Results 1 to 11 of 11
  1. #1
    mtbr member
    Reputation: 3034's Avatar
    Join Date
    Apr 2006
    Posts
    746

    training center vs motionbased comparing 3 years of data

    I ran a motionbased total since its going away soon
    for 3 years here's some totals (75%mountain 25% road)

    training center motionbased
    miles 8728 8776
    hours (moving) 912 1267 (mb must be total not moving)
    ascent 979561 1136025
    descent 992193 1137573
    avg speed 9.6 9.0

    for ascent training center adds 10% to MB adds 20% to the numbers the edge unit shows. Inflated resume

  2. #2
    trail rat
    Reputation: slocaus's Avatar
    Join Date
    Jul 2005
    Posts
    7,649
    MotionBased does some distance and elevation corrections that Training Center does not. GTC has no ability to do corrections that I can see, nor does it have any way to choose the corrections method and amount. TopoFusion is the best at this, in terms of choosing the method and amount of correcton. SportTracks has the ability with a plugin, and seems to be ok, but TopoFusion is my go to.

    MotionBased is known to be way over inflated, and Connect does not have the controls, but seems more accurate when looking at its reports compated to TopoFusion.

    MTBR previous discussion

    TopoFusion explanation
    CCCMB web - Facebook
    SLO trail maps - conditions

    "The physician heals, Nature makes well" - real fortune cookie

  3. #3
    Rep Power: Infinity
    Reputation: NateHawk's Avatar
    Join Date
    Dec 2003
    Posts
    11,198
    The values reported on the GPS are not the "true" values, because if you use GPS altitude, the accuracy is pretty poor. If you use a barometric altimeter, the reported altitude will "drift" over time from weather systems moving through, large changes in elevation, and that sort of thing. The more time the ride takes, the more drift the receiver experiences.

    I honestly don't know how MB Gravity did its thing, but I do know that it completely ignored the reported elevation values and replaced them with its own. Where it got them...I dunno, probably from some DEM somewhere.

    Unless you go out there with survey gear, I think pulling elevation values from a specified set of coordinates off of a DEM is going to be the most accurate solution. I can't speak for training center, as I don't use it.

    I do know that GC is or will be using some correction algorithm, and it's possible that Training Center does something similar.

  4. #4
    mtbr member
    Reputation: 3034's Avatar
    Join Date
    Apr 2006
    Posts
    746
    I disagree, The raw data from the edge is more accurate than motionbased and training center. the DEM is not acccurate around trees and steep lopes/canyons. I ride in the Sierra foothills (Calif) and
    its gets pretty sketchy.

    If you derive elevations based on horizontal location overlaid on a DEM model then your vertical accuracy is dependent on horizontal accuracy as well. To many places for errors to add up
    such as trees, canyons,cliffs, poor dem, etc

    my ridding buddy uses an altimeter bike computer and we agree very closely on vertical data when we ride. this is raw data in the edge.



    Quote Originally Posted by NateHawk
    The values reported on the GPS are not the "true" values, because if you use GPS altitude, the accuracy is pretty poor. If you use a barometric altimeter, the reported altitude will "drift" over time from weather systems moving through, large changes in elevation, and that sort of thing. The more time the ride takes, the more drift the receiver experiences.

    I honestly don't know how MB Gravity did its thing, but I do know that it completely ignored the reported elevation values and replaced them with its own. Where it got them...I dunno, probably from some DEM somewhere.

    Unless you go out there with survey gear, I think pulling elevation values from a specified set of coordinates off of a DEM is going to be the most accurate solution. I can't speak for training center, as I don't use it.

    I do know that GC is or will be using some correction algorithm, and it's possible that Training Center does something similar.

  5. #5
    Rep Power: Infinity
    Reputation: NateHawk's Avatar
    Join Date
    Dec 2003
    Posts
    11,198
    Do you realize the errors inherent in ANY elevation calculation? Sure, there are errors in a DEM (they're calculated based on aerial photo stereography, btw), but I posit that the errors on the GPS are greater. I've already detailed the drift from the barometric altimeter. How often do you calibrate your barometric altimeter? Moving speed makes a difference, too...how fast can the altimeter adjust to your change in altitude? Using GPS altitude has even greater potential for error. Just because that much error is inherent in the method of calculation.

  6. #6
    mtbr member
    Reputation: 3034's Avatar
    Join Date
    Apr 2006
    Posts
    746
    I was thinking of way to really test this variation in data (solo ride today - lots of time to think)
    I was thinking of using a 10' long pvc pipe, fasten the garmin edge to the top, go out in an open area and raise, lower the pole 20 times (tilting it ) to get 200' of vertical in the unit.
    perhaps move a little horizontally between lifts to sweeten the data

    then compare raw edge, MB , GC and training center.

    anyone have any suggestions before I run my test?

    ps Natehawk, I'm land surveyor by profession so I'm familiar with measurement accuracy and GPS.

