On 11/16/2014 04:22 PM, Vesa wrote: > On 10/31/2014 06:44 AM, Vesa wrote: >> So, we're pretty much ready to release 1.1, but... >> >> Ok, so I don't really like to rip on other people's code, there's >> definitely problems with some of my own code as well so can't really >> point fingers... however, lately, I've been thinking that the current >> implementation of logarithmic models is flawed. > Ok I looked at this today and it seems that the issue is a bit trickier. > I don't think we can fix it without some serious changes to how > automations work, which I'm not sure if we should even make. > > I remembered wrong about knobs, seems that knobs already handle > logscales just like they should. But with automations, well... there's > the problem of interpolation, ie. if we linear-interpolate on the raw > values, then the curve won't be linear in the logarithmic scale... so > the only option is for the values be interpolated and then convert to > logarithmic, like we're doing now. > > It's not pretty but it works, so we should probably leave it alone for > now. Backwards compat will suffer but oh well. At least we have logscale > knobs (which aren't really actually even logscale, just a sort of > approximation, because real logscale wouldn't work with zero or negative > values). > > So no need to hold back 1.1 for this I guess.
In 2.0, we might handle this by making it so that automation patterns only store "relative" values, ie. values between 0-1, corresponding to min/max values of the model. This way automations would be portable, and the same automation could control multiple knobs easily. Then the scale type could simply be taken in account when the values would be scaled from the 0-1 range to the model's range. ------------------------------------------------------------------------------ Comprehensive Server Monitoring with Site24x7. Monitor 10 servers for $9/Month. Get alerted through email, SMS, voice calls or mobile push notifications. Take corrective actions from your mobile device. http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk _______________________________________________ LMMS-devel mailing list LMMS-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lmms-devel
