Hi, I am currently working on a university project, I am using motion detection to play video. I have this working fine that when different points are triggered different clips play. However. I was wondering if someone new a way I could stipulate the direction that the motion has come from. The idea I have is that if someone walks along the camera then a ballet dancer animation will follow them, however if they turned around i would want to flip the animation so the dancer still follows them. Hope this males sense, and I would be grateful for any help.

Many thanks, Sarah

Hey guys,

I tried playing around with this patch to try and figure out what's going on (I don't really do video in Pd, so pdp is pretty foreign to me), so I'll just throw into this conversation a few things I've noticed. First, the problem with [pdp_mgrid] spitting out 1-10 is probably because you're forgetting to send that [50( message to it. Hook a [loadbang] up to it so that it happens automatically when the patch opens.

Also, I messed with this using my webcam, and it seems to jump around somewhat erratically. (This, by the way, is after using some of Arif's most recent suggestions. I didn't implement the specific coordinate thing, though, because I'm not exactly sure what you want to do with it, and it seems the 1-5 and 10-50 thing may be solving that problem for you anyway...?) My guess is that [pdp_mgrid] is sending out information for every cell that's detecting movement, and with a resolution of 50 it makes sense that many cells are detecting movement at once. This is going to make it harder to detect the direction accurately, as it will likely jump back and forth. Raising the threshold helped somewhat. Also, I played around with using cyclone's [mean] to accumulate x-coordinates and compute the average, with the assumption that the most active cells would be spit out more often. I don't know if this is the best way to go about it, but it seemed to help, though it will take some tweaking to figure out a good balance with the threshold and the number of x-coordinates to average. Here's how I set it up:

[inlet x-coordinates]
|
| [r clearit]
| |
| [clear(
|/
[mean] <-left outlet spits out average, right is number of numbers accumulated
| /
| [sel 20] <-set argument to the number of x-coords you want to average
\ /\
X
/ \
[ f ]
|
[outlet x-average]

I had the outlet go to the direction subpatch. Hope that helps at all.

hi,

thanks to both of you for all your help. My patch is slowly getting there (and yes still a bit messy ) It is a combination of things you have suggested and some help from people within the University.

Right now i have added a metro counter in on the right of the patch, so that whenever there is no activity for 9 seconds a screensaver plays. I have also got the direction sensor to almost work. However, right now it is not playing any of the files. They are all coming up at the right time into the gem window, however are just showing the first image and not playing the movie. I have included a zip file with all the movies as well. Do you know what may be possible for this problem?

Many thanks again,

Sarah

ps. Arif, I am at Dundee University

http://www.pdpatchrepo.info/hurleur/directionPatchneweee.pd

hi,

thanks to both of you for all your help. My patch is slowly getting there (and yes still a bit messy ) It is a combination of things you have suggested and some help from people within the University.

Right now i have added a metro counter in on the right of the patch, so that whenever there is no activity for 9 seconds a screensaver plays. I have also got the direction sensor to almost work. However, right now it is not playing any of the files. They are all coming up at the right time into the gem window, however are just showing the first image and not playing the movie. I have included a zip file with all the movies as well. Do you know what may be possible for this problem?

Many thanks again,

Sarah

ps. Arif, I am at Dundee University

http://www.pdpatchrepo.info/hurleur/directionPatchneweee.pd

for some reason it did not like my zip file so have just posted the PD files. I f you have any movies you could link up to see if it works for yourself that would be great

Thanks

Sarah

hi,

Forget my last question about the movie not playing...was just me being silly! the only problem I have left is that I want to control which numbers are banged at a time so that if two people are walking along then it wont keep jumping between them, It will just stick with the one. So if the first bang has been initiated, there will only be two other bangs that can be played after it (ie the ones to either side)

Is this possible?

Thanks again
Sarah

I think playing around with [moses] will help you here. Maybe one for the minumum and one for the maximum (i.e. n-1, n+1, where n=cell number) and update the arguments each time a new cell is sent.

How does moses exactly work?

Right now I am using split, but this isn't working the way i need. I have changed the x coordinate to 1-5 as having it up to 50 was just too sensitive.

so basically what i need is that if the number three has just been triggered I need it to play either 4 or -3. Does that make sense, been trying to explain it to other people as well. I have an image that might make it clearer for the direction of the movement that i need, which i will attach. It shows the direction of the movement, and which numbers will be triggered after each other.

Any idea would be welcome, or if you could expain if moses would be able to control it for two separate numbers, that wont be in scale.

Thank you

Sarah

http://www.pdpatchrepo.info/hurleur/Picture_6.jpg

The way I was thinking of using [moses] is like how [split] works:

[inlet]
|
[moses 2] <--minimum
|
[moses 5] <--maximum+1 -- left outlet is <, right >=
|
[outlet]

is the same as:

[inlet]
|
[split 2 4]
|
[outlet]

I was thinking of the inputs to [moses]/[split] as coming from the cell numbers before they go into the direction patch, not the numbers triggering the bangs. That's where the possible second person will be detected, so that's where it should be dealt with. The arguments would then updated to be relative to the incoming cell number.

Hi, thanks for the help before,

I seem to have everything down that I want to.

The splits are meant to be controlling some of the numbers that can be played after each other. Need to fiddle around with them slightly more though.

However. I am still having trouble with the motion sensor. Can anyone tell for me if it has everything in or not?

Thank you all. getting to the final countdown for my degree show now...yikes!

http://www.pdpatchrepo.info/hurleur/grid_trackingnew.pd

Hi, thanks for the help before,

I seem to have everything down that I want to.

The splits are meant to be controlling some of the numbers that can be played after each other. Need to fiddle around with them slightly more though.

However. I am still having trouble with the motion sensor. Can anyone tell for me if it has everything in or not?

Thank you all. getting to the final countdown for my degree show now...yikes!

http://www.pdpatchrepo.info/hurleur/grid_trackingnew.pd