One more MegaDrum firmware to test

[Rubis]

I am putting the finishing touches on my first drum now, but haven't gotten the settings working the way I want them just yet. Once I figure this stuff out, I will test this.

[Rubis]

Here it is. I took 20 measurements for each position, from all around the drum. My trigger is fully DIY, but I think most of the inconsistencies are from me, not the drum. It has a layer of pet screen sewed around 1/4" copper tubing with another layer simply draped over the top for the head. The rims are made from 3/4" birch plywood. Cone is 3 layers of 1/2" very soft poron foam, cut with an electric knife and then sanded.

EDIT: forgot to attach the data.

EDIT 2: just realized that my data is for the old test firmware. I will try again later.

[elrules]

Seeing Rubis testing, it could make sense if positional sensing was based in just a threshold. For example, for him, a threshold of 40 will give him more or less consistent differentiation between center and edge hits, and Superior 2.0 or BFD 2.0 could trigger different samples.

[Rubis]

Seeing Rubis testing, it could make sense if positional sensing was based in just a threshold. For example, for him, a threshold of 40 will give him more or less consistent differentiation between center and edge hits, and Superior 2.0 or BFD 2.0 could trigger different samples.

Did you check the other pages? That may hold true for hard hits, but medium and light hits return much lower values.

Also, I set retrigger to 20, and min scan to 50 for these tests, in case that makes a difference.

EDIT: just realized that my data is for the old test firmware. I will try again later.

[elrules]

I have modified the document you sent to take the average value for each test (light, medium, hard). Then each underlined value would be a wrong triggering (if we set a threshold to that average value, for each velocity range) and value for edge hits is higher than threshold and value for center hits is lower than threshold.

What I propose is to define 3 different thresholds for three different velocity ranges. Or even being able to assign a curve to it (instead of several fixed values), for example for your experiment, 11, 33, and 38 are the threshold values for each velocity => that will fix in a Log curve.

I found by setting those values for your experiment, that most of the hits will trigger correctly the right position (center or edge snare). Don't worry if not all the hits triggers correctly, even with my Roland TD-12 does not trigger nice all of them

[Rubis]

The drum I tested this on was my first attempt at building one from scratch and I used the old test firmware by accident. If my sloppy drum and drumming skills can get decent detection already, I think this may be very doable.

The drum I used in that test had no separation from the rim whatsoever. Vibrations from the head>rim>bottom>piezo may be causing a slight interference, not to mention I don't have any kind of rack or stand for it yet (I was just holding it which probably threw it way off). I should have some more drums built soon with a tweaked design that should work a little better. I will test the newer firmware then.

I use a log curve on the drum. Do you think that the curves would be the same for positional sensing and velocity?

[elrules]

Well, let's wait to see what dmitri thinks

[Rubis]

Dmitri, when using the second test firmware, do you want us to mess with SnareVar rim and center values to get it as even and accurate as we can, or do we leave it at the defaults so you have a consistent set of data to work with?

[dmitri]

It doesn't matter right now. The data is inconsistent right now and varies too wildly between different pad makes to an extent that I cannot put into production firmware and claim it supports it. I wonder if Roland modules get consistent results with only specific type of pads and mileage may vary with any others?

[lajox]

It doesn't matter right now. The data is inconsistent right now and varies too wildly between different pad makes to an extent that I cannot put into production firmware and claim it supports it. I wonder if Roland modules get consistent results with only specific type of pads and mileage may vary with any others?

I had an old (TE3.0 / 3.1) Hart Dynamics ProSnare sometime ago and positional sensing wasn't working very well with a TD8 and later a TD20. If I recall correctly Hart mentioned somewhere in their documents that PS is not supported. Found this post in the vdrums forum http://www.vdrums.com/forum/showpost.php?p=161996&postcount=2