Complex numbers and MatrixOp

Share icon

Fri, 10/21/2011 - 08:50 am

Hello,

While working on some code that utilizes complex numbers in matrices, I noticed that the MatrixOp/c operation doesn't always seem to work properly. To simplify an example, I made two complex matrices (B and A, 2x2) and performed MatrixOp/o/c C=B x A to perform matrix multiplication.

Afterwards, I changed one number in matrix B and repeated the operation. In some cases, matrix C was calculated correctly, while in other cases it did not update at all, even after repeatedly issuing the command.

I've compiled a word document showing the matrices used, changes made, and resulting calculated matrix C.

Any thoughts to this problem? Am I performing the matrix multiplication incorrectly?


Thank you. Strange MatrixOp glitch Complex matrix.doc
Hello Proland,

proland wrote:

I've compiled a word document showing the matrices used, changes made, and resulting calculated matrix C.


If you think that MatrixOP does not give you the correct results please send an *IGOR* experiment containing the necessary information to support@wavemetrics.com

A.G.
WaveMetrics, Inc.

October 21, 2011 at 09:38 am - Permalink

It seems that a variety of mistakes led to my misunderstading. When I had my code print the indivdual matrices, I assumed the values printed in the command history were by row (turns out that they print by column for some reason). As a result, I was comparing the off-diagonal elements to the wrong cell in my manual calculation.

In addition, when doing the test described in my original post, it seems that if you do not click outside of the cell that was changed in the table, it isn't actually updated. So while I saw a new matrix in the table, the previous values were used in the calculations since the actual wave hadn't been updated. Sorry for the confusion.

October 24, 2011 at 01:23 pm - Permalink