Page 1 of 1

Real Drum like response from e-pads

PostPosted: Wed Sep 09, 2009 8:21 pm
by jmcdougall
Like others here, I am not a professional muscian but rather a musical hacker/hobbyist. Over the past while, I have seen much discussion about positional sensing and how to deal with rim/bow/edge etc. I think it is possible to get to real life-like drum response with e-pads but serveral things have to change and much of it is not pad related. Here are the changes that I feel have to take place:

1) The midi spec have to move from its current 32K/sec to at least 128k/sec. The current standard is way to slow to handle the data volume that realistic drum response would require:
2) Peizo technology is not the most effective way to meet the requirements. A pad has to generate two signals - one beeing the force of the impact and the second being the distance from center of the impact. Note I am NOT saying positional. I think the focus on XY positioning is irrelevant. My perspective being that a drum head is normally an equal tensioned surface. As you move from the center to the edge, there is an increase in pitch as the displacement of the head is reduced and therefore the frquenciy of vibration (pitch) increases. One has only to detect this position relative to center. In theory, if there was a ring piezo (not I mean the piezo itself was a ring and a typical center piezo, very fast scanning along with some good math, one could calculate the relative position of the hit from the timing of the sensing of the hit at the two piezos. However, this whole process would be very easy to implement with a FSR potetiometer. In essense, a two layer FSR where one layer has constant resistivity across its surface and is used to determine the hit velocity and the second layer varies in resistance from center to edge thereby determining position (pitch)
3) The midi spec already has what is need for implementation. The is a midi message called poly pressure. It is designed for the aftertouch pressure on each key ( not to be confused with channel pressure that applies to all keys). Given that from item 2) one can detect both the hit ( note on and note velocity) and the pressure (poly pressure) sending these messages immediately one after another would achieve the natural variability of the sound most are after.

However, this is where item 1) comes into play. If you read through other related forums, you will find that most software, even that based on applications like BFD running on high powered workstations, just barely keep up with an active drummer. Part of this is still computer horsepower handling samples but the other part is the throttle of midi 32K. When the connection approaches speed saturation, the message transfers simply slow down.

One can imagine that if 3) were implemented, the volume of midi messages will double and the midi port will look like a garden hose that has a firehose plugged into it.

On the other side, the drum playing hardware/apps must also handle these messages and their volume. Even in the keyboard world there is only a handful of hardware synths and a very few VSTs that do anything with the poly pressure message. This side would also have to upgrade to handle the message and to do so will also require a significant increase the processor capacity.

I would welcome any comments

Re: Real Drum like response from e-pads

PostPosted: Tue Feb 15, 2011 2:19 pm
by daniele99
By now, buy a Mandala Pad (300 Euro, i think it uses your FSR config 128 rings). For Midi, i think we must wait for MIDI II.
;)
By the way, your post is very interesting. If the majors will decide to invest in FSR.