Which hardware setup do I need?

[ldb]

Hello guys,

I stumbled upon DSP Trigger by searching for "positional sensing rim trigger" and found the youtube video - I must say I am amazed. Based on the video it seems to be an excellent piece of software, so I want to give it a try.

Now as I understand, I plug the piezo into a line in jack of say my sound card, and it will detect the hits. But of course a normal sound card will only handle one or two pads, which is not enough for a drum set. I also know, that there are various USB audio interface avaialble, but I wonder which one would be optimal?

Do you have any suggestions what to look for? Are there any say 10 channel audio interfaces at reasonable prices out there? If yes, will my laptop (quadcore) be able to handle all the signal processing plus the sequencer with an acceptable latency?
What about hihat controllers? Would it be possible to have them polled somehow? (With some adapter converting resistance to voltage.)

Or did I get it wrong, and DSP Trigger is only meant to process one pad, or maybe two - and the other pads (the less important ones, like toms and crash) are connected with a (cheap) drum module?

LDB

PS: yeahtuna, I also have some questions to you as the developer of the software, which are not really of public interest, so I do not want to post it here. Could you please get in touch with me per e-mail? (The e-mail address is in my profile.)

[ldb]

And one more:
Would it be a theoretical option to do a "multiplexing" of the audio signal?

Imagine a special connector, which will plug into a line in jack and has 5 inputs. It will multiplex in the following way:

T is one period

1/18 T: 100% signal
1/18 T: 0% signal
1/18 T: 100% signal (this pattern is used to give a synchronisation possibilty to the software)
1/6 T: input 1
1/6 T: input 2
1/6 T: input 3
1/6 T: input 4
1/6 T: input 5

Or maybe some frequency modulation?

Cheers,
LDB

[Rob]

For best results from DSP Trigger you'll want to use it in combination with a cheap drum module simply because of the difficulties in regards to hihat controllers as you pointed out.

I've played around with the idea of plugging in a hihat controller into the audio input, but was not able to consistent and reliable results. There's probably a way to do it, but electrical engineering just isn't my strong suit.

If the hihat pedal was patched through a MIDI Keyboard of some other MIDI device that accepts an expression pedal, you could get around needing a module. I would need to implement some code to detect to pedal and splash articulations, but that certainly would be doable without much fuss.

In regards to multiplexing, while I'm sure it's technically possible, it's not something I'd be willing to look into for DSP Trigger.

Rob.

[ldb]

I have checked some audio interface prices for interfaces with more than 5 inputs, and having all drum pads connected directly is definitely an option, unless you're willing to spend a lot.

I think I will try it with a single pad first, then using two inputs, and if that works out well with my pad, I will investigate whether it is somehow possible to sample multiple signals over one interface channel with frequency modulation, separate them with some additional software on the computer, and feed the single channels to DSP Trigger.

LDB

[Rob]

I read up a bit more on multiplexing and it seems like a very interesting idea. You have hardware that will do the multiplexing on the way in?

[ldb]

Not at the moment, and my electronics knowledge is not enough for designing such hardware - those signal processing classes at school are quite a while ago now.., But I will talk to some of my friends, who are more into that topic, what they think about it.

I also believe that probably as an alternative to TDM (time division multiplexing) described above maybe it would be easier to go with frequency domain multiplexing. Though I can not really judge which one is easier to implement on hardware/software side.

In latter case one could amplitude-modulate the trigger signal (http://en.wikipedia.org/wiki/Amplitude_modulation) and put more triggers in separate frequency bands on one audio input. In DSP Trigger you would convert the whole signal to frequency domain with an FFT, cut out one of the subbands, transform back to time domain with iFFT, and do an envelope detection (http://en.wikipedia.org/wiki/Envelope_detector) to get the original signal.
(At least that's how I would imagine it with my minimal SP knowledge

I think the main factor for the question whether it will work or not is the bandwidth needed for correct processing of one trigger signal, and the bandwidth provided by the audio interface.

I believe the carrier frequency has to be a lot higher than the typical frequencies of the trigger singnal, and is probably limited to say 20.000Hz bz the audio interface.

On the other hand trigger outputs do not seem to produce quite a high bandwidth, especially if simplified by a low pass filter by cutting of the parts not needed for interpretation by DSP Trigger? (http://www.edrum.info/theory.html) Can we say that any component above 1000 Hz is irrelevant?

I will assume - probably the assumption is wrong - that a carrier Frequency of about 10.000Hz is still enough to modulate a signal with 1000Hz max.

If so, in the simplest form of AM one trigger would use about a 2000Mhz wide range of the output freqency range. Let's leave a gap between the ranges for each trigger, that would make it possible to modulate up to 9 triggers on one audio input.

Again I must say I am not an export of this matter - but I hope that I will be able to find someone to discuss this with.