  7. #7
    Rep Power: Infinity
    Reputation: NateHawk's Avatar
    Join Date
    Dec 2003
    Posts
    11,198
    Quote Originally Posted by 3034
    I was thinking of way to really test this variation in data (solo ride today - lots of time to think)
    I was thinking of using a 10' long pvc pipe, fasten the garmin edge to the top, go out in an open area and raise, lower the pole 20 times (tilting it ) to get 200' of vertical in the unit.
    perhaps move a little horizontally between lifts to sweeten the data

    then compare raw edge, MB , GC and training center.

    anyone have any suggestions before I run my test?

    ps Natehawk, I'm land surveyor by profession so I'm familiar with measurement accuracy and GPS.
    I'm studying GPS and GIS in graduate school, so I'm no stranger, either. Been using them since SA went off in 2000.

    Using a method where you KNOW (based on measurements/hand calculations) and comparing is a good idea, but I'd modify it a little to generate more elevation change (would be easier to compare, I think). I'd probably double the length of the pole, and raise/lower enough times to get 1000' of vertical. You should probably hold the pole in each position for at least 30 sec, too, to make sure the GPS has 'settled'.

  8. #8
    trail rat
    Reputation: slocaus's Avatar
    Join Date
    Jul 2005
    Posts
    7,649
    Quote Originally Posted by 3034
    I disagree, The raw data from the edge is more accurate than motionbased and training center. the DEM is not acccurate around trees and steep lopes/canyons. I ride in the Sierra foothills (Calif) and
    its gets pretty sketchy.

    If you derive elevations based on horizontal location overlaid on a DEM model then your vertical accuracy is dependent on horizontal accuracy as well. To many places for errors to add up
    such as trees, canyons,cliffs, poor dem, etc

    my ridding buddy uses an altimeter bike computer and we agree very closely on vertical data when we ride. this is raw data in the edge.
    I'd love to see some reference links to support this. I am not trying to be a smart azz; I have read all I can find on GPS accuracy over the last four years of owning my three, and am not a professional, just an extremely curious (ok, maybe obsessed) user.

    One thing I have learned is that there are many theories on how to correct the raw data to get more accurate results, and all have their shortcomings, of course. I have tried using the different methods in TopoFusion and SportTracks, where I can change settings, and then comparing to known mapping points for actual elevations vs my GPS readings. I will recalibrate my 60CSx at known points on rides, but the 705 has no calibration abilities, but I do like to look at what it reads when I have a known point, especially a benchmark.
    CCCMB web - Facebook
    SLO trail maps - conditions

    "The physician heals, Nature makes well" - real fortune cookie

  9. #9
    Rep Power: Infinity
    Reputation: NateHawk's Avatar
    Join Date
    Dec 2003
    Posts
    11,198
    Quote Originally Posted by slocaus
    I will recalibrate my 60CSx at known points on rides, but the 705 has no calibration abilities, but I do like to look at what it reads when I have a known point, especially a benchmark.
    My 76CSx has a setting that lets the GPS use GPS altitude to calibrate the barometric altimeter. I use it because I usually don't keep track of benchmarks for calibrating. It's not perfect, but keeps things somewhat close, anyway. Does the Edge 705 use something like this automatically?

  10. #10
    trail rat
    Reputation: slocaus's Avatar
    Join Date
    Jul 2005
    Posts
    7,649
    Quote Originally Posted by NateHawk
    My 76CSx has a setting that lets the GPS use GPS altitude to calibrate the barometric altimeter. I use it because I usually don't keep track of benchmarks for calibrating. It's not perfect, but keeps things somewhat close, anyway. Does the Edge 705 use something like this automatically?
    I cannot find any info. I have tried Google, cruised the MotionBased, GroundSpeak, and new Connect forums, no info. I would guess it does.

    There is a known issue with the 705 calibrating slowly, many seem to think it can be addressed in firmware, but has not happened yet. I turn mine on and leave it on a table outside for 15-20 minutes before I ride, and that seems to work.

    On my 60CSx, I use the manual calibration to a know elevation when I head from home, or at trail heads where I can get a known elevation. A few rides I do have benchmarks that I have located, just for the purpose of checking the 705 and recalibrating the 60CSx. Yes, I am a geek.
    CCCMB web - Facebook
    SLO trail maps - conditions

    "The physician heals, Nature makes well" - real fortune cookie

  11. #11
    Rep Power: Infinity
    Reputation: NateHawk's Avatar
    Join Date
    Dec 2003
    Posts
    11,198
    Quote Originally Posted by slocaus
    There is a known issue with the 705 calibrating slowly, many seem to think it can be addressed in firmware, but has not happened yet. I turn mine on and leave it on a table outside for 15-20 minutes before I ride, and that seems to work.
    Minor annoyance for me, I suppose. It usually takes about that long for me to get ready. Guess I'd just have to remember to start up the GPS first thing. Sounds like it might be doing some kind of auto calibration, but my 76CSx only takes a minute or so to start up and get running. Not sure what the Edge would be doing to take so long.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •