Author |
Message |
Fozzie
Joined: Jun 04, 2004 Posts: 875 Location: Near Wageningen, the Netherlands
Audio files: 8
G2 patch files: 49
|
Posted: Wed Aug 11, 2004 12:57 pm Post subject:
"keyclick" problems with G2 engine |
|
|
Maybe this has been discussed already on the email list, but I couldn't find it. When I play certain patches like Epiano's or others with fairly quick attack settings (see also http://electro-music.com/forum/post-17539.html#17539), I hear short clicks or crackles in the attack phase of the sound that are not consistent each time I hit the same key. Is this normal (and depending on the patch), or is it my engine?
I have ruled out other sources of this problem, as it can be heard also when headphones are plugged in directly in my engine. Hope someone can help, I'm a bit frustrated with it. |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24081 Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
|
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 11, 2004 1:36 pm Post subject:
|
|
|
thanks, Jan. It will probably be the osc phase, indeed, as there does not seem to be any overloads anywhere. Thanks for pointing me to the right post, I could faintly remember something about it, but couldn't find it. |
|
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 11, 2004 1:44 pm Post subject:
|
|
|
Yes indeed, this works (sorry, still quite a synth-newbie). New problem, though: if you play more notes than polyphony allows, these clicks are heard when notes are 'stolen'. Is there any solution to these? It is not the release setting on the ADSR's, that I checked already. Anyone? |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24081 Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
Posted: Wed Aug 11, 2004 2:08 pm Post subject:
|
|
|
Far less easy ...
something can be done but will cost polyphony and it will make you have new problems and complaints as you'll still run out of voice resources and this will always have an audible effect.
You can use envolope modules that have a yellow output that goes active when the envelope is finished. The gate signal for the envelope should be passed through some logic that ensures gate pulses can only go through when the envelope is finished.
Someone made such a patch, but I can't remember who, a quick look through the EM archive didn't get me there, maybe it was on the mailing list but I don't remeber when or who so that's hard to find ...
Voice stealing will still continue to do it's work though and it willl kick in earlier because of the higher patch load.
Better to buy an expansion board, or learn to count your fingers :-)
Jan. |
|
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 11, 2004 3:10 pm Post subject:
|
|
|
I don't care so much that note stealing occurs, but I mind the audible noises along with it. Is this clavia territory that I'm discussing now, or is this still patching stuff?
Trying to keep off the pedal sure works, but hey, I don't want to learn to play too properly (I would have inserted some emoticon here, normally, but I won't, Jan). |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24081 Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
Posted: Wed Aug 11, 2004 4:59 pm Post subject:
|
|
|
Sorry, I didn't read very well, it's the clicks on take over ...
Ok some thoughts about that.
When the previous note is still sounding and at a random moment somewhere in it's cycle it gets reset to zero a click will result indeed with an undetermined amplitude.
There are two possble reasons for the signal to go to zero suddenly.
Either because the envelope gets reset and jumps to zero (but that's a selectable option on some or all of the envelope modules)
or because the oscillator gets reset and will start at zero again.
For the second reason syncing the envelopes to the oscillator instead of the other way around might work. Such syncing could be made by using a sample and hold in the envelope gate signal. The S&H should be triggered from an oscillator output.
For one oscillator ... in this particular patch there are more, some synching between the osc's would also be needed maybe, which then might kill some expression in turn.
I'd try to experiment along such lines.
There is of course a far much simpler way to get rid of the clicks, simply drawn them in reverb ...
Jan. |
|
Back to top
|
|
|
|