Replies: 5 comments 15 replies
-
|
Every trackpoint in openrowingmonitor tcx is a stroke. But in fact in a tcx file, a trackpoint should not be stroke driven but timestamp driven. Is it possible to have a trackpoint every second in tcx file (or in rowingdata csv) generated by openrowingmonitor ? |
Beta Was this translation helpful? Give feedback.
-
I am not familiar with tcx conventions but had quite some experience with the FIT SDK. Recently garmin pushed out its official SDK to javascript. Currently currently it only allows to decode js files but I suppose it is just a matter of time that encoding will follow: As a side note, there is an issue with the official SDK, namely that the record message does not include the stroke_distance for some odd reason, but the actual message number exists (87)... so the SDK needs to be extended with the garmin tool (FitGen) if that field is need.
I think this is the behavior of the Conecpt2 log as well when you export the data into fit. I wonder what is the use case for the need to have datapoint in every second. I mean apart from HR I only see a more spikier pace without actually added information. Can you provide more reasoning for this? |
Beta Was this translation helpful? Give feedback.
-
With tcx, the unwritten convention for indoor rowing is that there is a trackpoint for every stroke. As almost all metrics exposed in the tcx are stroke driven (speed, watts), that makes more sense. There are multiple unwritten conventions we found out the hard way, like having to start with a starting trackpoint (so no time and distance covered yet), otherwise the strava data is wrong. I must say, I haven't tested the effects of deviating from the stroke based approach. However, intervals.icu complains about the lack of datapoints (for me, one around every 3 seconds), but I don't see how increasing the frequency would make the data better (as aside heartrate, all other metrics are calculated only twice per stroke).
For rowingdata it is an official option for their side to produce a time-based option (i.e. one entry per second). As they work with on the water rowers, it makes sense to process GPS data that way. On a more conceptual level, I'm not certain how to deal with it yet for an indoor rower: some more advanced metrics are only valid at the end of specific moments in a stroke (Handle force metrics for example). I could ask the people on their side what they expect there in that situation, to prevent messing up force curves and similar cool features. I would consider it a very good option once we sort out some internal issues in Open Rowing Monitor. The current TCX and RowingData recorders already could handle a second time-update frequency: they accept the input as is. Getting the data there is a bit more complex: @Abasz and I am looking for a way to continuously update consumers of data and let them sort out what they like themselves. We could add another update timer, to notify the file writers. |
Beta Was this translation helpful? Give feedback.
-
|
The main problem with stroke as a trackpoint is that the dateTime format should only contains second, the format does not allow fractional seconds. Same rule apply to fit record timestamp (coded as an integer). So actually the milliseconds are lost when you upload your tcx or fit file message to strava or garmin connect. When you use a smartwatch to record and indoor rowing session with ant+ or ble input from a rowing machine and/or heartrate sensor, the smartwatch create a tcx or fit file with trackpoint (tcx) or record message (fit) with a periodical timestamp every x seconds. I did not find any stroke_distance in Profile.xlsx from garmin sdk, i only found a avg_stroke_distance which can be included in session or lap message, but not in the record message. Do you have any Concept2 log fit file export to share so i can examine it ? |
Beta Was this translation helpful? Give feedback.
-
I was talking about tcx files, i agree with you that sdk is clear on that matter with fit messages. |
Beta Was this translation helpful? Give feedback.
-
Can i ask if you used a connect iq app to connect via ant fe-c and which one ? |
Beta Was this translation helpful? Give feedback.
-
|
If that watch supports ant fe natively, there is no need for an additional connect iq app. You can connect directly and use native data fields. I would recommend to refer to the user manual of the watch how to connect to fitness equipment. https://support.garmin.com/en-GB/?faq=rlynVxyyoq9dcSZGw5YFo8 Of course you need to use indoor rowing Activity type and the watch needs to support it |
Beta Was this translation helpful? Give feedback.
-
|
Ok, thank you. |
Beta Was this translation helpful? Give feedback.
-
|
I've tried today, and connected my watch to the rower as a "trainer" with success. Here is sample dump of 2 fit records : I love this ANT+FE-C mode. Great job ! |
Beta Was this translation helpful? Give feedback.
-
|
Thank you, i'll take a look at those files this afternoon. |
Beta Was this translation helpful? Give feedback.
-
First of all: great idea! It was on my wish list for a long time, as Garmin's data isn't as accurate as I like it to be (there are delays in my ANT+ setup, which cause time and distance to be inaccurate). I always considered it a furute extension of the tcx and RowingData file writers, as you want to pack as much data into it as possible, and OpenRowingMonitor has more data than the tcx. But I realize that getting a working solution is always the primary goal. In my persuit for decent setups, I found several solutions that could provide the basis for a native js ORM extension: Perhaps this helps you further along. |
Beta Was this translation helpful? Give feedback.
-
This one looks promising. |
Beta Was this translation helpful? Give feedback.
-
|
Dumped 8324850505_ACTIVITY.fit.txt with fitdump.pl, field 87 of record messages are related with stroke distance but some values are weird : they are all about beween 800-1000mm, but 303mm seems to be really weird (i think a stroke was missed because distance jump from 1635.00 m to 1655.99m). |
Beta Was this translation helpful? Give feedback.
-
|
As I said, this is a merge based on time stamp, which is error tone as the PM5 recording and the watch recording are started and ended at different times, it is not that accurate. I wanted to show this to make a point that two fit files can be merged in a way that you record the same data each second until new arrives, but I would not rely on it :) a better example is the 22 August recording from PM5 via ant fe-c. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Hello,
I have been playing with openrowingmonitor tcx files.
I've forked and modified https://github.com/beavis69/Coxswain2Fit to make it compatible with openrowingmonitor tcx files.
I was able to import the fit file to garmin connect without errors.
Another great tool is https://github.com/mrihtar/Garmin-FIT to dump fit files.
But actually i think that tcx files have some issues :
Beta Was this translation helpful? Give feedback.
All reactions