|
Jul 2013
3:36pm, 12 Jul 2013
271 posts
|
Moleshome
Relax, take a chill pill, get the kids, have a great weekend.
|
|
Jul 2013
11:01am, 16 Jul 2013
287 posts
|
Moleshome
Bump.
Weekend over, back on your heads.
|
|
Jul 2013
5:08pm, 16 Jul 2013
13,229 posts
|
fetcheveryone
I did 10 x 400m intervals this morning, so it was the ideal time to see how the speedmarks behaved.
I've made two changes, which should help. I should reiterate that when you do a stop-start run, as I did with the intervals, and when you couple that with running the exact distance of one of the speedmarks, it's hard to get 100% accurate results, because of the way in which the data in the TCX is presented.
That said, I've tightened things up a bit.
1) The code steps through every lat/long pair in the data, working out the distance from each one to the next, and stuffing the points into a buffer until it reaches the distance it's interested in (let's say 400m for the sake of this example). When it's collected 400m of route (which in reality will be more like 409.89358945 metres), it looks at the start and end time of that segment, to work out the equivalent 400 metre pace (400/409.89*time). If that beats the best time on that route, it stores it as the new best. In either case, it can then ditch the first point in the buffer. HOWEVER, it occurred to me that it can *keep* dropping the first point in the buffer *until* the total distance stored falls below 400m. In the example, if the first point accounted for 7 metres, and took 2 seconds, then we'd still have 400m in the buffer, but we'd have a time that was two seconds faster.
2) Secondly, when I've done processing the individual points on the route, I clear the buffers and pass in the lap information. So from here on, if your laps say you ran an 8:24 mile, then you can expect the speedmarks to show that you've been at least that fast.
Returning to my initial apocalyptic warning, TCX files contain [Track] nodes (usually one for each lap, but if you stop/start your Garmin it also generates them). When a [Track] is started, the TCX records the start time, but not necessarily your location at that exact point. So although it can report a lap time of X:YZ, there's not always exact evidence to back it up. To give you an example, my fastest 400m today was 1:44, but the [Trackpoint]'s inside that particular [Track] start at 2514 metres, and finishes at 2897 metres. The next [Trackpoint] is inside the next [Track], which is me recovering and not running very fast at all.
|
|
Jul 2013
7:54pm, 16 Jul 2013
3,724 posts
|
paul the builder
Why the "409.89358945 metres"? Just because lat/long pairs aren't especially close, and that's the next one that's over 400m from the starting one? Or is there something more that I'm missing?
Seems an awful lot of decimal accuracy to write down, that's all 
|
|
Jul 2013
8:20pm, 16 Jul 2013
13,240 posts
|
fetcheveryone
409 - yes, points in the file are far enough apart to produce this, sometimes even more.
Lots of DP - perhaps not relevant to my point, but in general I avoid rounding until the last possible moment.
|
|
Jul 2013
12:34pm, 17 Jul 2013
292 posts
|
Moleshome
So how does the process work?
Presumably the importer talks to the GCP, what does the GCP return? A TCX file, a FIT file?
What happens when you get footpod data, presumably that has no lat/long pairs?
|
|
Jul 2013
12:47pm, 17 Jul 2013
293 posts
|
Moleshome
PS. If I've got my Garmin set to one second recording then it's storing points a lot closer together than 400m. Does GCP give you all those points or does it summarise them in some way?
|
|
Jul 2013
12:57pm, 17 Jul 2013
3,727 posts
|
paul the builder
400m! Think you misunderstood Fetch's post. That implies that points can be something like 10m apart, so the next one in a *long series* might take you to 409m, the previous one in the series having been a little under 400m from the start.
|
|
Jul 2013
2:18pm, 17 Jul 2013
294 posts
|
Moleshome
Doh. Yes, I misunderstood your comment, the large data point is the additional 9m not a single 409m segment.
|
|
Jul 2013
2:45pm, 17 Jul 2013
414 posts
|
Toadie
This is working for me now, by the way, and it's fun! Also I didn't realise until the other day that I could now use the garmin communicator plugin with my 110, so I've been enjoying that. Much better. Thanks!
|
