Square snaps

[Dirk le Roux]

What program are you using...?

Hi Gordy

I trace the lines in Archicad and print pdf from there using Foxit Reader PDF Printer. Its default settings printing there is 'vector', not 'raster'. When you open the pdf you will see lines load as lines, not a bunch of pixels.

The workflow goes like this: import footage into Tracker and make sure that time and dimension settings are correct. Trace the line markers at regular intervals. I use rather large time intervals, which helps combat ripple. Copy the desired data to Excel and create the charts. The first chart I published for this thread has both the unsmoothed data and their moving average trend lines. The subsequent charts show just the values as copied from Tracker (I averaged between different marker tracks for each leg). For the fingerprint composites, I copy regular frames from Tracker into Archicad, trace and output as described.

Regards, Dirk

[gordonjudd]

I trace the lines in Archicad and print pdf from there using Foxit Reader PDF Printer.

Dirk,
I can imagine tracing all of those line shapes took a lot of tedious effort and admire your perseverance in producing such interesting overlays of two waves propagating on the same line .

I was hoping that you had found a way to pick out line shapes automatically as I have found edge detection methods don't produce useful results in getting isolated line shapes in MATLAB.

Thanks for all of your effort.

Gordy

[Dirk le Roux]

I will be most interested in the calculation you come up with to quantify the tangential velocity of the line going around a curved path.

I have devised a 2d equation which, given the incoming and outgoing legs' velocities, works provided the bend's path of travel is aligned with that of the markers on the incoming leg. If not aligned, the 1/2(V1-V2) or 1/2(V1+V2) type of approach becomes a complicated waste of time. Moreover, given that one will need for that simpler 2d scenario to measure at least the speed and direction of at least one point on each of the two legs, plus the direction of the bend's travel (to check validity), it will be much simpler to just measure the velocity of a point on either of the legs and that of the bend. The resulting bend-vs-leg or leg-vs-bend relative speed easily and more reliably give the bend's propagation speed and conversely the tangential speed (its magnitude here being the operative interest).

So far, the physics observations I’ve contributed along this experiment were, on the one hand, pointing out the tendency of the rear-most incoming leg’s trajectory dictating bends’ motion in space and on the other hand positing the question of the mid leg’s apparent levitation. I now add a few observations/musings.

Legs’ acceleration: With Graeme’s snap cast previously analysed, the fly leg continually accelerated upwardly (though with decreasing intensity). With the square snaps, the measured fly leg speed for a while decreased apace with the rate of the other legs’ decrease before gradually diverging until finally appreciably accelerating once the fly end has left the ground. Note that Graeme’s snap cast initially had a fly leg speed of virtually zero. In contrast, the square snaps did not, the fly leg already having had considerable speed by the time that the final acceleration by the rod tip was made.

Cheers,
Dirk

[Dirk le Roux]

Form frozen in space: What is it about the apparent frozen-in-space occurrence with snap casts? During an apparent freeze period, incoming (into a bend/feature) and exit leg velocities are equal or roughly equal in magnitude, meaning that upstream propagating loops and bends appear to be stationary or slow-moving in the earth frame.

Wave propagation: Features/disturbances originated from the rod end propagate upstream and are easy to observe, being seemingly frozen. Fly end originated disturbances (ripple from the curled slack on the ground) propagate downstream and as such in this scenario are more challenging to follow. Frame by frame, one can track the disturbances/waves travelling in both directions and observe that these do indeed propagate and do pass each other without difficulty. Continued propagation should mean continued line tension between the two ground contact points.

[Dirk le Roux]

The following observations may illuminate the mid leg’s apparent levitation after the rod input has completed:

• Studies have shown that on similar phenomena (leaping and falling chains and ropes) interactions at the ground pick-up and deposit points and the dynamics at direction changes are largely energy-conserving.

• Line markers’ height vs time plots deviate from parabolic, showing that a line particle’s motion while airborne during these snap casts is not that of a free projectile.

• Once the previously mentioned speed equalisation has occurred, the fly leg goes up as fast as, and over time faster than, the rod leg goes down. This has obvious geometric consequences for the bend positions and the portion of the line in between.

• Acceleration magnitudes measured as markers travel along the line’s configured path, especially where markers round bends, dwarf the gravitational acceleration g by up to 50 to 1.

• As the bends neither contract nor radiate in space, it follows that the large centripetal accelerations there are opposed by equally significant centrifugal effects.

[Dirk le Roux]

The non-problem of slack:

With these square snaps, the line already in motion appears to effortlessly extract from the stationary curls of slack line piled on the ground. This brought me to experiment with fishing casts (admittedly on grass so far – I haven’t had a chance to test these on water) performed from such an initial arrangement of slack. I have found that efficiency-wise a cast made from a pile of slack works surprisingly well, on the lawn at least on par with static roll casts. Even better when first making a backcast from a strategically placed pile before delivering the forward cast.

Lefty’s “God won’t let you make that cast unless you first get the fly moving” comes to mind. With these casts, the fly lies motionless long after my cast input was made. God does let me make them. With a 4-weight line in the confines of my back garden, I can from a piled 30 ft of fly line out the rod tip comfortably lift into a backcast, then forward shoot to accurately enough land the fly 50 ft out.

Why then the failures familiarly associated with casts started with a line having slack somewhere? I think the problem is not slack per se, but how it is configured. I think the slack configuration in the fail situations are discontinuous, such that the line in motion (having momentum) suddenly meets a significant section of straight line at rest and is unable to move all that mass virtually at once. With a small curl configuration such as you should be able to place with a piling snap, line in motion continually picks up only a small next segment of line.

All the best,
Dirk

[Graeme H]

Interesting stuff Dirk. Thanks for sharing it.

(I'm looking forward to getting back into the action soon.

Cheers,
Graeme

[Paul Arden]

Hi Dirk, the roll cast Is usually taught as a method of removing slack. And I agree that it’s interesting how with one cast (Roll) it’s fine to have and with another (overhead) it’s best not to have

I think is has to do with quantity, relative quantity and it’s position. For example slack positioned such that it interferes with the casting stroke is always a problem - assuming that it’s a significant amount. Also there becomes a point when the amount of slack being moved by the loop leads to a problem.

I never agreed with Lefty on this but maybe he’s up there discussing it with Odin right now

Cheers, Paul

[Lasse Karlsson]

Aitor did a good comparison between an overhead and a rollcast, the same energy needed to make the overhead roll out made the rollcast fail. I think its one of those things we tend to forget. Applying the same amount of force to an overhead cast, will remove as much slack line as it would in a rollcast.

Cheers
Lasse

[Paul Arden]

Never saw that but it’s an excellent example. A good question I had 25 years ago at my APGAI exam was do I need more or less force with a smaller D-loop. Ok it was a good question back then

Cheers, Paul