Sounds like what LV2 is doing with its CV ports. This is a good idea.
Vesa kirjoitti Sun Apr 06 2014 12:19:47 GMT+0300 (EEST): > Tell me does the following make sense: > > - AutomatableModel contains a pointer to a buffer > - When a controller is connected to a model, it gets passed that buffer > pointer > - new buffer is created at connection time (deleted when connection is > disconnected) > - Every period, the controller writes the control values to the buffer > > So far, nothing else is done with that buffer. Only when some thing > (plugin, etc.) wants to read sample-exact values from that model, then > the values in that buffer get translated from the controller's -1-1 > scale to the model values, whole buffer at once. Thus: > - The model will have a method that returns a buffer pointer > - This method first goes through the buffer and translates the values to > the model's scale, then returns a pointer to the buffer, so the > recipient can then use those values for whatever > - When no controller is connected, the method returns a null pointer, > which the plugin/etc checks for, and falls back to the regular value() > method > > I think this might be faster than reading values from a method one by > one. Instead of calling a method in automatablemodel (which maybe calls > other functions) per-frame, we could just retrieve a buffer of values > from the model, per-period. Also this way, we still won't bother > translating all the buffer values if the recipient doesn't want > sample-exact data. > > Does this make sense? > > ------------------------------------------------------------------------------ > _______________________________________________ > LMMS-devel mailing list > LMMS-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/lmms-devel > ------------------------------------------------------------------------------ _______________________________________________ LMMS-devel mailing list LMMS-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lmms-devel
