I'm noticing an issue and not sure if it's RaceRender or GoPro. When I import the video and GPS data, RaceRender is showing a weird triangle path that I didn't take at the start of the map and the speedometer is off. It's done this both as a single video import and as a collection. Is this a known issue? Is it RR or GoPro? Appreciate your help.
I'm running a Win 10 machine and RaceRender v3.7.3.
Which GoPro 8 firmware do you use?
Current is 1.6 (there is also a Labs firmware 1.7!).
I use 1.7 and do not have the problem.
Thanks for the reply.
Currently have 1.6. Just downloaded 1.7 and will give it a go.
Just for information - the LABS 1.7 firmware is more of a beta firmware ... it is not certain whether the existing features will be adopted in a later official version!
Changing the firmware from 1.6 to 1.7 did not help.
Here's a link to the screenshot: https://www.dropbox.com/s/b8c6mdfb0g...31.28.png?dl=0
I think there might be two possibilities here:Quote:
Originally Posted by sonicdj
1. This could be the GoPro synchronizing with the satellites before it can fully triangulate its position. It might not have a lock on all of the satellites when it first turns on. You do appear to have open sky, but there are a number of trees lining the road. I don't think this should be a big issue, but it might actually be causing the second possibility.
2. One (or more) of the GPS coordinates are recorded or extracted incorrectly causing RaceRender to think that nearly everything is in one location, but one or two rows are in another country. This ends up looking like a small map at one side and then a huge line to the row that is incorrect. RaceRender does not seem to be good at detecting these anomalies and just ends up showing a giant line to the one bad row instead of just ignoring it.
This thread lists some of the details, but not everything there may apply to you. Since you don't have laps, perhaps you could click the Laps button, then click Edit Lap Times in the pop-up, then delete everything in the textbox and add:
Lap 000: 00:00:10.000
to the lap times (or perhaps longer than 10 seconds to get beyond the coordinate[s] that cause the issue). This will cause RaceRender to interpret the first 10 seconds as staging time which don't count towards the map display.
Alternatively, you could try to open up the CSV in Excel or Google Sheets to try to find a row that is very far away from the rows around it for Latitude and/or Longitude. In this instance, very far away would be anything greater than 0.1 degrees away (and probably not more than 0.01). So if you see something that jumps from 41.5 to 0.2 and back to 41.5, then you could try just deleting the 0.2 row.
@HoboBob:Quote:
Originally Posted by HoboBob
I think this is another case of bad data sneaking in. As I wrote in that post (https://forum.hptuners.com/showthrea...1#post567086):
Since we can't assume perfection in hardware, the software -- RR -- should correct for this. It would be a whole lot easier than having the user open a spreadsheet each time a straight line appears as the path of travel, searching for bad data, and adjusting it. I mean, that kind of task is exactly the kind of work computers are good for.
This is still my opinion. That is, since a GoPro device is acquiring this data in real time, it's almost impossible (barring some tricky internal buffering calculations) for GoPro to catch any of its own mistakes. OTOH, RR can have a look at ALL the data as it imports it, and with some simple programming can figure out that some rows of data are way out of whack. At that point, if I were doing the programming -- and I'm a retired geek -- I would emit a dialog box, saying something like, "A few rows of data appear to be in error. OK to average out those rows, with the data appearing before and after?"
Do you know whether anyone on staff has considered such defensive analysis and programming, or am I talking to a wall here?
Weston seems to be the only staff member that posts here, and he doesn't seem to post much recently, but he has said that he reads the forum posts (also from your previous thread).Quote:
Originally Posted by wspollack
I would absolutely agree that this is a way that RaceRender could be improved and could help users who might be a little less technical. I guess that could be a deviation from a concept that RaceRender just accepts the input as given and doesn't try to be smarter than the incoming data, but I don't think that's the paradigm that RaceRender typically uses especially with regards to interpolating GPS data and trying to improve accuracy/timing. I get the impression that there is quite a bit of code regarding handling of GPS data, so I don't know how simple a feature request like this would be. Perhaps it's just a preprocessing step when importing the data originally? I've seen a number of posts requesting translating data from one unit to another; so some sort of preprocessing seems like it would be particularly helpful there too.
As a current young-ish software developer, I've thought about offering my services perhaps even pro bono even if just to have a better product that I can use myself. I'm not sure how my current employer might take to that since they have a bunch of rules about things like that...but I would absolutely love to be able to contribute if possible. RaceRender works really well for what I need it to do (i.e. rendering professional looking data on top of simple videos), and for the things that it doesn't, I write a script to clean up my CSV prior to importing, and that seems to work well enough for me. For less technical users, that might not be an option, and I'd hate for something so useful to die out. I suppose something else would pop-up in its place though.
I too am having problems with GoPro data files. After extracting and merging I get some bad satellite locks. These occurred when travelling in and out of tunnels in mountains. I immediately thought it was a GPS sat lock issue. If anyone has an 'edit-csv' script, I would like to give it a try. I was also looking at making several copies of the merged data and setting start and end times to omit the problem sections of the route, overlying several 'transparent' TrackMaps with different file 'inputs'. Not sure if that would work though. Any comments gratefully received (I'll make a new post of this as well ;-)
Paul
Attachment 101983