Is there an error in wconv1.m used to compute CWT in Wavelet Toolbox?

In the wavelet toolbox, the continuous wavelet transform uses a function called wconv1.m to convolve the (integrated) wavelet with the signal to be analyzed. However, this function uses the line:

y = conv2(x(:)',f(:)',shape);

to vectorize and then convolve the two. I think this should be...

y = conv2(x(:).',f(:).',shape);

(i.e. element-wise transposition) as the existing form also conjugates the signals before convolution. The result is thus conjugated with respect to the correct CWT.

The CWT is often used for real signals, or is used only to generate a scalogram (using absolute value of CWT). In both these cases the difference I point out is irrelevant.

However, if I am correct, the conjugation error will cause serious problems in any case where the CWT is used for phase analysis or where causality is important.

Do I have this correct?

regards Donnacha