Zen Engine Sound Design Quirks
29 Nov 2020, 11:29Some Notes of Sound Design quirks with the Zen Engine:
1. When dealing with Structs that share a single filter and amp:
One partial (the 2nd one) loses it's ability for envelopes and LFOs to modulate the filter and amp (rather than those things resulting in summed results)
*would be nice if these modulation sources weren't just lost in a future update*
2. When dealing with modulating effects, different partials set to do such (in the mod matrix) end up in conflict rather than summed results.
What I mean by this, is say one partial has an envelope modulating some effect parameter while another partial has System Control 1 set to modulate the same parameter, rather than the results of these modulations being summed, one partial will over take the other [near] completely resulting in adjustments of the overtaken partial only having very brief momentary recognition of it's value before things snap back to the value set of the other partial. I'm unsure of how the priority is established, but I "think" it's a matter of which ever partial does a modulation first gets the priority (but haven't bothered testing to verify).
3. Delayed partials DO NOT recognize (mono) legato play and will always retrigger.
(would be nice if Roland added an option to have such action on legato play enabled OR disabled)
4. The second wave slot is ALWAYS used (regardless of if explicitly set to be or not). As such, stereo detuning and stereo width difference parameters can ALWAYS be implemented and used on sample based Osc (even when only one waveslot is set to be used). This DOES NOT apply to PCM-Sync, Noise, Supersaw, or VA.
*it's been mentioned some, but will note again here that the second wave slot does count as a voice, and subsequently every active partial takes up at least 2 voices (this continues to stand even if a partial has been set to a blank sample memory area); there is no such thing as a single voice partial. Of related end, to avoid unnecessary voice eat up when sequencing external gear, all partial switches on clips intended for external gear should be set to OFF.*
*5. (this one is documented, though arguably not well explained) When partials are set to cycle/random mode, a partial can be used consistently if it has a full velocity range and like wise can have partially over lapping parts with other cycles to the extent that it's velocity range shares the velocity range of other cycles.
guessing there may be more little (undocumented) quirks and very interested in what else others may have stumbled across.
1. When dealing with Structs that share a single filter and amp:
One partial (the 2nd one) loses it's ability for envelopes and LFOs to modulate the filter and amp (rather than those things resulting in summed results)
*would be nice if these modulation sources weren't just lost in a future update*
2. When dealing with modulating effects, different partials set to do such (in the mod matrix) end up in conflict rather than summed results.
What I mean by this, is say one partial has an envelope modulating some effect parameter while another partial has System Control 1 set to modulate the same parameter, rather than the results of these modulations being summed, one partial will over take the other [near] completely resulting in adjustments of the overtaken partial only having very brief momentary recognition of it's value before things snap back to the value set of the other partial. I'm unsure of how the priority is established, but I "think" it's a matter of which ever partial does a modulation first gets the priority (but haven't bothered testing to verify).
3. Delayed partials DO NOT recognize (mono) legato play and will always retrigger.
(would be nice if Roland added an option to have such action on legato play enabled OR disabled)
4. The second wave slot is ALWAYS used (regardless of if explicitly set to be or not). As such, stereo detuning and stereo width difference parameters can ALWAYS be implemented and used on sample based Osc (even when only one waveslot is set to be used). This DOES NOT apply to PCM-Sync, Noise, Supersaw, or VA.
*it's been mentioned some, but will note again here that the second wave slot does count as a voice, and subsequently every active partial takes up at least 2 voices (this continues to stand even if a partial has been set to a blank sample memory area); there is no such thing as a single voice partial. Of related end, to avoid unnecessary voice eat up when sequencing external gear, all partial switches on clips intended for external gear should be set to OFF.*
*5. (this one is documented, though arguably not well explained) When partials are set to cycle/random mode, a partial can be used consistently if it has a full velocity range and like wise can have partially over lapping parts with other cycles to the extent that it's velocity range shares the velocity range of other cycles.
guessing there may be more little (undocumented) quirks and very interested in what else others may have stumbled across.