Modify .GPX elevation data?

[+rakeck]

I thought this would be a simple web search that would yield a number of paid and free apps, but ..

 

If I have a .gpx file from a Garmin device that contains elevation data (altimeter), and I'm looking to replicate real-world documented data as closely as possible, and I've determined that one "benchmark point" (summit) in the path file is 20 feet too high, how can I programatically modify all the elevation points in the gpx file to be the original elevation value, minus 20 feet?

 

I don't mind paying a reasonable amount for an app that works well.

 

Thanks

Edited June 30, 2013 by rakeck

[Colorado Bear]

Open the file with Excel, set up a function for the elevation - 20 then copy and paste special or past values. I did a seach on the Web and found several reference many of which are right here in the forums.

 

Colorado Bear

[Grasscatcher]

.....and then when you get through editing your data, you'll have only one point in the whole file that you "Know" is correct and all the rest will then be questionable. I would bet that many will actually be off more than the 20 ft "correction" you just made.

 

And really, you don't "Know" that the one "summit" point is correct, do you? Is it a Benchmark with "adjusted elevation"? Did you calibrate your unit just before you started?

 

There is a summit near me that all the topo maps and highway signs say is 13383 ft, yet at the very top of the summit there is an official Benchmark that says 13420 ft. Which is correct?

 

Also think of this....20 ft is easily within the tolerance range of what you should expect from your GPS.

[+rakeck]

I hear you grasscatcher, but I'm just interested in getting a profile image and associated text to "equal" what the world associates with a given pass ... the sign at the top. Yes, I'm finding a lot of published pass elevations don't match with current technology findings, but if a sign at the top says 10,000 feet, then the elevation profile peak should say 10,0000 just so as not to confuse people.

 

the altimeter on a modern GPS is still more accurate (and without the wild erroneous ground peaks and valleys) in a 15 minute time span than any google earth elevation profile you can create in mountainous areas with only low res imagery

[Grasscatcher]

My point was,.... and still is....., that you would have more accurate overall data if you only corrected/ changed the very few points at the actual summit, and not adjust the whole data set by a fixed uniform amount. A "fixed amount" adjustment would almost guarantee greater inaccuracy.

[+GeoTrekker26]

the altimeter on a modern GPS is still more accurate (and without the wild erroneous ground peaks and valleys) in a 15 minute time span than any google earth elevation profile you can create in mountainous areas with only low res imagery

I don't think that statement is true for a consumer grade GPS, especially those made by Garmin. Garmin still quotes 15 meters as the average positional accuracy for their devices. For altitude, Garmin says satellite calculated figures can be off by +- 400 feet. Others have posted in various forums that Garmin elevation calculation are less accurate that other brands.

 

This plus/minus variance happens at each reading, so to make an adjustment of readings for several locations based on what was observed on one location is absolutely wrong. Should you actually do that, you should tell anyone that might use your data that the numbers are crude approximations only and have been made meaningless by unfounded manipulation.

 

If each GPS was consistently off by the same margin at every reading, the device could be calibrated once and there would never be any error. That isn't the case.

[+JBnW]

Following on with GeoTrekker, here is an altitude graph collected by a 62s during a down-and-back river run in a boat last week. The river is wide enough to offer clear, unobstructed views of the sky, fresh batteries, GPS calibrated, etc. etc. etc.

 

 

Now as fun as it may sound going over 60 foot waterfalls in an airboat as suggested in the graph, in reality there are no such features in this river. It was all flat, calm water. There are three locations along this stretch with "known" elevations at USGS Gauging Stations; which would be six total matching points (three down and three back). Needless to say, comparing those with measured elevations is an evaluation of randomness. So, simply changing all elevation points in a dataset an equal amount so that one point matches a published elevation does not address the inherent error of the entire dataset.

[+rakeck]

This is what i get with my garmin. The profile is fine, and I understand there may some 'errors' but overall the device logs beautifully smooth and I'm OK with the results but the sign at the top of trout Creek says 9,487 feet. I use WAAS and I've learned that the altimeter needs to be bypassed when operating at significantly fast elevation changes (above 4.5%/60 MPH) or the results will be wildy off. I'd still like to modify all the points to get me to an image that reflects 9487. On balance the elevation rise is more or less accurate and the dips and such are more or less accurate. I'm not trying to go to the mars with this project, i simply want to edit my recorded data, put it into a profiler, and get an end result that matches with the real world summit sign. I've messed around with TrackProfiler but it doesn't want to display anything in feet/miles even though I've modified the option setting, and when I attempt a global elevation change, it doesn't seem to do anything.

 

Edited July 13, 2013 by rakeck

[+rakeck]

I guess I'll do it this way it works but is slow.

Using GPSVisualizer ...

 

With utility, convert my .gpx to .txt file on my desktop

open the .txt file with Excel, make necessary elevation edits, remove temp column data, re-save (keeping .txt format)

Convert the .txt file to a .gpx file with utility

Save .gpx file from web into a different folder on my PC

Done

[Grasscatcher]

This is what i get with my garmin. The profile is fine, and I understand there may some 'errors' but overall the device logs beautifully smooth and I'm OK with the results but the sign at the top of trout Creek says 9,487 feet. I use WAAS and I've learned that the altimeter needs to be bypassed when operating at significantly fast elevation changes (above 4.5%/60 MPH) or the results will be wildy off. I'd still like to modify all the points to get me to an image that reflects 9487. On balance the elevation rise is more or less accurate and the dips and such are more or less accurate. I'm not trying to go to the mars with this project, i simply want to edit my recorded data, put it into a profiler, and get an end result that matches with the real world summit sign. I've messed around with TrackProfiler but it doesn't want to display anything in feet/miles even though I've modified the option setting, and when I attempt a global elevation change, it doesn't seem to do anything.

 

 

NO.....that chart is NOT what you get with your Garmin.....

That is what you get with your Garmin data when plotted with "whatever" software that has however many unknown filters to manipulate the data .

 

"..... and when I attempt a global elevation change, it doesn't seem to do anything.".... See, even the software knows that you shouldn't do that!

[Grasscatcher]

I guess I'll do it this way it works but is slow.

Using GPSVisualizer ...

 

With utility, convert my .gpx to .txt file on my desktop

open the .txt file with Excel, make necessary elevation edits, remove temp column data, re-save (keeping .txt format)

Convert the .txt file to a .gpx file with utility

Save .gpx file from web into a different folder on my PC

Done

 

And at that point you will have a totally INACCURATE dataset that has absolutely NO relationship to reality......but you will be satisfied and happy, and of course, that is the most important thing.

 

Enjoy your exercise in futility.....

[puzz]

... I've messed around with TrackProfiler but it doesn't want to display anything in feet/miles even though I've modified the option setting, and when I attempt a global elevation change, it doesn't seem to do anything. ...

 

I've fixed this bug on trackprofiler. The graph now will be in miles and feets (you can try even nautical miles if you want). See for example here or here.

 

If you want to add a fixed value to the elevation, login go to your track, select "Elevation errors" and then "Add/substract elevation", ant then write for example +30ft or -30ft.