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

replace

[> 10]
|
[select 1]

with [select 2] (2 being the specific coordinate). You are right in hooking up that same value into both [pd direction] and [select 2]. However if the coordinates moves across far to quickly it may jump past 2 and not get noticed by select. So you have to find a different way of catching it.

Can you think of a better way?

hi, thanks for all the help so far. just one more question...:)

I have put all the stuff in but it doesn't work. What i have done is made my x grid dimensions to 50 so that there was more control with more numbers. I then want to divide it by 10 so that the numbers outputting from the grid would be 1-5. However right now it is spitting out numbers 1-10! What I then want to do is that when the drection patch is reading that the direction is going right, it will times the number by 10 and when it goes left it will times it by 1. Then I can input the files at numbers 1-5 and 10, 20, 30, 40, and 50.

Right now, I am not sure if the direction patch is working properly. So am slightly stuck.

The other things I would like my patch to do and I dont know if it is possible. Is that if nothing has been banged for 2 minutes then it plays another movie (like a screensaver)

Also, for the concept of my project, if one person is walking along then I don't want another person to walk along straight after and for the animation to get confused, so is it possible that if one bang goes off it freezes the other bangs except from the ones side by side?

Sorry I know I am being pretty picky with what I am wanting. This is my 4th year university project and as you can see I am leaving it a little late and can't find the answers anywhere.

Many thanks for everything and would appreciate any help I can get.

Thanks

Sarah

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

hi, thanks for all the help so far. just one more question...:)

I have put all the stuff in but it doesn't work. What i have done is made my x grid dimensions to 50 so that there was more control with more numbers. I then want to divide it by 10 so that the numbers outputting from the grid would be 1-5. However right now it is spitting out numbers 1-10! What I then want to do is that when the drection patch is reading that the direction is going right, it will times the number by 10 and when it goes left it will times it by 1. Then I can input the files at numbers 1-5 and 10, 20, 30, 40, and 50.

Right now, I am not sure if the direction patch is working properly. So am slightly stuck.

The other things I would like my patch to do and I dont know if it is possible. Is that if nothing has been banged for 2 minutes then it plays another movie (like a screensaver)

Also, for the concept of my project, if one person is walking along then I don't want another person to walk along straight after and for the animation to get confused, so is it possible that if one bang goes off it freezes the other bangs except from the ones side by side?

Sorry I know I am being pretty picky with what I am wanting. This is my 4th year university project and as you can see I am leaving it a little late and can't find the answers anywhere.

Many thanks for everything and would appreciate any help I can get.

Thanks

Sarah

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

sarah, if no one beats me to it I'll try have a crack at it tomorrow, just have been so busy with my own stuff and too tired right now. I understand you have those looming deadlines.

btw, You're not 4th year at Brighton are you?

ahh, your patch is starting to become a mess.

Plus, you're still implementing the directionPatchneweee.pd wrong. Get rid of the [> 10] altogether. (I only used that for MY debugging purposes). now it works fine.

if divide by 10 isn't working properly, you're not dividing it right somehow.

do this

[nbx]
|
[/ 10]
|
[nbx]

and you will see it works correctly

To do what you want is straight forward. The output of [select] only outputs a bang (strangely [route] doesn't seem to do the job) so you need to recall the co-ordinate value [v] or /[r]. if using send/receive receive into the cold inlet of yet another [f] (one for each direction) box then the output of [select 0 1] into their relevant hot inlets. this way the [f] boxes are constantly storing the co-ordinate value, and when banged send it on it's way to either

• or
• now you can continue with business as usual.

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