www.megadrum.info

I'm setting up my DIY electronic kit and after spending a few hours tweaking the setting I have a few questions regarding the MinScan parameter.

1. I have 10" DIY mesh pads with centrally placed piezos and Yamaha PCY135 piezo+switch type cymbals. What would be the 'scientific' way to set the correct MinScan value for these pads? My problem is that when I tweak my settings, I often change MinScan too, which adds an extra parameter to the equation. But I think MinScan should be a 'Set once and then forget' parameter. Is that right?
2. How can I find out or measure the correct value for MinScan? In the online docs for PS it says: "With Positional set to 2, MinScan must be set so that to cover a full half wave which is ~30-40 (3-4ms) for Roland PD-125X." The docs about MinScan recommend a value of ~20 for rubber pads and ~30-50 for mesh pads. I think I would need to set MinScan more accurately than that.
3. So far the most 'scientific' method I tried was to plug my pad's jack plug into my soundcard's mic input and record a few hits with a DAW. (It is a poor man's oscilloscope...) This way I was able to determine the length of the full half-wave in milliseconds (by counting the number of samples in the wave). But what shall I do with this value? Shall it be my maximum for MinScan? Or the minimum? Do I need the length of the first half wave, at all? If not, what shall I measure?
4. There is one sentence in the docs about MinScan which I don't understand: "it will keep sampling it for MinScan period of time before marking the signal as registered and making it ready to be sent over MIDI next MegaDrum scans (scan time is set by 'Latency') all channels for registered signals." First of all, what does "marking the signal as registered means"? Can someone (Dmitri?) explain the whole mechanism a little more? Also, part of the sentence is missing where the red text is.
5. What problems can it cause if I have a completely wrong MinScan setting (because it is too high or too low)?

The reason why I think I need more accurate values than '~30-50' is because in reality MinScan seems to influence a lot of settings. MD seems to be very sensitive for a correct MinScan setting. For example it influences level detection accuracy. It can also make my DynLevel and DynTime settings work (or break them). It is also critical for correct Positional Sensing. It also seemed to solve the hotspot issue of my PCY135 cymbal.
So I would really appreciate the answers to my questions and a general explanation of how MinScan works and how to set it correctly.

halifax30 wrote:I'm setting up my DIY electronic kit and after spending a few hours tweaking the setting I have a few questions regarding the MinScan parameter.

1. I have 10" DIY mesh pads with centrally placed piezos and Yamaha PCY135 piezo+switch type cymbals. What would be the 'scientific' way to set the correct MinScan value for these pads? My problem is that when I tweak my settings, I often change MinScan too, which adds an extra parameter to the equation. But I think MinScan should be a 'Set once and then forget' parameter. Is that right?
   

MinScan is individual for each pad/cymbal. Once you set a correct MinScan for a pad/cymbal you don't need to change it any more.

How can I find out or measure the correct value for MinScan? In the online docs for PS it says: "With Positional set to 2, MinScan must be set so that to cover a full half wave which is ~30-40 (3-4ms) for Roland PD-125X." The docs about MinScan recommend a value of ~20 for rubber pads and ~30-50 for mesh pads. I think I would need to set MinScan more accurately than that.

You don't have to set MinScan very precise. MinScan 20 and 21 will make only 0.1ms difference in latency.

So far the most 'scientific' method I tried was to plug my pad's jack plug into my soundcard's mic input and record a few hits with a DAW. (It is a poor man's oscilloscope...) This way I was able to determine the length of the full half-wave in milliseconds (by counting the number of samples in the wave). But what shall I do with this value? Shall it be my maximum for MinScan? Or the minimum? Do I need the length of the first half wave, at all? If not, what shall I measure?

MinScan should be set so that it covers a peak of the first half wave. If the peak of the first half wave is in 2ms after the start of a signal then you should not set MinScan lower then 20 to guarantee that MegaDrum's sampling for the signal covers the peak,

There is one sentence in the docs about MinScan which I don't understand: "it will keep sampling it for MinScan period of time before marking the signal as registered and making it ready to be sent over MIDI next MegaDrum scans (scan time is set by 'Latency') all channels for registered signals." First of all, what does "marking the signal as registered means"? Can someone (Dmitri?) explain the whole mechanism a little more? Also, part of the sentence is missing where the red text is.

MegaDrum considers a signal ready processing and sending over MIDI only if from the start of a signal (crossed Threshold) it was sampled for at least MinScan/10 milliseconds. Once it was sampled for MinScan/10 milliseconds it is marked as registered and ready for processing and sending over MIDI. Checking if an input has a signal registered is done every Latency/10 milliseconds.

What problems can it cause if I have a completely wrong MinScan setting (because it is too high or too low)?

If MinScan is set to low then you may get incorrect velocity values for the same signal level (hit strength). If you set MinScan too high you increase overall latency from the particular input. Overall latency for an input is between MinScan and MinScan+Latency (in milliseconds: between MinScan/10 and (MinScan+Latency)/10)

Hi Dmitri,
I finally got around to record my pads with my PC to see how their waveform looks. I simply connected my pads to my PC's microphone input one by one and recorded light, middle and hard hits.
I still have to analyse my results, but here is a quick one. The picture shows a hard hit on my kick drum. This is a DIY 13" kick with Pearl mesh head and a 35mm piezo placed at the edge (inside the shell) with a foam cone (actually it is a pyramid with 4 sides and a chopped tip).
My question is, what should be my MinScan for such a waveform? Shall I set it so it covers the highest peak? Does MD take into account the highest peak during the sampling time of MinScan? It is interesting to note that the first highest peak is negative.
My current setting for my kick is MinScan=30, which is about 132 samples at 44.1kHz. This is marked on the picture.

Yes, MinScan must cover the highest peak. If an input doesn't have a precision rectifier then negative signals will be ignored.