Acoustic prediction accuracy

Computer model and acoustic prediction accuracy …

Using an actual room (rehearsal room “208” at Leeds college of music) we both “computer predicted” with and without acoustic treatment.  And “physically measured” with and without acoustic treatment for comparison. The result is a measure of our acoustic prediction accuracy.

This room is deliberately small, to make the calculative predictions as hard as possible and show the largest error !
The traditional “Sabine” method of reverberation time calculation requires an average of 100 room reflections to be remotely accurate. This small test room was deliberately chosen to have far less than the required number of reflections. Which, would normally mean, by “Sabine” calculation methods, large prediction errors. Our unique “bi-calculative” and “reverse ray trace” system however, is highly accurate, right down to 100 Hz or so.  Well proven over thousands of projects world wide from home cinema design to recording studio design and beyond.  Our computer prediction accuracy is so high mainly due to our unique “bi-calculative” method. This compares two separate calculations, starting with the original Sabine T=(0.161 x V)/A formula.  A second, rather more complex formula, which takes the geographical position of absorbent surfaces into account is run in parallel.  A cross correlation between them result in a final answer that is dependably and absolutely accurate.

acoustic prediction accuracy graph.jpg


RED Computer predicted, room untreated
LIGHT GREEN Actual measured room untreated
BLUE Computer predicted, room treated
DARK GREEN Actual measured room treated

Measured versus predicted: acoustic prediction accuracy

Above 500 Hz = no error
250 Hz = minimal error & undetectable by ear
125 Hz = minor error & largely undetectable by ear
63 Hz … actual measured traces are flatter than those predicted, so aural outcome better than that predicted !. This low frequency prediction error is due to our “bi-calculative” modeling algorithms inability to represent bass absorption in sufficient detail, but is of no practical consequence since the real outcomes are better than predictions !.

In short … Our acoustic predictions are what you actually get !