From: Dave (dfisher_at_pophost.com)
Date: 2002-03-02 21:20:09
On Sat, 2 Mar 2002 20:30:42 +0000, Jim Peters wrote:
>Right, I've just had a chance to look. The '-' bug isn't the only one
>you're seeing -- do you see those thin lines across a couple of the
>displays. There is something weird going on at a few very specific
>frequencies that I haven't tracked down yet. It stays as a single
>line even when you enlarge. It also isn't there all the time. I
>wonder if somehow I'm corrupting the FFTW tables. I have an idea
>about what might be doing this.
I did notice, but forgot to mention it. At first I thought it was just a
display corruption, but then the "line" followed even while shifting octaves up
and down. But as you, I found that it comes and goes. I wish I had a better
understanding of the FFTW routines, or I could offer something more helpful.
:-(
>> Anyway, I have put instructions on how to use the data files with
>> BWView on the page.
>
>Thanks. It's very useful to have some more data files to test with,
>especially as your files show up some new strange things going on.
>For instance, your files show a huge amount of high-frequency junk
>between 256/3 and 256/2 -- however it seems to come and go as the
When you say 256/3 -- where exactly do you mean? If you have a byte offset, or
time correlation, I can look at the data via BioGraph and perhaps be able to
provide some more feedback to you this way.
>position is moved -- so it might be a signal going in and out of phase
>with my convolution function. Eeek! Making a nice wide window helps
>somewhat. But I'm still not clear whether it is your data or my
>analysis. I'll look at this some more.
The subject was pretty good about not moving during the recording session. He
was having a fun time seeing how he could make all the lines of the FFT chart
"explode" by clenching his jaw and changing his facial expression. I let him
have fun with this, but asked that he not do that during the session. I did
notice a couple of places, though, where the amplitudes soar and it kind of
looks like what I see when there is body movement introduced into the
recording. So maybe he snuck one or more in while I was not looking. :)
>> I'll tell you Jim--I am really struck with just how *beautiful* the
>> images are. I think you have created a new art form. Now, if I
>> could just fully understand how to interpret that art... :)
>
>I think it's really tempting to get lost in the shapes -- it ends up
>being like fractals. I was doing that, too. It helps to have a
>purpose for looking at the file, I think! Anyway, I have some gaudily
>coloured displays for the next version.
lol! That sounds splendid. Yeah, I'll have to get past the brainwave-gazing
stage and dig into the math behind the art.
>Regarding your subject, I have no idea. There is delta all over the
>place, at about the same level as the alpha waves. But I'm just
>starting to learn about interpreting this stuff. I'm also beginning
>to understand Jim's need to see both channels at the same time.
I know; I was struck by the amount of delta, but as you, am still learning to
interpret what I am seeing.
Dave.
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:39 BST