alternative to [buddy] [match] hence to massive input
hi!
please have a short look at my example:
as soon as all cells (coordinates) have actually been set -and only then- I seek for output action. I'm having trouble finding an object that knows what to expect and reacts once this pensum is achieved.
I'm familiar to [buddy] and [match] - but in a case of 120x120 pairs of cell coordinates rather cumbersome..
or someone tells me how to easily extract my matrice's 14.400 cell coordinates in a list :)
thankful for any hint..!
all the best
-jonas
some more info is needed as to what you want to look for / compare to, but I'd play with jit.spill/jit.iter and zl group/compare
hey timo,
thanks for having a look..
as soon as every cell of my matrice has been set, it should start from the beginning (clearbang, so to say)..
jit.iter is my friend, as jit.spill only outputs my matrice's values.. and [zl group]/[zl compare] also sound quite promising..
the only confusion at this point: the coordinates of a 120x120 jitter matrix don't fit into my [message] object :)
any suggestions on how to deal with all this data (respectively giving [zl compare] something to work with)?
cheers!
-jonas
hey timo!
I did go for [jit.iter] - although, in the end, I decided to work with my matrice's values rather than its cell coordinates. this way I didn't need the help of [zl] after all
here's my patch so far, if you're interested (of course it gets neater the bigger the matrice is - but it would take ages to show how it works).. it's a bit messy
as soon as all cells have been "whitened", the procedure starts from the beginning.
cheers!
-jonas
I'm probably misunderstanding what you are after since what I have here might be more easily done if you use jit.noise...
But if I understand correctly, you would like to set each pixel of your matrix with a value and once every possible location has been computed you want an output. Well, this should work then... of course it's not overly exciting right now since I am randomly assigning 0 and 255 as values... to make it actually interesting and useful is up to you :)
ahoy,
there are a few other things which can be optimised here.
1.)
[pak setcell 0 0 val 0]
2.)
remove all numberboxes! you dont need them they only eat up CPU.
3.)
with a metro at 2 milliseconds and no need to sequence dance music
or audio for a theater application but "only" standalone video work, you
might want to put a [defer] behind your metro.
and/or turn off overdrive. (in case you run audio at all somewhere in the runtime.)
4.)
oh, and instead of random 4, + 1 you could use expr ... i am not exactly
sure if that is really a performance boost ... dont use float values inside
the expr rand then it might be, simply because you saves one connection
between objects ... metro likes that alot.
-110
sorry the pak idea is wrong, it would change the patch. but you can put it in a pack and let
input 2 trigger the pack.
:)
are you sure you need the [clip]?
hey roman!
with [clip] my coordinates at least stay within my matrix
thanks for the improvements, I'm not really familiar with performance issues (yet)
3) qmetro should do..?
4) show me :)
@MIB: I'm familiar with jit.noise, but in this case prefered the "ant tunnel effect" (you'll know what I mean when using (much) larger matrices). this will actually be exhibited on a television (in combination with MSP sound).. so you have a starting point, and after some minutes black screen turns to white screen :)
i dont know offhand but i think [expr rand(15)] should do the same than your two objects.
(check if output is correct.)
and voila - one message less has to be sent on maxmsp level from one object to another.
the correct synthax is: rand(1,5)
what i told you - as well as what you wrote - cant work. :)
p.s. wow it is really hard to see that comma on my new flatscreens. on my CRTs no problem.
"p.s. wow it is really hard to see that comma on my new flatscreens. on my CRTs no problem."
you can say that again! didn't notice it until now :)
but nonetheless: http://jonaswolf.com/rand2.tiff (sorry)
now i am confused.
aha, ok. it is actually expr random()
try this [expr "random(1,5)"]
and here's how I would do the matrix checking
btw do you know [urn]?
and [drunk]?
thanks, timo..! not a day without learning =)
and yea, I'm familiar with [drunk] & [urn]
I'll upload an example video later on of how it's meant to be - I actually like the way the cells spread.. the "fuller" it gets, the less changes appear on screen. I think it's a bit neater than having my screen "whitened" after exactly 16 bangs. and the larger the matrix gets, the more calming, even at high speed.. (metro [2] i. e.)
..even though not much happens in my jit.window, I know there's alot going on :) and it's funny to wait for the last gaps to close.. it's gets nice starting from 30x30 matrices
here you go :)
I like the end bit, not only due to the music clearing with proceeding cell spread (max @ableton)
at the point were the picture almost freezes one can still witness an event. sometimes it takes ages.. and it's kinda cool when it eventually does take place (although my patch restarts after the song has finished - whether the matrice has completely filled or not. I'm expecting people to run off after a minute anyway, hehe)
once more, a big thanks for help 'n hints!
all the best
-jonas
haha, thanks... the power of random love
Hi,
since you posted in max forum, you got a bit max-centric replies. Massive input=Jitter.
I'd generally try to stay away from jit.spill/jit.iter and stay in matrix land as long as possible.
Here's a solution you might get in jitter forum:
good point, I was already looking for jit.sum or a sum option in jit.expr, but I forgot about jit.3m
!
elegant. thanks, nesa..
assuming I had the jitter part "settled", I was initially seeking for max objects (cf. topic title).. but you're right, it's clearly a jitter playing field
all the best
-j
ps: oh, and while we're "at it" - [qmetro 10] as good as [metro 10 @defer 1]?
for anyone who's interested, here's "my" patch which belongs to the power of love clip posted above - with nesa's refinement: