FreeTrack Forum
|
Hi!
I built myself a 3-Point-Cap from an old baseball hat, and it works really well. Now, to get the most out of it, I want to give FreeTrack its exact dimensions. My problem is, I can only measure two sides (b,c) and calculate a third one (a). I am no math genius, but I am stuck with two variables (x,y) that are codependent, but can be anything between zero and a because I don't know any angles and I don't see any solution to this.  So how do you get the most accurate values for x and y, which you have to enter in FreeTrack? Any ideas? Since the cap is in my way, simply using a ruler is not sufficient. Cheers, Zapp |
|
|
Class : Apprenti Off line |
|
|
@slaver1
Are you sure this link was meant for me? Because I have Win7 x64 and run FreeTrack 2.2.0.279 with a Microsoft VX3000 Camera and have absolutely no problems with it... I'm rather amazed how well it works  @dewey1 Thank you! That's exactly my problem. The only formula I know to calculate x and y is x²+y²=a² when I measure one side and calculate the other, I end up with a range of more or less x ~ 95 to 115 and y ~ 90 to 110mm So yeah, I guess you can say I suck at using a ruler  |
|
|
Here are three 3D models with variable heights and corresponding X-axis.
I am sure one of these should be close enough!  |
|
|
Class : Apprenti Off line |
No. Link is just in my sig.....  |
|
@slaver1
Ups. I totally didn't see your comment before But in the meantime I figured out that measuring one side and calculating the other one makes the error even worse. So I went to get some Chinese food the other day, built myself some right angles out of chopsticks and measured my entire head with them... and it really makes a difference in FreeTrack. So I recommend to anyone to spend some time into measuring your devices, because now, my head-movements are much more accurate than before Thanks for the help! |
|
|
I went with x=112 and y=102, which leaves me with an a=151.
But given the fact that I based my calculations for x and y on a value which already added two possible measuring errors, I am not surprised that the (hopefully) real values do not match the calculated ones. x y cAnd what I have learned from all this is, that Error Propagation really works and chopsticks are better than math |
|
|
zapp;
You can't fault the math. The real problem was the accuracy of the individual measurements. A tolerate of 10% is fairly large. If you had a 1% tolerance in your measurements, the results would have been much better. Here is the 3D model with your latest measurements.  |
> Stats
1 user(s) connected during the last 10 minutes (0 member(s) and 1 guest(s)).