Working with the Voice Stealer (Polyphony Allocation)
How can I control voice allocation, so I can choose which programs will have voices stolen if I run out of polyphony?
The K2 has its own dynamic voice stealing algorithm (stealer). Although you can't specify which programs have priority over others, you can edit the programs to make them less likely to be stolen by the stealer.
- If you want a sound to be prioritized so it won't be stolen, make sure the key is held down throughout it's duration. If sequencing, make sure the note's duration is as long as the sound needs to be around. Notes held only by the sustain pedal have a somewhat lower priority, and those held with nothing (but with a long release) have an even lower priority.
- A more drastic approach is to use a mono program for drum loops, sound effects etc. since mono programs are given maximum priority, although it limits you to one sound per MIDI channel. Depending on your application you may be able to make (one or more) monophonic programs, combine it with the rest of your sounds using a performance setup. Be aware that when using ignore release with mono programs, make sure "Legato Play" is Off on the COMMON page.
- The stealer gives the low key on a channel extra priority.
- A certain amount of priority comes from the amplitude of the layer. You can usually pad the signal somewhere in the algorithm by 18 dB, then turn it up on the AMP page by 18 dB. This makes it look hotter to the stealer, but there is no gain difference unless you are using a distortion type DSP. You could go on turning up the AMP page, and turn it back down on the output page, but you'll quickly run into clipping.
- A certain amount of priority comes from the AMPENV level. If you have a looped sample and use a decaying amp envelope, the priority decreases as the note decays. So if you are playing an unlooped decaying sample (i.e. a ten second sample of a chime), and you edit the envelope so that it doesn't decay, the priority remains about the same during the actual decay that is within the sample recording.
Beyond that, the priority is controlled by the amp page and by the amp envelope, so you can increase priority (without affecting volume) by turning up the amp page and using a pad somewhere in the algorithm.
- You can also set one or more of the envelope "hold" flags on the LAYER page, ThrAtt and TilDec.
Need more info? Check out the following links: