Hello: I've downloaded the FIT SDK 21.38.00 and I'm trying to work with the decode example. I'm running in Linux and I've managed to compile the C++ example using g++ compiler and build the library as $ cd FitSDKRelease_21.38.00/FitSDKRelease_21.38.00/cpp $ for i in `ls *.cpp`; do g++ -Wall -O3 -c $i; done $ ar rcs libcppfit.a *.o After that I have the static FIT library called libcppfit.a in the FitSDKRelease_21.38.00/FitSDKRelease_21.38.00/cpp/ folder. Then, I compile the decode program as (the order is called from the same folder as before) $ g++ -Wall -O3 -I. -L. examples/decode/decode.cpp -o decode -lcppfit I have now a working decode binary. In this link ( drive.google.com/.../1WlbPG-4KCB_dSAnks4vztoAp8ZHyC_TN you can download two FIT files: file-garmin-edge530.fit: file logged with a Garmin Edge 530 with the cycling dynamics activated and with a Favero Assioma DUO powermeter file-van-rysel500.fit: file logged with a Var Rysel 500 GPS with the cycling dynamics activated and with a Favero Assioma DUO powermeter When I convert these files with the decode binary I obtain strange results related to the left_right_balance field. For the file-van-rysel500.fit I obtain logical results as, for example, Field10 (left_right_balance) has 1 value(s) Val0: 49 Field10 (left_right_balance) has 1 value(s) Val0: 51 Field10 (left_right_balance) has 1 value(s) Val0: 48 But for the file file-garmin-edge530.fit I obtain values like Field11 (left_right_balance) has 1 value(s) Val0: 176 Field11 (left_right_balance) has 1 value(s) Val0: 181 Field11 (left_right_balance) has 1 value(s) Val0: 183 These values are clearly wrong as it must be decoded with the mask 0x7F, and after that I obtain (in Python): >>> 176&0x7F 48 >>> 181&0x7F 53 >>> 183&0x7F 55 wich are correct values. So, the question is: is it a bug in decode.cpp program? How it is possible the difference in decoding, and the Garmin FIT file the wrong answer? Thanks

Running cpp/examples/decode/decode.cpp program and strange behavior with two FIT files from different devices

Hello:

I've downloaded the FIT SDK 21.38.00 and I'm trying to work with the decode example. I'm running in Linux and I've managed to compile the C++ example using g++ compiler and build the library as

$ cd FitSDKRelease_21.38.00/FitSDKRelease_21.38.00/cpp
$ for i in `ls *.cpp`; do g++ -Wall -O3 -c $i; done
$ ar rcs libcppfit.a *.o

After that I have the static FIT library called libcppfit.a in the FitSDKRelease_21.38.00/FitSDKRelease_21.38.00/cpp/ folder.

Then, I compile the decode program as (the order is called from the same folder as before)

$ g++ -Wall -O3 -I. -L. examples/decode/decode.cpp -o decode -lcppfit

I have now a working decode binary.

In this link (drive.google.com/.../1WlbPG-4KCB_dSAnks4vztoAp8ZHyC_TN you can download two FIT files:

file-garmin-edge530.fit: file logged with a Garmin Edge 530 with the cycling dynamics activated and with a Favero Assioma DUO powermeter
file-van-rysel500.fit: file logged with a Var Rysel 500 GPS with the cycling dynamics activated and with a Favero Assioma DUO powermeter

When I convert these files with the decode binary I obtain strange results related to the left_right_balance field. For the file-van-rysel500.fit I obtain logical results as, for example,

   Field10 (left_right_balance) has 1 value(s)
       Val0: 49 
   Field10 (left_right_balance) has 1 value(s)
       Val0: 51 
   Field10 (left_right_balance) has 1 value(s)
       Val0: 48

But for the file file-garmin-edge530.fit I obtain values like

   Field11 (left_right_balance) has 1 value(s)
       Val0: 176 
   Field11 (left_right_balance) has 1 value(s)
       Val0: 181 
   Field11 (left_right_balance) has 1 value(s)
       Val0: 183

These values are clearly wrong as it must be decoded with the mask 0x7F, and after that I obtain (in Python):

>>> 176&0x7F
48
>>> 181&0x7F
53
>>> 183&0x7F
55

wich are correct values.

So, the question is: is it a bug in decode.cpp program? How it is possible the difference in decoding, and the Garmin FIT file the wrong answer?

Thanks

Top Replies

Ben FIT 17 days ago in reply to jgpallero +2

The decode.cpp example just dumps all the field names and values for a mesg to the console. The example does not manipulate the values in anyway. You can see this in the PrintValues(FieldBase&) method. There…
jgpallero 17 days ago in reply to Ben FIT +1

Thank you for your answer. I reply here also your answer in the other thread (forums.garmin.com/.../questions-about-left-right-balance-and-its-decodification).

If (left_right_balance & 0x80 == zero), then…

All Replies

0 Ben FIT 19 days ago

See my answer to your other question here on how to decode left/right power values.
RE: Questions about left/right balance and its decodification

0 jgpallero 18 days ago in reply to Ben FIT

Thank you for your answer. But in this case I have not modified the file decode.cpp, it is the one provided with the FIT SDK. And this file produces right values for left_right_balance for one file and wrong values for other. And curiously, the right values (49%, 51%, 48%) are the ones stored in the Van Rysel FIT file and the wrong values (176%, 181%, 183%) are the ones stored in the Garmin file.

0 Ben FIT 17 days ago in reply to jgpallero

The decode.cpp example just dumps all the field names and values for a mesg to the console. The example does not manipulate the values in anyway. You can see this in the PrintValues(FieldBase&) method. There is nothing specific about the left_right_power fields in that method. It prints out the value as is.

If the left_right_power fields are of interest, then you can add code to the OnMesg( fit::RecordMesg& record ) method to apply the masks.

For the example values from each file that you provided, when the masks are applied the values look as expected:

49 & 0x80 = 0 = left side
49 & 7F = 49% left side

176 & 0x80 = 0x80 = right side
176 & 0x7F = 48 % right side

0 jgpallero 17 days ago in reply to Ben FIT

Thank you for your answer. I reply here also your answer in the other thread (forums.garmin.com/.../questions-about-left-right-balance-and-its-decodification).

If (left_right_balance & 0x80 == zero), then the value is the balance for the left leg

Are you sure of that? As I understand, this is contrary to what standard says. In the file Profile.xlsx of FitSDKRelease_21.38.00 (and in prior versions) we can read:

left_right_balance uint8
    mask  0x7F   % contribution
    right 0x80   data corresponds to right if set, otherwise unknownCode

This means, as I can interpret, that if left_right_balance & 0x80 == zero the value is not the balance for the left leg, but it is unknown, i.e., useless.

For example, for the two files I gave as example (drive.google.com/.../1WlbPG-4KCB_dSAnks4vztoAp8ZHyC_TN), if I upload them to GarminConnect I obtain correct balance values for the data contained in file-garmin-edge530.fit, i.e., for the file generated with Garmin device. But for the file-van-rysel500.fit the balance values are omitted, i.e., Garmin Connect discards them as left_right_balance & 0x80 == zero when it reads the file, so Garmin Connect applies what the FIT standard says in Profile.xslx