Pedal Balance in the file exported by ROUVY is not parsed

[Yi-Ting Chen]

I uploaded the FIT file exported by ROUVY to fitfileviewer.com for analysis, and I can see the data in the left_right_balance column.
Then import the same FIT file to GC, but GC cannot parse out Pedal Balance data.

The attachment is a record for reference, I hope it can be helpful.

[Mark Liversedge]

On Tuesday, 21 December 2021 at 09:14:22 UTC Yi-Ting Chen wrote:

I uploaded the FIT file exported by ROUVY to fitfileviewer.com for analysis, and I can see the data in the left_right_balance column.
Then import the same FIT file to GC, but GC cannot parse out Pedal Balance data.

That's right, the ROUVY app does not generate compliant files, the same issue is present with Garmin Connect.

I would contact the app publisher and raise it as a bug with their app.

We could change the code to support their badly formatted files, but it would impact other users with well formatted files.

If there is a way of supporting both in safe way I'd be happy to implement.

This has come up a few times recently.

Mark

[Yi-Ting Chen]

Hi Mark,

Thanks for your reply.

I also tried to report the problem to ROUVY, hoping to make the function more complete.

liver...@gmail.com 在 2021年12月21日 星期二下午7:28:28 [UTC+8] 的信中寫道：

[marcen]

A suggestion on how to handle the data might be.

After the query if right is set to one and the answer is no then as an intermediate step provide a list where devices are inside that show the correct l/r distribution but are not set to right one.
Devices that are not in the list will still be set to n/a.
The devices that are in the list will be checked once by hand if they are correct.

To be able to check if this makes sense, attached is a fit file that is recorded with the Wahoo element.
In the middle of the unit is round about one minute only with the left leg were driven.

The 100% value in the l/r distribution is equal to load only with the left leg and 100% left torque effectiveness.

Unfortunately, I'm at the Wahoo support not progressed. The have not seen the Probelm with itself because in the Wahoo app the l/r distribution is displayed correctly. Maybe someone else has more luck with the support.

[Ale Martinez]

El martes, 28 de diciembre de 2021 a la(s) 17:57:48 UTC-3, marcen escribió:

A suggestion on how to handle the data might be.

After the query if right is set to one and the answer is no then as an intermediate step provide a list where devices are inside that show the correct l/r distribution but are not set to right one.
Devices that are not in the list will still be set to n/a.
The devices that are in the list will be checked once by hand if they are correct.

To keep that list updated would be a nightmare, it is not just the devices, but all the APPs generating FIT files.

If we had to do something about this, which I am not convinced at all,  I would prefer to put a global check  like "Enable Non-Compliant L/R Balance" similar to existing one for Smart Recording and let the user make the choice.

 

To be able to check if this makes sense, attached is a fit file that is recorded with the Wahoo element.
In the middle of the unit is round about one minute only with the left leg were driven.

The 100% value in the l/r distribution is equal to load only with the left leg and 100% left torque effectiveness.

Unfortunately, I'm at the Wahoo support not progressed. The have not seen the Probelm with itself because in the Wahoo app the l/r distribution is displayed correctly. Maybe someone else has more luck with the support.

 

[marcen]

Thank you for the answer.

Okay, does not make sense.

Will try to formulate my questions and comments better.
How would it be better?

[Ale Martinez]

El miércoles, 29 de diciembre de 2021 a la(s) 20:51:21 UTC-3, marcen escribió:

Thank you for the answer.

Okay, does not make sense.

Will try to formulate my questions and comments better.
How would it be better?

I think your formulation was ok, but the proposed solution seems too difficult to maintain, specially for a problem created by others faulty implementation of an industrial standard intended to allow interoperability.

Try to import a non compliant FIT file to Garmin Connect (Garmin controls the specification, BTW), and you will find this is not a GoldenCheetah problem.

[marcen]

Thank you for the explanation.

I also see that the file must comply with the specification.

With Garmin Connect I had told the support. I did not see as a problem of GC.

I try again with the Wahoo support.

[Ale Martinez]

El jueves, 30 de diciembre de 2021 a la(s) 19:45:39 UTC-3, marcen escribió:

Thank you for the explanation.

