{"items": [{"author": "Edward", "source_link": "https://www.facebook.com/jefftk/posts/473216142735690?comment_id=473230266067611", "anchor": "fb-473230266067611", "service": "fb", "text": "Function-discovering software here: http://creativemachines.cornell.edu/eureqa", "timestamp": "1357873590"}, {"author": "James", "source_link": "https://plus.google.com/106345404829653994850", "anchor": "gp-1357877827488", "service": "gp", "text": "The gain/level graph looks piecewise linear, with a change in slope at exactly 0.5, and a zero somewhere around 0.1.\u00a0The panning behavior appears to be logarithmic - the level on each side appears to be c*log(d*pan). I figured this out by looking at the graph of L vs Pan at medium gain.\n
\n
\nThere are two possible orderings which these functions can be applied in, and you still have to fit the constants, but I think that's enough degrees of freedom chopped off to make it tractable.", "timestamp": 1357877827}, {"author": "David Chudzicki", "source_link": "https://www.facebook.com/jefftk/posts/473216142735690?comment_id=473252909398680", "anchor": "fb-473252909398680", "service": "fb", "text": "Looks like you're writing in Python? So maybe something like this: http://docs.scipy.org/.../scipy.interpolate...", "timestamp": "1357878936"}, {"author": "Jeff Kaufman", "source_link": "https://plus.google.com/103013777355236494008", "anchor": "gp-1357908721782", "service": "gp", "text": "@James\n\u00a0\"piecewise linear\"\n
\n
\nYup. \u00a0Slope of 1.84 for gain>=50% and 4.13 below that. \u00a0Using this to predict gain from level 'i get: \nhttp://www.jefftk.com/1818vsl-predict-gain-from-level.png\n which seems pretty good.", "timestamp": 1357908721}]}