False Trigger Problem / AltFalseTrSupp to Yes new or old?

[talrusher]

hi,

im very confused bout the AltFalseTrSupp option. in the documentation is written, that if i set it to yes, the old algorythm will do the job. in the megadrum manager and opn the forum ive read something different. what is now true? with enabled or set to yes, is the new or old falsetrigger algorythm working?

thx

[dmitri]

From the changes log in Latest MegaDrum firmware version:

Changes in version 20110703:

1. A new main false triggering suppression algorithm. The main goal was to improve false triggering suppression without killing fast/press rolls. With the new algorithm the false triggering suppression works much better than the original and is easier to configure especially for mesh type pads. The old false triggering suppression is now AltFalseTrSupp. See http://www.megadrum.info/forums/viewtop ... f=3&t=1630 for details but bare in mind that in this topic the new algorithm was still activated when AltFalseTrSupp was set to Yes. I will updated the MegaDrum configuration shortly to reflect the changes.

Which means the new algorithm is used by default (AltFalseTrSupp is set to No).

[talrusher]

thank you... im trying to setup my meshpads to trigger fast, but without any false notes... now for more than one month, without any good solution. even with my alesis trigger io (that module sounds even without false notes less good than md) dont had significant falsenotes... the whole feelling was till today really betterr, because of having full controle over the toms, without any noticable retrigger or blocking.... that way it is till today not possible for me to get my pads trigger with less than 3 false notes every hit... for that it is on my 12" snare and 10" toms... hihat etc... are working really good...

...tried the old false triggering algorithm, and it really fells better with that. even if i get more than 5 false notes on a hit. it sounds better and fells better... and for now, i can even set retrigger to more than 1 ! with the new algo, had the poroblem with missing or inconsistent hits. dyntime setting more than 12 is really noticable. feels and sounds not good i think... retrigger more than 1 feels strange, because of missing hits. dynlvl not really helps at all... think it has no funktion or not a big one. raising minscan even to a higher lvl than 80 makes the hits only sounding a litte better. best results around 50. meassured my pads halfewave, it is positive and takes only 3ms, but still missing hits. with old algo have many false notes but witout missing hits and it feels and sounds relatively ok... for an unknown reaseon. pads are not to cold or to hot. every pad is around 850 highlvl

on the alesis trigger io, every pad is working like a charm, without having any option than retrigger... but it isn nessessery to set it higher than 1.

wtf? im really confused bout that... really love my md, dont want to go back to trigger io

plzzzzz help me out of this ...

many people have fun with their mds, but me not....

whats the problem? how to solve it?

[dmitri]

1. What pad make/brand/model are you having the problem with?
2. Exactly what settings are you using? (Note, there is no such settings as "I tried everything).
3. Make a video showing the problem.
4. How did you get HighLevel 850?

[talrusher]

to test this issue, ive pluged in my trigger io, and with it, its much better. absolutely no falsetriggering, not a single note. every hit is noticed without missing something or inconsistancy....

have diy mesh toms...
my settings:
latency 40
treshhold 40
linear
minscan 50
dyntime 12
dynlvl 10 (think does nothing)
retrigger 1 (if i set it to something else, i sometimes miss hits)
alttrigger no

for now,i have every hit one falsenote and sometimes 4-8 ... tried rewiring, rebuilding pad etc...
i noticed that if i set alttrigger to yes i can set the retrigger option to more than 1. if alttrigger is set to no, i miss often hits if retrigger is more than 1.
sometimes the falsenotes disapear, for an unknown reason and then they come back... dont know why.

have a positive halfwave. at the moment the hits are very consistent, sometimes it feels like a compressor is on...

hilevel is around 850 with two resistors. one 90k and second 10k. gain is 8.

[dmitri]

to test this issue, ive pluged in my trigger io, and with it, its much better. absolutely no falsetriggering, not a single note. every hit is noticed without missing something or inconsistancy....

And it still registers even very light hits?

treshhold 40
...
for now,i have every hit one falsenote and sometimes 4-8 ...

What is the velocity of the main note and the velocities of the false notes?

have a positive halfwave. at the moment the hits are very consistent, sometimes it feels like a compressor is on...

I.e. velocities are inconsistent? Seems like the pad is still hot. What HiLevel do you get if you set HiLvlAuto to Yes and then make a couple of dozens strongest hits?
Or the piezo signal frequency is very low (half wave is longer than 5ms) in which case you have to increase MinScan.

hilevel is around 850 with two resistors. one 90k and second 10k. gain is 8.

Is the sensitivity with this threshold, gain and hilevel the same as with IO?

[talrusher]

And it still registers even very light hits?

yes, can make a role just stressless with the lightest hits i can make and the io registers (midi velocity 6 - 10)
ok... ive tested my snare pad with the io intensly. noticed that the io displays 3 false notes on every hit. the falsenotes are usualy not noticable by hearing. the triggering is really fast, but noticed that they are no mid range notes, even if i set curve to loga3 or max 4. that is why the io sounds so shi...y.
with my md32 newest firmware:
gain 8
curve loga1
treshold 30
hilvlauto no (850) <- this is the end. the pad is not to hot!
retrigger 8
dynlvl 15
dyntime 16 (with 24 or even more have only 1-2 falsenotes, not noticable by ear)
minscan 40 <- 50 without changing something (the halfwave is positive and only 3,8ms long. meassured with cubase)

...the light hits are triggered really better, but even now not as god as the io and im getting many many falsenotes 4-8 on every hit.

treshhold 40
...
for now,i have every hit one falsenote and sometimes 4-8 ...