I also see that the file must comply with the specification.

With Garmin Connect I had told the support. I did not see as a problem of GC.

I try again with the Wahoo support.

BTW, the argument: "They have not seen the Problem with itself because in the Wahoo app the l/r distribution is displayed correctly." doesn't hold water IMHO. When a company decides to implement a standard not controlled by itself, the main reason tends to be interoperability with other systems, specially if the specification is controlled by a competitor.

[Alan Benstead]

You can add FulGaz to the list of companies that do not meet the Garmin FIT file specification for Pedal Balance. I just sent them a support request to see if they will modify their FIT file generation, hopefully they will.

I must say I can understand why there is some confusion. From the specification snippet that Ale sent me some time ago, it states "right 0x80 data corresponds to right if set, otherwise unknown". The "contribution" can only come from either the right or the left pedal, so I think it would be better if the specification read "right 0x80 data corresponds to right if set, otherwise LEFT".

Anyway, will wait and see if FulGaz replies.

[Ale Martinez]

I am curious about which power meter are using people experiencing this issue and how it is connected (ANT+ or BTLE). Jjust trying to understand if the problem is generated by the way these APPs/Head Units generate the data, or they are passing through non-compliant ANT+ messages generated by the PM.

[marcen]

The two power meters are connected to Ant+ and are recorded with the Wahoo Element.
First is a Power2Max NGeco and the second is a Favero Assioma duo. Both power meters do not show the l/r distribution.

I had test a Garmin 530 and the Favero Assioma duo were connected to Ant+.
Since the l/r distribution was displayed in GC and Garmin Connect.
With the Garmin I had not made a test, whether the left or right Bien is set to 100%, if only one leg is driven. The same with Power2Max NGeco and Garmin 520.

[Alan Benstead]

I am using the Favero Assioma Duo dual sided pedal based power meters. When riding on the indoor trainer using the FulGaz app, the connection is via BTLE. The Left/Right balance is displayed correctly on the virtual head unit that FulGaz have in the app. I can change the effort from each leg and it matches correctly in the app. Importing the FulGaz FIT file into GC does not give any L/R Balance. Using the FitToCSV.bat tool I am able to see that L/R Balance is included in the FulGaz generated FIT file, but Bit 8 (0x80) is not set and therefore the data is ignored by GC.

When riding outdoors, I use a Garmin 830 to record data from the Assioma pedals via Ant+ and there are no issues importing and displaying the data in GC. Prior to getting the Garmin 830 I used a Wahoo Elemnt Bolt to record data, I am almost certain that was via Ant+. Looking at the Wahoo FIT file data I can see that Bit 8 is not set and hence not read into GC (after the change that was made to GC early last year).

I still have the Wahoo Elemnt Bolt, so if you want me to perform any tests please let me know.

[Yi-Ting Chen]

My power meter is ASSIOMA DUO. When I connect to the Bryton Aero60 head unit via ANT+, the GC can correctly display the L/R balance of the FIT file.

After using Bluetooth to connect to the Rouvy app of the iPad, the GC cannot parse the L/R balance of the FIT file exported from the Rouvy website.

Ale Martinez 在 2022年1月2日 星期日下午4:22:01 [UTC+8] 的信中寫道：

[MyJunk Junk]

Just wanted to add that Karoo (at least one of it's head unit) is also one of the companies that is not adhering to the FIT specs.There's more details here as I really went down the rabbit hole.

https://forum.intervals.icu/t/difference-in-left-right-balance/3627/27

To share, Garmin + Assioma Duo, when connected via ANT+ and BT, both works correctly. FIT is uploaded to Both Garmin and GoldenCheetah and LR balance data is displayed properly.  (however, there is a difference in how the ride's LR summary data)

Garmin shows 45/55 and GoldenCheetah is 51.7/48.3 (the only way I managed to get close to 45/55 is to omit when there's no pedaling, 100%L and 100%R data points)

If there's a need for some FIT files, links to dropbox are also in the above forum. 

[MyJunk Junk]

https://github.com/GoldenCheetah/GoldenCheetah/commit/5a93a2c62f68c7c1b36873d7897d3e319c08e706

Wondering if this will also fix the LR balance summary computation?