Printable View
New user... Bought deluxe today.
I took a multi-part gopro camera video and imported it into RR. It shows the "track" as a straight line. If I take a single file it will show me an actual path.
Is this a limitation of the file size / data points in the csv file?
Is there a better way to superimpose a GPS path onto my videos?
Thanks!
HW
I've seen that happen when my data file was 5-10+ hours or when I tried using the RaceRender Data File Format features without starting the CSV with "# RaceRender Data".
If you're not seeing the issue when importing individual files, then it's probably easiest to import them individually and just synchronize them appropriately (or render separate videos and combine them later).
HLW,
What are the end points of the "straight line"? I believe that you will find that there are GPS data points with very low "accuracy". You will likely find these at the end point of your "straight line" some distance from where you actually were. It looks like a straight line because all the "real" GPS points are concentrated at the other end from the few erroneous data points. The GPS accuracy for each point is embedded in the GPS data from your Go-Pro but Race Render does not use it. The single file(s) that shows the proper track do not have "bad" GPS data in them.
I found the same problem and spent a considerable amount of time to figure out what was happening. I made a request (several months ago) that RR needs to parameter that automatially IGNORE GPS data points that do not meet a certain accuracy but have not received any acknowledgement of my request or suggestions from the developers of RR on how to deal with it.
Glenn
I've had this happen to me when RR has selected my fastest lap from a multiple lap session. However, that lap may have been a short lap or possibly a short segment from pit out to the start/finish line. Double click on the "track map" under "Display Objects". That will bring up a Display Object Properties window. Next to the Draw Selected Lap, select a different lap to be displayed.
Good catch! I just got one where my "best" lap was actually just a sector time, and selecting "Lap 1" to be displayed instead of "Best" caused the correct map to be drawn.Quote:
Originally Posted by WhtnWrench
Background: RaceRender, Hero4, iPhone
I've been using (Ultimate) RaceRender (RR), along with TrackAddict (TA), both purchased several years ago.
In those old days, I'd use a GoPro Hero4 Silver for motorcycle rides and the occasional track day. I'd use my iPhone6 for the GPS data acquisition, with TA running on the phone. Back at home, I'd use RR to generate the videos, with the output from the GoPro and TA. Never had any problem, except that the iPhone is a little slow to react to, say, speed changes.
Background: Current:
Lately, I've been testing with a new Hero7 Black (and I have to say, with built-in GPS and "HyperSmooth," new software-based stabilization, the Hero7 is an excellent improvement vs. the Hero4). This is nice, because that GPS chip means that I no longer need to use my phone at all (which would have been running TA).
This setup worked fine, without incident, for a few weeks.
Hit the Straight Line:
For the first time, a couple of days ago, I encountered this straight-line problem.
I typically use a RR template that lists speed, altitude, and the track map (which, when not on an actual track, is just the path of travel). That day trip from a couple of days ago included a few breaks and a lunch stop. Using the data downloaded from the Hero7, I was able to generate final videos in RR without incident for the morning and late afternoon.
However, for one of the in-between riding chunks, I got that straight line for a path of travel. Take a look at the upper right corner of the first attachment here, which is a screen shot of the beginning of the final video generated via RR (and that travel path doesn't change, throughout the video).
Found the Problem:
Thanks to @HLW starting this thread, and @GChong providing significant analysis, I knew what to look for.
I saved that project, and quit RR (in order not to possibly have a file opened by two programs at the same time). Next, I opened the projects data file -- in my case, called GX010009-RR_Data_Merged.csv -- in LibreOffice (although any spreadsheet program will do).
I then sorted the data, looking for the high and low values of latitude and longitude.
The outliers for both latitude and longitude (and altitude, but not speed) were contained in one row ... about a third of the way through 39,209 data rows.
(This was a 36-minute, 20-second, video, so we're talking about a file entry every 0.055 seconds or so, if my math is correct.)
The values for the time, latitude, longitude, altitude, and speed for this row were 1555181476.46056, 0.000001, 0.977798, 25.32, and 28.499. The second attachment here is a screen shot of that area of the .csv file, so you can see a bunch of the normal values before and after that errant row.
Fixed This Particular Problem:
Rather than delete that row, and possibly alter the synchronization, I took the average of the prior and following cells, and put those values in the columns for latitude, longitude, and altitude.
The third screen shot shows this change.
I saved the altered .csv file, quit LibreOffice, started up RR, and opened that saved project.
The travel path was no longer that fictitious straight line. And then I went on to successfully generate a final video, with the template widgets included; a screen shot from that final video is the last attachment here.
The Future?:
What @GChong suggested (see post #3) seems to me to be the proper solution here, i.e., have a way for RR to ignore data that's pretty obviously wrong. I mean, that one bad row was probably the result of a firmware or hardware glitch in the Hero7.
(I don't know how to look at the data in the camera's .MP4, but I'm assuming that RR does this extraction correctly. And I went through this entire process, described above, several times, using saved-off data, new projects, etc., and each time the .csv file was generated the same way.)
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.
In my opinion, there should be some user-settable limits, such as the latitude or longitude changing by more than X (e.g., 1) degree between successive rows, and the altitude changing more than Y meters, and the speed changing by more than Z m/s, that sort of thing.
Further, when RR encounters these conditions, it should let the user know that nnn rows were corrected, set to the averages of adjoining cells, something like that.
The Support:?
When I first bought RR and TA, the forum support seemed much quicker to me. I haven't been on the forums much, until I got my Hero7, but reading a lot of threads, I agree with @GChong in that support seems to be lacking.
Maybe the real money -- and hence time spent -- is in the hardware side of things. When I first made my purchases, this was pretty much a software-only outfit. Looking around now, there's a lot of hardware -- car add-on devices -- for sale. That may be good for business, but perhaps HP Tuners should invest some of the profits into hiring more support staff, in terms of both the forum and (software) product maintenance and improvements.