MIDI mapping in DS2
-
- KVRer
- Topic Starter
- 13 posts since 20 Feb, 2004
I love the sound of DS but I am very unhappy with the way MIDI mapping is implemented. when I change a parameter with a midi controller the range of the target seems to be split into 128 equally spaced steps independently from the function of the control. this makes it impossible to use midi for certain controls like amp attack time or filter cutoff since the steps are far too big with low values and too small for high values. could this be changed to a non-linear behavior?
-
- KVRist
- 440 posts since 13 Mar, 2009
-
- KVRer
- Topic Starter
- 13 posts since 20 Feb, 2004
I can confirm that it works good with the model parameters. problem seems to be only with the 2nd row of parameters (LFO1..., AMP..., DIST...).
I have so far only used DS in Live. I did some more testing now with Live and Max 6 (using the [vst~] object) with interesting results:
* Live 9.1: if i map the same CC to several parameters only the first one reacts on changes ("first" means the first parameter in the list of parameters which I made visible to MIDI remote control by using the Configure button).
* same parameter reacts differently in different hosts. for example:
AMP Attack:
Live 9.1: CC 1, 2, 3, 4 -> 79, 157, 236, 315 ms. range 0-10000.
Max 6: CC 1, 2, 4, 4 -> 7.87, 16, 24, 31 ms. range 0-1000.
OK, obviously different by a factor of 10 but still both linear. i think some exponential function would be the way to go.
* Max 6: (at least) the following parameters in DS2 are sending value above 1. when changed. these parameters cannot be controlled in Max since [vst~] only accept 0.-1. for parameter values: LFO/FM SOURCE, LFO Speed, LFO waveform, DIST Filtermode, DIST Filter, distortion algorithm, FILT cutoff, filter mode, RESON master, Freq1+2+3, SHIFT coarse freq, coarse freq mod source, fine,
I have so far only used DS in Live. I did some more testing now with Live and Max 6 (using the [vst~] object) with interesting results:
* Live 9.1: if i map the same CC to several parameters only the first one reacts on changes ("first" means the first parameter in the list of parameters which I made visible to MIDI remote control by using the Configure button).
* same parameter reacts differently in different hosts. for example:
AMP Attack:
Live 9.1: CC 1, 2, 3, 4 -> 79, 157, 236, 315 ms. range 0-10000.
Max 6: CC 1, 2, 4, 4 -> 7.87, 16, 24, 31 ms. range 0-1000.
OK, obviously different by a factor of 10 but still both linear. i think some exponential function would be the way to go.
* Max 6: (at least) the following parameters in DS2 are sending value above 1. when changed. these parameters cannot be controlled in Max since [vst~] only accept 0.-1. for parameter values: LFO/FM SOURCE, LFO Speed, LFO waveform, DIST Filtermode, DIST Filter, distortion algorithm, FILT cutoff, filter mode, RESON master, Freq1+2+3, SHIFT coarse freq, coarse freq mod source, fine,
-
- KVRist
- 440 posts since 13 Mar, 2009
Thanks for your reply!
Makes sense. One of two things needs to happen. Either we add MIDI control mapping into the plugin itself or we stick with host mapping and convert all parameters internally to a 0..1 range. Otherwise there is no opportunity (or limited opportunity) for hosts to perform a non-linear mapping that makes practical sense.
Max 6 isn't doing anything to understand the internal parameter values and just mapping to 0..1.
Right now I feel more inclined to go with the latter and convert parameters to 0..1. That way they will automatically mirror the non-linear mapping in the UI.
Which begs the question why did we do that in the first place. I think it allowed some older hosts to provided better feedback in their automation editors.
Thanks!
AudioSpillage
Makes sense. One of two things needs to happen. Either we add MIDI control mapping into the plugin itself or we stick with host mapping and convert all parameters internally to a 0..1 range. Otherwise there is no opportunity (or limited opportunity) for hosts to perform a non-linear mapping that makes practical sense.
Max 6 isn't doing anything to understand the internal parameter values and just mapping to 0..1.
Right now I feel more inclined to go with the latter and convert parameters to 0..1. That way they will automatically mirror the non-linear mapping in the UI.
Which begs the question why did we do that in the first place. I think it allowed some older hosts to provided better feedback in their automation editors.
Thanks!
AudioSpillage
-
- KVRer
- Topic Starter
- 13 posts since 20 Feb, 2004
i vote +1000 for putting the non-linearity into DS. is there any host at all that can do non-linear mapping? and how would it know which kind of non-linearity fits which parameter?
if you would want to go the extra mile you could even let the user control the curveness of the exponential function - for each parameter individually. if that's not luxury I don't know what is. =)
it makes sense that Max doesn't do anything since it's much more programming language than software. although it would be nice if it would deal correctly with values >1.
if you would want to go the extra mile you could even let the user control the curveness of the exponential function - for each parameter individually. if that's not luxury I don't know what is. =)
it makes sense that Max doesn't do anything since it's much more programming language than software. although it would be nice if it would deal correctly with values >1.
-
- KVRist
- 440 posts since 13 Mar, 2009
In AU there is some mechanism for providing a suggested non-linear mapping but it won't match the mappings we use in the UI. I'm not sure how many hosts follow it either to be honest.
The model parameters work fine because they are all 0..1. This is because those parameters vary depending on the model. However, I think we'll move everything over to 0..1 mapping internally. Might also help when we add the 'Randomizer' back in.
Thanks!
AudioSpillage
The model parameters work fine because they are all 0..1. This is because those parameters vary depending on the model. However, I think we'll move everything over to 0..1 mapping internally. Might also help when we add the 'Randomizer' back in.
Thanks!
AudioSpillage
-
- KVRer
- Topic Starter
- 13 posts since 20 Feb, 2004
sounds great.
since you didn't address it i just want to make sure you didn't read over this bug report:
* Live 9.1: if i map the same CC to several parameters only the first one reacts on changes ("first" means the first parameter in the list of parameters which I made visible to MIDI remote control by using the Configure button).
since you didn't address it i just want to make sure you didn't read over this bug report:
* Live 9.1: if i map the same CC to several parameters only the first one reacts on changes ("first" means the first parameter in the list of parameters which I made visible to MIDI remote control by using the Configure button).
-
- KVRist
- 440 posts since 13 Mar, 2009
Strange, will look into that. Presume this isn't happening with others plugs?pure wrote:sounds great.
since you didn't address it i just want to make sure you didn't read over this bug report:
* Live 9.1: if i map the same CC to several parameters only the first one reacts on changes ("first" means the first parameter in the list of parameters which I made visible to MIDI remote control by using the Configure button).
AudioSpillage
-
- KVRer
- Topic Starter
- 13 posts since 20 Feb, 2004
not happening with other plugins here. I also tried Live 8 and DS1 - same misbehavior.
however, sometimes (rarely) one of the "overshadowed" parameters might move a bit.
however, sometimes (rarely) one of the "overshadowed" parameters might move a bit.
