Smooth plugin automatability (with test results)
-
- KVRist
- Topic Starter
- 45 posts since 8 Jun, 2010 from Los Angeles
I wrote up a post on our approach to automation & how it compares to other plugins: http://tonal.goodhertz.co/why-automatability-matters/
-
- KVRian
- 573 posts since 1 Jan, 2013 from Denmark
Thanks for your commercial. To actually make a thread: Where is the responsibility - the DAW or the plugin? Many DAW's operate to varying degrees of smooth interpolations (in technical terms; how often setParameter() is called), which theoretically can alleviate this issue. I also recall one member here advocating the use of linear smoothing which produce less appealing plots, but avoids issues of unwanted oscillations existing with cosine interpolations, etc.
What is your opinion on reactiveness vs. smoothness? You can produce even better plots on your plugins by smoothing automations even more, but given we're operating in a domain where non-linearities and artifacts are desirable; I know I for one have often created interesting musical/creative effects by exploiting sudden/extreme automations, especially in combinations with complex filters.
Perhaps the best solution is letting the user set the amount of 'smoothness'.
What is your opinion on reactiveness vs. smoothness? You can produce even better plots on your plugins by smoothing automations even more, but given we're operating in a domain where non-linearities and artifacts are desirable; I know I for one have often created interesting musical/creative effects by exploiting sudden/extreme automations, especially in combinations with complex filters.
Perhaps the best solution is letting the user set the amount of 'smoothness'.
-
- KVRian
- 1379 posts since 26 Apr, 2004 from UK
There is only so much a DAW can do. If the algorithms used by the plugins are not made for parameters automation, there is nothing the DAW can do (typically the form used for IIR filters).
-
- KVRAF
- 7402 posts since 17 Feb, 2005
What does it matter that the amount of distortion added via automation is small, if it is audible? A single number doesn't describe the quality of the automation filter, just the same as everything else.
Also I really question taking any approach to the problem that doesn't consider the algorithm that the parameter smoothing is applied to. Where do these numbers come from?
Also I really question taking any approach to the problem that doesn't consider the algorithm that the parameter smoothing is applied to. Where do these numbers come from?