I'm making a webm creator for 4chan,

I've generated some samples and created a multiple regression equation

How do I change the equation to determine the correct bitrate for the target file size of 3072?

also for really high iq, how do I effectively adjust the bitrate if the file is too large or too small?

*I collected the data by creating a few webms with different bitrates, and recorded the different variables from the source video

Here's my findings :

fileSize = (duration * durationCoef) + (frames * framesCoef) + (origSize * origSizeCoef) + (bitrate * bitrateCoef) + const

These are the regression results:
durationCoef = 66008226.82677362
framesCoef = -1118782.5575664348
origSizeCoef = 2.6696120497622724e-05
bitrateCoef = 7.069421090497821e-05
const = -510.20052375


the bitrate coefficient is derived from the bitrate that was passed to the webm creator.

obviously I would like to change the bitrate so as to achieve a 3mb webm

also my r2 in the model was .74