| Re: Make TV a function of X? [message #18468] |
Fri, 07 January 2000 00:00 |
wbiagiot
Messages: 59
Registered: January 1999
|
Member |
|
|
In article <854pqn$lpa$1@nnrp1.deja.com>,
raouldukey@my-deja.com wrote:
>
> I am thinking I will have to make some sort of fake data set by
> making a fake time array and put the data in where it exists, and
> fill it with zeros where it doesn't. Yuck! This has to be
> a common problem, so has anyone else thought of a solution?
>
> Cheers
Raoul,
Well here's a long shot. I took interest in your problem because there
is a possibility that I might encounter something similiar.
Try a thumbnail approach:
In window A, display the main waveform over the entire time span
resampled to some reasonable data length (2k-8k pts ?). Make sure that
every pulse GROUP is represented on the plot, even if you have to do it
artifically - avoid aliasing. This is kinda like a thumbnail. Enable
the window for mouse events. Allow the user to place the mouse anywhere
on the plot (choosing a pulse train of interest) and click. Decode the
event, obtain the X position of the mouse and translate it into units of
your X scale. That should determine which group of pulses is being
targeted. Knowing that, pop up a secondary plot window showing that
pulse train in detail - possibly using a plot with a slider on the X
display range if the number of points is large.
The alternative is to have the 2nd window track the mouse in
realtime - decoding on the fly and presenting the data continuously.
The 2nd window would only update if the mouse is over an 'active
region'.
Obviously, this is just an outline with alot of gaps to be filled in. I
have used some parts of this idea on images (which are waveforms anyway)
and am very satisfied with the results. Learned alot of IDL
functionality in the process.
Good Luck,
Bill B.
--
"They don't think it be like it is, but it do."
Oscar Gamble, NY Yankees
Sent via Deja.com http://www.deja.com/
Before you buy.
|
|
|
|
| Re: Make TV a function of X? [message #18470 is a reply to message #18468] |
Fri, 07 January 2000 00:00  |
raouldukey
Messages: 15
Registered: November 1999
|
Junior Member |
|
|
Sorry for the confusion, but you have figured out what I am trying
to do.
>
> raouldukey@my-deja.com writes:
>> I have a data set that is irregularly sampled, and I
>> need to make an image-type plot of this data, and was
>> hoping to get some suggestions on the best way to do this.
>
> It's not totally clear what you are doing here, but I assume that each
> of your samples is actually a *vector*, and that they are irregularly
> sampled in time. But maybe it's more appropriate to say that you have
> uniformly sampled data separated by irregular gaps. Right?
Right! I am dealing with radar sampling. When a pulse is transmitted,
the return echo is uniformly sampled, and is the same from pulse to
pulse. However, the frequency (pulses/sec) of the pulses changes, plus
there are gaps where it was turned off for whatever reason.
are gaps
The image plot style is perfect for this sort of experiment, using
time for the x axis and range for the y axis.
>
> I have had a hard time representing the same kind of data. My best
> technique so far is to label the X axis as "sample number," and let
> them appear uniformly sampled. Then I place a vertical separator
> lines whenever gaps would be. Finally, you might want to label each
> continuous segment with the time if you have room. Not perfect, but
> what can you do?
>
Thanks for the suggestion, however, it won't work in this case. I
have to have a meaningful time axis, so that I can compare to the
other instruments that were operating. Also, I am dealing with
thousands of pulses, so I can't label the time easily.
I am thinking I will have to make some sort of fake data set by
making a fake time array and put the data in where it exists, and
fill it with zeros where it doesn't. Yuck! This has to be
a common problem, so has anyone else thought of a solution?
Cheers
Sent via Deja.com http://www.deja.com/
Before you buy.
|
|
|
|
| Re: Make TV a function of X? [message #18475 is a reply to message #18468] |
Thu, 06 January 2000 00:00 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
raouldukey@my-deja.com writes:
> I have a data set that is irregularly sampled, and I
> need to make an image-type plot of this data, and was
> hoping to get some suggestions on the best way to do this.
It's not totally clear what you are doing here, but I assume that each
of your samples is actually a *vector*, and that they are irregularly
sampled in time. But maybe it's more appropriate to say that you have
uniformly sampled data separated by irregular gaps. Right?
I have had a hard time representing the same kind of data. My best
technique so far is to label the X axis as "sample number," and let
them appear uniformly sampled. Then I place a vertical separator
lines whenever gaps would be. Finally, you might want to label each
continuous segment with the time if you have room. Not perfect, but
what can you do?
> BTW, I am using Craig Markwardt's plotimage.pro proceedure,
> to do my image plots, and I recommend it highly!
Why thanks!
Craig
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
