Double inputs not registering

[Mistilru]

only one don or kat can be registered at a time. cannot do double hits as only one side will register

[LuiCat]

This is the trade off for using a shared envelope threshold under current version. The first few versions of this may have implementations with independent threshold, but tested buggy for me.

Currently I don't have a testing environment to fix this. Will keep this issue open until time.

[karabambus]

Is it buggy because Arduino can read analog inputs only one at a time?

If that is the case - is it possible to solve this by outputting double hit (two hits simultaneously) if two analog inputs pass the threshold in short span of time (short enough that it outputs two hits when you want it and not double hit, and vice versa)?

[LuiCat]

Is it buggy because Arduino can read analog inputs only one at a time?

Yes but no.

It is true that Arduino Leonardo/Micro, the atmega32u4 ones, are only capable of reading one analog pin at a time. However, this is not the true cause of the problem.

The design consideration and problem with the threshold is that, not everyone's drum is perfectly made. Most people may not deal with the sound sealing between sensors well. This makes the drum to pickup hits from multiple sensors. As a result of this, we cannot simply pick any input that passes the threshold.

As now, ignoring drum hits that is too close in time is designed by intention, to overcome the issue above.

If you have a professional drum (even better than taiko force v5 ones, since those drums may still having crappy sensors) and would like to try a workaround, you can lower the amount of threshold that each drum hit raises. This is done be modifying float k_threshold = 1.5 to a value lower than 1.0.

[LuiCat]

BTW if anyone is interested, AnalogReadNow.cpp/.h contains an implementation of analog read that switches the read pin early, to avoid idle waiting during the read.

This allow us to read the input pins one at a time at relatively higher frequency without too much error in the read sensor level.

[Kamehamehaas]

Hello,

I am looking into building my own taiko for the switch. I want to make sure what we are saying is the cause of this is the floatk_threshold limit?

Also I took a look at the .ino code, and while I could not follow 95% of it, I think I noticed there are no independent thresholds for each sensor? I assume if we wanted a sensor to be more sensitive we would increase the sen level?

I am by no means an expert and your code seems to be the easiest way to build my taiko for the switch, and I just want to get it to a level where I can play more of the harder levels

[LuiCat]

Hi @Kamehamehaas , you can modify sens in the code to adjust the sensitivity of your sensor. If your sensor is giving signals too strong, try to lower it. The input signal will be scaled by that value.

By the way, the recommended threshold value is just above the noise level that all your sensors produce. If the threshold is low enough but not giving unwanted double inputs, then you are good to go.

[Kamehamehaas]

@LuiCat I tried to edit my comment when I read the read me again but was too late. So thank you!

But has there been any further breakthrough on the double hits? The whole reason I want to build my drum is because the switches motion controls cannot capture fast movements well so this code being able to recognize rapid or double hits is crucial to me lol

[LuiCat]

@Kamehamehaas , I think I can try out an older solution with separate sensor thresholds when I get my taiko drum back.

[Kamehamehaas]

Sounds good @LuiCat , I am honestly going to try an get to a testing stage through serial monitoring myself this weekend. So I can't also be a test dummy if you're interested

[Kamehamehaas]

Also a possible dumb question. If I have a Bluetooth module we can wire to the Arduino, could this be wireless? I don't know how the joy con library works

[LuiCat]

@Kamehamehaas I'm interested in bluetooth solutions as well. I remember there is libraries on GitHub that reverse engineered NS's bluetooth protocol.

In another way, maybe you can get a cheap 3rd party joy on, and wire it up to Arduino. This might work as well, but need some soldering works.

[karabambus]

@Kamehamehaas I think code recognises rapid hits well, and without double hits it is still playable.

Concerning the problem that @LuiCat mentioned with sensors picking up "noise" from other hits - I think my drum has pretty good isolation of individual sensors so I can also test it out. I was really lazy, and everything worked fine so I didn't try to experiment with separate sensor threshold version of code. Maybe today or tomorrow Il' try to post an update how it worked!

[Kamehamehaas]

@LuiCat my original plan was to solder into a controller, but I saw this and I had everything already to deploy so wanted to give it a go.

@karabambus thanks for the input for the rapid hits! That's good to know. If you don't mind me asking, how did you isolate your sensors? Also any testing you do and share would be great to know 😁

[Kamehamehaas]

Any update on isolation or double hits? I've spent the weekend still testing my isolation set up and didn't try the code yet