wasn't going to post this initially but I remembered a plugin i made earlier that raises a real number to an /imaginary/ power and I thought about FFT with complex number pairs, and put two and two together and it worked somehow.
http://ift.tt/2bgmk8f
http://ift.tt/2bgmk8f
bottom drive parameter is the main drive, eventually it decimates into dc offset/asymmetry hell but before that there's a sweet spot that sounds pretty good. The top drive just loudens the input sample before convolving it, not sure if it affects things but it's there for the curious. If tanh slider is on, there is a saturation stage before everything else, which might sound nice if you know it's the last effect in your chain, or otherwise. Dc slider adds dc offset. Bypass bypasses the convolution.
You can also use mousewheel up/down to toggle tweaking a bunch of behind the scenes parameters (192 at this FFT size, would double if anyone can handle the next biggest FFT), which are done all at once cause they're probably not very effective/noticeable doing one at a time. Basically they go off at different angles according to the mouse movements, some will overdrive, others will under-drive. Just try it, it'll make more sense when actually doing it.
The concept is a convolution plugin a la Nebula, but with this an impulse/texture is generated every sample(!) and varies according to the input sample which is nudged around with all those 192 parameters. I don't quite understand Nebula/volterra kernels but I assume that it is a convolution with a few impulses to choose from based on input volume, and some kind of crossfade/morph between them? This plugin, with any luck, will create a unique sound image/variation for every possible floating point value, i.e. not static and hopefully as dynamic as possible despite not using real captures of analog equipment
http://ift.tt/2bgmk8f
http://ift.tt/2bgmk8f
bottom drive parameter is the main drive, eventually it decimates into dc offset/asymmetry hell but before that there's a sweet spot that sounds pretty good. The top drive just loudens the input sample before convolving it, not sure if it affects things but it's there for the curious. If tanh slider is on, there is a saturation stage before everything else, which might sound nice if you know it's the last effect in your chain, or otherwise. Dc slider adds dc offset. Bypass bypasses the convolution.
You can also use mousewheel up/down to toggle tweaking a bunch of behind the scenes parameters (192 at this FFT size, would double if anyone can handle the next biggest FFT), which are done all at once cause they're probably not very effective/noticeable doing one at a time. Basically they go off at different angles according to the mouse movements, some will overdrive, others will under-drive. Just try it, it'll make more sense when actually doing it.
The concept is a convolution plugin a la Nebula, but with this an impulse/texture is generated every sample(!) and varies according to the input sample which is nudged around with all those 192 parameters. I don't quite understand Nebula/volterra kernels but I assume that it is a convolution with a few impulses to choose from based on input volume, and some kind of crossfade/morph between them? This plugin, with any luck, will create a unique sound image/variation for every possible floating point value, i.e. not static and hopefully as dynamic as possible despite not using real captures of analog equipment
fakewarm (tape-y plugin)
Aucun commentaire:
Enregistrer un commentaire