loading the experiment list.

We plan on separating the saved data from the experiment configuration soon, so a saved state would be a zip file containing the experiment configuration along with a binary file containing the recorded data. In fact, the student who implemented the new iOS parser has already implemented this on iOS as well, but since version 1.1.0 took so long and this still had to be implemented in Android, this feature was postponed. I am hoping to get around to this around autumn.
Reply
#7
A little update about the fix:
I have tested the new formula by comparing with the odometer of my car over 50km and they agreed to within 3%, which I attribute to using the average earth's radius in our formula instead of the actual geoid elevation.

Unfortunately, Apple has rejected the update because they came around to read the GPS experiment description and objected to our explanation on how Android and iOS give slightly different information (the text has been there for ages and is even slightly outdated, but maybe they came across this because our release notes now mention GPS). Therefore it will take a few more days until this is settled, although we might release the Android version sooner without waiting for Apple.

By the way, the disputed info is actually rather relevant to this thread:
Since version 1.1.0 we include all the location data in the GPS experiment, which not only includes satellite data, but mobile and wifi-based location as well. Since these data points are less accurate, they may increase the calculated distance further, so you may want to enable the setting "only use satellite data" in the experiment's menu. This is not possible on iOS though (and we might not even be allowed to inform iOS users about this problem in the future).
Reply
#8
(07-22-2019, 09:34 AM)Sebastian Staacks Wrote: Yes. I assume, you have seen this on iOS? In theory, this might happen on Android as well, but it should hardly be noticeable, but the problem is worse on iOS because of the way, the experiments are loaded on iOS. The problem is that the recorded data is saved in a single file along with the experiment description and therefore all the data from every experiment has to be parsed when loading the experiment list.

We plan on separating the saved data from the experiment configuration soon, so a saved state would be a zip file containing the experiment configuration along with a binary file containing the recorded data. In fact, the student who implemented the new iOS parser has already implemented this on iOS as well, but since version 1.1.0 took so long and this still had to be implemented in Android, this feature was postponed. I am hoping to get around to this around autumn.

True, this is on iOS. This is no big deal since the data can also be exported, so there's no real need (for me) to save a state with lots of data.
Reply


Forum Jump: