Author |
Message |
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Wed Aug 24, 2005 1:28 pm Post subject:
fast retriggering problem |
|
|
I was using the G2 engine today with my band members, among other things as an e-drum module triggered by an electronic drumkit (yamaha dtxplorer). The problem is that the module of the drumkit uses a standard notelength for the midi-out signal that the pads generate (each pad on a fixed note). When the same pad is hit fast, the new note-on signal is sent before the old note ended. Problem with the G2 is that it now skips some of the fast trigger signals, because the gate from the midi note-receive module stays high over the overlapping notes (hope I am making any sense here).
I can't find a way to let the note-receive module (or other note/trigger/gate receiving module) to act only on the 'positive edge' of the note-on signals with just a very short pulse (shorter than the midi note duration). Am I missing something? Is it possible on the current G2? |
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004 Posts: 7678 Location: The Hague, NL
G2 patch files: 3
|
Posted: Wed Aug 24, 2005 1:45 pm Post subject:
|
|
|
That sounds like a implementation problem of the drumkit.
Some stuff does do this on purpose though, Live has a "hidden" mode that sends midi like that; note on when a loop gets started, a note on with velocity1 for each repeat, note off at stop. It's meant for syncing up VJ stuff.
What you could do is detect velocity changes and work from there.
take your velocity output, send it to a one sample delay that you clock with a inverter fed back on itself.
Make a "equals" out of two compare modules (x=y if (x is greater then or equal to y) and (y is greater then or equal to x) and use it to compare the velocity to the delayed velocity.
If those two aren't the same there must have been a new note.
You might also be able to fix this by using more voices; I think a "note on" should spawn a new voice even without a "note off". That may lead to side-effects so be carefull. _________________ Kassen |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Wed Aug 24, 2005 1:55 pm Post subject:
|
|
|
Thanks Kassen, good tip concerning the velocities. I am indeed quite annoyed by the way the drum module works; I can do some adjustment on the note-on length, but it can't be set as short as I like. edit: By the way, the drum module works OK with softsampler plugins, there the triggering can be very fast (so the implementation is not that incompatible with other tone generators).
The polyphony setting: I tried that, but unexpectedly I couldn't detect an improvement. Maybe because it's the same note-number.
I think I will try your tip in combination with the ordinary triggering combined in an OR combination gate. That way there's a better chance that two hits that have the same velocity (accidentally)are detected both. |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Wed Aug 24, 2005 2:59 pm Post subject:
|
|
|
This is the patch as I see it. It uses quite a few logic modules for 1 trigger note. Can I be more efficient? The green modules are the actual velocity-change detectors. I struggled a bit with detecting both positive and negative changes in the velocity value, I tried with a levelconverter module but didn't succeed, hence the double detection of positive and negative changes.
Description: |
screendump of relevant part of the patch |
|
Filesize: |
63.46 KB |
Viewed: |
20152 Time(s) |
|
Description: |
Detect fast triggering of one note by changes in velocity; demo patch |
|
Download |
Filename: |
Fasttrigger testpatch.pch2 |
Filesize: |
1.53 KB |
Downloaded: |
1356 Time(s) |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24075 Location: The Netherlands, Enschede
Audio files: 277
G2 patch files: 320
|
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004 Posts: 7678 Location: The Hague, NL
G2 patch files: 3
|
Posted: Wed Aug 24, 2005 3:58 pm Post subject:
|
|
|
Fozzie wrote: | I struggled a bit with detecting both positive and negative changes in the velocity value, I tried with a levelconverter module but didn't succeed, hence the double detection of positive and negative changes. |
That could be a bit tighter. This pc is Linux and so won't run the editor so bear with me.
We have X and Y. X is the velocity output, Y is the delayed velocity, ok?
You take the compare A-B module (in the level or the switch section). It outputs "1" if one signal is higer to or equal to the other signal.
You take two of those. On the top one you connect X to the left input and Y to the right and on the bottom one you do it the other way around. Then you put a AND over those two modules.
What's going on there? Well, the only case in which both are true is if X is equal to Y. You "logic invert" that to get a pulse at every change.
This is a bit tighter. It just detects changes, not what the change is, but for that it's realy usefull and tight. If you want changes measured then you substract X from Y (use a mixer and invert the Y input) and you get the change on the output, often quite usefull too.
If all else fails you could always close-mike the drumpads and take a RMS (root mean square, a envelope folower) with a very short time folowed by a compare to a static value to that signal. That might catch hits too but it's a little expensive.
Trick matter. Good luck! _________________ Kassen |
|
Back to top
|
|
|
deknow
Joined: Sep 15, 2004 Posts: 1307 Location: Leominster, MA (USA)
G2 patch files: 15
|
Posted: Wed Aug 24, 2005 7:50 pm Post subject:
|
|
|
fozzie, i don't have my g2 hooked up at the moment, but is it possible to connect the gate from the note detector to an envelope generator, and set the eg to "retrigger", and use that as your gate?
deknow |
|
Back to top
|
|
|
Kassen
Janitor
Joined: Jul 06, 2004 Posts: 7678 Location: The Hague, NL
G2 patch files: 3
|
Posted: Thu Aug 25, 2005 3:02 am Post subject:
|
|
|
Blue Hell wrote: | This is what I use for change detection. |
Nice. That even saves a inverter compared to the classical method! _________________ Kassen |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Thu Aug 25, 2005 4:21 am Post subject:
|
|
|
Blue Hell wrote: | This is what I use for change detection. |
The rectifier , I clearly overlooked that cute module. Thanks @ Kassen as well for your solution. |
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Thu Aug 25, 2005 9:28 am Post subject:
|
|
|
i also see an implementation prolem for clavia there.. a modular synth actually should be able to process 2 overlapping notes on the same pitch..It should retrigger on the next note on command...we should report that to clavia...they have done drumtriggers for years... i dont think the dd drum stuff works much different...
Have you tryed the different poly and mono modes of the G2? |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Thu Aug 25, 2005 10:03 am Post subject:
|
|
|
3phase wrote: |
Have you tryed the different poly and mono modes of the G2? |
Tried mono, legato and polyphony up to 4, all gave the same results (subjectively tested though) |
|
Back to top
|
|
|
deknow
Joined: Sep 15, 2004 Posts: 1307 Location: Leominster, MA (USA)
G2 patch files: 15
|
Posted: Thu Aug 25, 2005 10:16 am Post subject:
|
|
|
...i might be misreading, but wouldn't the "monokey" module (in the in/out group) send a gate everytime a key is pressed (if it is "last" mode)?
deknow |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Thu Aug 25, 2005 12:32 pm Post subject:
|
|
|
deknow wrote: | ...i might be misreading, but wouldn't the "monokey" module (in the in/out group) send a gate everytime a key is pressed (if it is "last" mode)?
deknow |
Haven't tested, but I don't think it works. The 'last' option makes sure the last notevalue is output at the pitch out, but in quick fills this value doesn't change. And the helpfile describes the gate out just as the gate outputs on other modules (the gate goes high upon note-on and stays high until the (last) note is released). Thanks though. |
|
Back to top
|
|
|
deknow
Joined: Sep 15, 2004 Posts: 1307 Location: Leominster, MA (USA)
G2 patch files: 15
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Thu Aug 25, 2005 12:56 pm Post subject:
|
|
|
deknow wrote: | yes, you are right, the gate will stay high. another approach might be (i don't use the nord with a kbd or midi input, so i don't understand this aspect so well) using the status module in poly mode, and using the voice number output to trigger a gate.
this might help....it's a method i use to prevent the pitch tracker from jumping octaves (not a direct solution to your problem, but it might lead you to something).
deknow |
i am not sure if another voice is selected when no trigger occurs..but its worth investiating.. but again..
this sounds like an issue clavia should take care of.. even when i havent run in the phenomen yet it can happen any time..lots of devices work the way as the described drumtrigger..can happen even with interslot midi ..but maybe its sorted for that...??
We have still a 1.x os..
the instrument looks pretty finished but in the details its actually still in development..who informs clavia about the issue? |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Thu Aug 25, 2005 2:22 pm Post subject:
|
|
|
3phase wrote: | .who informs clavia about the issue? |
I'll drop them a line. By the way, the velocity trick does its job very well, triggering has improved so my problem has become of much less importance. |
|
Back to top
|
|
|
ian-s
Joined: Apr 01, 2004 Posts: 2669 Location: Auckland, New Zealand
Audio files: 42
G2 patch files: 626
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Thu Aug 25, 2005 3:34 pm Post subject:
|
|
|
g2ian wrote: | This is a little curious. I was looking for a solution to this problem when I discovered that, at least with the internal routed MIDI, it works anyway.
This test patch uses pulse modules to create an overlapping stream of MIDI note on/off for the same channel/note number. You can see from the LED’s that the next note starts before the previous completes. Yet the note detect responds to each note on.
Maybe there is another problem, when doing MIDI controlled lighting in the early 80’s, I had a problem with lost data due to very long MIDI cable runs. Every now and then, a spot light would ignore a note off and stay on between songs (annoying).
I suspect that the minimum note length for the drum module would be only a few MIDI clocks, very difficult to play fast enough to overlap. |
lots of early soundexpanders needed a certain gatetime to sound at all...
but..usually this time is adjustable in the trigger converter...
iv no midi setup wright now but i think its worth to check it...
maybe its a bug...maybe it was the trigger situation... |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Fri Aug 26, 2005 1:29 am Post subject:
|
|
|
g2ian wrote: | ...
Maybe there is another problem, when doing MIDI controlled lighting in the early 80’s, I had a problem with lost data due to very long MIDI cable runs. Every now and then, a spot light would ignore a note off and stay on between songs (annoying).
I suspect that the minimum note length for the drum module would be only a few MIDI clocks, very difficult to play fast enough to overlap. |
Well, when I control my softsampler plugin (DR008) on my PC I can trigger samples as fast as the internal sounds of the module, so cables are not the culprits.
The drum module is very basic (as is the manual ), and it is possible to adjust triggerlength but only between 0.1 and 9.9 seconds (don't me ask why so long). The standard setting appears to be ~0.5 seconds, and that was plain horrible with the G2, but no problems with the softsampler plugin. Therefore, it does seem to be a problem with external note-on detection (although with the 0.1 setting on the module and the solutions in this thread the problem is 98% solved, so to me it's more a matter of principle now).
I'll see if I have time tonight to do some less subjective tests by recording notes in my sequencer and sending these to the G2. Maybe that way I can be a bit more specific and scientific. |
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Fri Aug 26, 2005 6:35 am Post subject:
|
|
|
Fozzie wrote: | only between 0.1 and 9.9 seconds (don't me ask why so long). |
Its for cymbal sounds... just in case you like to know.
I do a test aswell when my midi is hooked up again.
A test with the most current OS is more usefull when searching for bugs. |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Fri Aug 26, 2005 6:54 am Post subject:
|
|
|
3phase wrote: | Fozzie wrote: | only between 0.1 and 9.9 seconds (don't me ask why so long). |
Its for cymbal sounds... just in case you like to know.
I do a test aswell when my midi is hooked up again.
A test with the most current OS is more usefull when searching for bugs. |
Forgot to mention: I updated to make sure this was no old version bug. I'm on 1.32 right now. |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Fri Aug 26, 2005 1:11 pm Post subject:
|
|
|
I might be on to something. If I sent whole notes from my sequencer every quarter note (4 overlaps all the time; doesn't seem to matter which note or which midi channel unsurprisingly) to the g2 everything is fine....except for the first 4 quarters. It starts with a beat on quarter 1, than three quarters of silence, and from then onwards every quarter.
I'll run some more tests to see what I can come up with. I forgot to mention, there was also something funny when sequences were triggered by the pads. There was a slight pause between hitting the pad and the start of the sequence. Now that I think about it, it might have started somehow on the end of note . Before jumping to conclusions, this could of course also be my shabby logic programming. I'll dig into it. |
|
Back to top
|
|
|
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Fri Aug 26, 2005 1:28 pm Post subject:
|
|
|
Weird?? I used a clock divider after the gate output from the note-receive module. When I use the output from the clock divider to trigger anything (even if divide=1), it triggers on the negative edge of the note-on (on the note-off if you will). Directly using the gate from the note-receive is OK.
I use the clock divider in its default dropdown setting. It should not reverse positive and negative edges, should it?
edit: in more tests (driving it with momentary switches) the clock divider always responds to negative edges, not on positive edges. BUG! |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24075 Location: The Netherlands, Enschede
Audio files: 277
G2 patch files: 320
|
Posted: Fri Aug 26, 2005 4:59 pm Post subject:
|
|
|
This is a bit strange indeed, when the clock divider would react on negative edges only it could not possibly divide by 1 I'd say ... do you do something with the reset input maybe ? _________________ Jan
also .. could someone please turn down the thermostat a bit.
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Fri Aug 26, 2005 7:10 pm Post subject:
|
|
|
Blue Hell wrote: | This is a bit strange indeed, when the clock divider would react on negative edges only it could not possibly divide by 1 I'd say ... do you do something with the reset input maybe ? |
I came along some problems with triggering on negativ edges aswell..
i reported something what looked as a bug to me to clavia but i wasnt able to really put the finger on the point..it was rather that i realized the sequencers are advanced also on release of a key or so..cant remember...
I wasnt sure if this was a bug or a feature in the situation and so i cant be sure if my formulations made clavia finding a problem or just seeing it as a user error..What is probably the majority of reports they get sended in...
It cant be a mistake to look at this item closer... Every hidden Bug we find makes our instrument better in the end...
Lazy me have to install the bloody midicables..I hate it...
USB midi would have been nice |
|
Back to top
|
|
|
|