hello, i just started out with MaxMsp and read about Jitter etc. but like i said, just started.

now, i'm also a Phd student in the biology department and i'm studying the aging process.

the two things aren't linked, but reading about Jitter gave me an idea i need to work out very quickly now (and i'm not able to do it myself because i'm noob deluxe)

i want to measure quatitatively the amount of movement over time.

so i have this group of animals (little worms) that lie still. but will slowly regain the ability to move. i want to measure the amaunt/gain of movement over time. (i have two mutant groups and want to compare them)... it is a small test, but i need this done quick...

i will just make video's of the process and use those as input for jitter to analyse.

so i would like to ask if someone has a patch like that? will jitter be able to do this?

like i said i need this quick, otherwise i would search and make all of it myself...

the two things aren't linked, but reading about Jitter gave me an idea i need to work out very quickly now (and i'm not able to do it myself because i'm noob deluxe) 

need-something-quick

yep start searching and start making a small patch

then send a patch and we will give advice....

>like i said i need this quick, otherwise i would search and make all 

yep start searching and start making a small patch
then send a patch and we will give advice....

>
>
>like i said i need this quick, otherwise i would search and make all 
>of it myself...
>
>
>


wohow Mattijs, that's almost exactly what i was looking for!

i tried to at least understand the basics of the patch.

as i see it the actual framerate of te mov file should be the limiting factor in this patch?

basicly this patch sends bangs (frames) when a certainamount of pixels/cells in a matrix change from one frame to the next? is this correct?

so to actually quantise the amount of movement i should maybe have a set of different thresholds and count the amount of bangs over time for each threshold value?

i take it that i will also be able to program a patch that can track individual subject? that would be very cool! because i could eg follow 10 individuals simultaneously from 1 video!

so to actually quantise the amount of movement i should maybe have a set of different thresholds and count the amount of bangs over time for each threshold value? 

Quote: GD wrote on Wed, 07 February 2007 17:33

----------------------------------------------------

> wohow Mattijs, that's almost exactly what i was looking for!

> i tried to at least understand the basics of the patch.

> as i see it the actual framerate of te mov file should be the limiting factor in this patch?

> basicly this patch sends bangs (frames) when a certainamount of pixels/cells in a matrix change from one frame to the next? is this correct?

> so to actually quantise the amount of movement i should maybe have a set of different thresholds and count the amount of bangs over time for each threshold value?

No, the amount of change is equal to the amount of bangs per second (with a maximum of 1 bang per 5 ms, due to the movie framerate). The timer object counts the time between bangs, so the number that is output by timer has an inverse relationship with the amount of change, so will get smaller with more change.

You need only one threshold (the one that's already present). Tweak this threshold to match the global amount of movement in your video, or in other words, the sensitivity of the system. Once you configured it to match your data you'll no longer have to change the threshold. The 65000 I came up with was with very rapidly changing video, I can imagine that bringing it down works much better for worms that move slowly (I didn't have that kind of test material ;) Test it and you'll see the change factor get smaller when you lower the threshold.

> i take it that i will also be able to program a patch that can track individual subject? that would be very cool! because i could eg follow 10 individuals simultaneously from 1 video!

That would definitly be possible, but a factor 100 more difficult to do it right. Look here

http://www.iamas.ac.jp/~jovan02/cv/objects.html

Quote: GD wrote on Wed, 07 February 2007 17:33
----------------------------------------------------
> wohow Mattijs, that's almost exactly what i was looking for!
>
> i tried to at least understand the basics of the patch.
>
> as i see it the actual framerate of te mov file should be the limiting factor in this patch?
>
> this is very encouraging...
>
> basicly this patch sends bangs (frames) when a certainamount of pixels/cells in a matrix change from one frame to the next? is this correct?

>
> so to actually quantise the amount of movement i should maybe have a set of different thresholds and count the amount of bangs over time for each threshold value? 

No, the amount of change is equal to the amount of bangs per second (with a maximum of 1 bang per 5 ms, due to the movie framerate). The timer object counts the time between bangs, so the number that is output by timer has an inverse relationship with the amount of change, so will get smaller with more change. 

>
>
> i take it that i will also be able to program a patch that can track individual subject? that would be very cool! because i could eg follow 10 individuals simultaneously from 1 video!

That would definitly be possible, but a factor 100 more difficult to do it right. Look here 

but in my case i have a movie were the first 5 minutes are without movement. than slowly more and more movement over time untill the animals recovered completely and move around like they normally would.

i need to compare the rate of recovery (so the regain of movement) between two groups.

that's why i came up with the different threshold levels... but indeed maybe that's not the best way to do it.

i just need to record the actual program output somehow (and put the changing output of the program on a timeline) that way i should end up with data (a curve) i can really use!

i need to compare the rate of recovery (so the regain of movement)  between two groups.

Quote: GD wrote on Thu, 08 February 2007 08:53

> i just need to record the actual program output somehow (and put the changing output of the program on a timeline) that way i should end up with data (a curve) i can really use!

I think I do. What about this? It writes the data to a coll every 500 ms. Shouldn't be hard to fit the data into a curve..

Quote: GD wrote on Thu, 08 February 2007 08:53
----------------------------------------------------

> i just need to record the actual program output somehow (and put the changing output of the program on a timeline) that way i should end up with data (a curve) i can really use!
>
> you understand what i mean?

need something quick