What is the velocity of the main note and the velocities of the false notes?

thats alternating... sometimes hitting 127 and getting back 120, 98, 3, 3, 5, 1 ..... and so on
sometimes hitting 90 and getting only 20 back as falsnote. noticed that the weak hits dont coming back so often as the strong hits...

have a positive halfwave. at the moment the hits are very consistent, sometimes it feels like a compressor is on...

I.e. velocities are inconsistent? Seems like the pad is still hot. What HiLevel do you get if you set HiLvlAuto to Yes and then make a couple of dozens strongest hits?
Or the piezo signal frequency is very low (half wave is longer than 5ms) in which case you have to increase MinScan.

i mean that the pad is really damn rocksolid consistent. i hit 100x with the same intensity and get 100x the same midi note. it is a bit strange to me, because dont knew such precision on the tirgger io.
the halfwave produced by this pad is only 3,8ms long so theres no reason to put minscan higher than 40, think 35 would be enough for absolute consistence, but i set it to 50 anyway because sure is sure...

the trigger io has no options for falsenote suppression, the only is retrigger. on setting this from 0-10 theres not really a change on outputting falsenotes. maybe the io was supressing 1 or 1,563333^ falsenotes, but they were not noticable to me. theres only gain as an opion to change the velocity, so think the hilvl is fixed. noticed this by trieing set the velocity right. it is nearly impossible:

tried to adjust the two modulues to trigger as equal as it was possible for me:
i could not understand this. could not believe that the io has better falsetriggersupression and tested things anyway... the whole night (damn, its really not the first)

for testing i raised the minscan interval to 80 -> 1 -3 falsenotes less outputting... ok, tried to raise latency to 100 -> 1-3 falsenotes less and together this make 6 - ... falsenotes less...
so on, i feed the md32 with this trigger settings:
treshold 30
gain 8
hilvl 850
retrigger 6 or even less
dynlvl 15 (lowering this gives me no noticable suppression in midi log so let stay. think can hear a difference but im not sure)
dyntime 8 or even 4 so the light hits are really nice and smoothy triggered. soft rolls, the softest hits i could make, are not longer a problem and the rest sounds normaly to me, a bit better and smoothier.
minscan 60 -> latency 80

think this does the trick. the md respondes in suppressing falsenotes as good as the trigger io in this case, without any cost in my opinionl. it sounds smoothier and triggers even softest rols like hell...

[dmitri]

I understand you got a good result after all.
Anyway, it looks like the signal produced by your pad on a hit subsides very slowly. In this case I would suggest first to tighten the head mesh.

After that I would suggest the following for a mesh pad which produces long half waves and slowly subsiding signal.

1. AltFalseTrSupp must be set to No (default) and HighLevel must be set correctly for the pad and the pad must not be too "hot' or it must be "cooled".
2. Set MinScan to ~40
3. Set Curve to Linear.
4. Set Threshold to 80 (yes 80, it will be lowered later).
5. Set Retrigger to 1.
6. Set DynLevel to 15 and DynTime to 60

Once it is set like this you should not get any "after hit" false notes. Obviously fast rolls will be missing hits with these settings.
After that, making a single strong hit at a time, start to reduce DynLevel and monitor MIDI Log.
Reduce DynLevel until false notes start to appear.
At this step raise DynLevel by 1 or 2.
Now start reducing DynTime until false notes start to appear.
At this step raise DynTime by 1 or 2 steps.
Now start reducing Threshold until false notes start to appear. Every time false notes start to appear increase DynTime. Don't reduce Threshold too much otherwise you won't be able to get rid of low velocity false notes.

At the end you should get the best balance between false notes suppression and good fast rolls. Of course you will want to change the Curve to your liking after that.

I wrote all of the above from memory of what I did with one of the mesh pad with a slowly subsiding signal.

[talrusher]

ive recorded the raw input sound from my pad to reproduce exact input for both drumbrains. ive feed both with this file and recorded the final triggered sound by superiordrummer2.

i found it very interesting to compare both drumbrains and maybe you are even interessted in seeing or listening what the exact differences between both are and which algorithm sounds better, trigger io or megadrum32

md32 retrigger6 ....wma is with discribed settings: latency 80 minscan 50 or 60 ... think sounds really good.

SOUNDFILES:
https://skydrive.live.com/#cid=EF315770276A6635&id=EF315770276A6635%215297

[talrusher]

I understand you got a good result after all.

...

At the end you should get the best balance between false notes suppression and good fast rolls. Of course you will want to change the Curve to your liking after that.

I wrote all of the above from memory of what I did with one of the mesh pad with a slowly subsiding signal.

i will give it a try. posted my pad raw outgoing sound for testing purpose... maybe you check the signal waves for making a diagnostic or listen to the triggered soundfiles with different settings and the exact same sequenz with TRIGGER IO to compare both drumbrains and getting suggestions or someting... if you want