CarlR

I have a patch where I am triggering four videos with a little sub-patch that uses the tempo object as a common time base. This works in conjunction with some counters and select objects. The patch allows me to reverse engineer simple generative compositions and then play the individual elements (usually four) on real instruments recorded as video. The video part of the patch is quite simple and uses messages received from the trigger patch.

Loop-lengths of all the videos are different and relate to the re-triggering of musical parts reflecting the settings of the trigger sub-patch.

My problem is that although the patch works and triggers the videos it is sometimes very slightly out of time. It will go back into time after a revolution of each loop has completed. Most of the time all is well and the rhythm of each part is synced. All it takes is a little glitch in the re-triggering of the video to make the next revolution slightly out.

I am sure this is not down to the trigger patch but to do with the re-starting of the videos. I have used the loadram message but this dose not help.

My media is light-weight at the moment, each video is a 320x240 medium quality photo jpeg. I would post the patch but as the media is so specific it would make no sense without it.

Basically I need better accuracy of the video re-triggering so that my patch plays back music with NO timing errors. The patch runs for up to about a minute sometimes without problems. I have 2gb of ram and a 2.16mHz intel cpu macbookpro. 

Sorry for not being able to be concise, but some things need a bit of detail. Your thoughts are always welcome

My problem is that although the patch works and triggers the videos it is sometimes very slightly out of time. It will go back into time after a revolution of each loop has completed. Most of the time all is well and the rhythm of each part is synced. All it takes is a little glitch in the re-triggering of the video to make the next revolution slightly out. 

My media is light-weight at the moment, each video is a 320x240 medium quality photo jpeg. I would post the patch but as the media is so specific it would make no sense without it.
Basically I need better accuracy of the video re-triggering so that my patch plays back music with NO timing errors. The patch runs for up to about a minute sometimes without problems. I have 2gb of ram and a 2.16mHz intel cpu macbookpro. 
Sorry for not being able to be concise, but some things need a bit of detail. Your thoughts are always welcome
Carl


accuracy-triggering-videos

Not seeing your patch makes it very difficult to

offer the kind of specific advice you appear to 

You could try deriving your metronome from audio

rate objects - some people don't use metronomes

at all for timing stuff. For some examples of this

approach in action, you should poke around in the

examples>sequencing-looping>audio-rate-sequencing-looping

Not seeing your patch makes it very difficult to
offer the kind of specific advice you appear to 
wish to have. 

You could try deriving your metronome from audio
rate objects - some people don't use metronomes
at all for timing stuff. For some examples of this
approach in action, you should poke around in the

I had a very similar problem, I'm not an expert on video so this is 

probably a not very elegant solution, but it works solidly. loadram 

did not work well for me either, btw. I also see you are already 

using counters... are you using them to run the video? if not, read 

I do not use the jit.qt.movie loop point messages as these only seem 

to work excelently in the tutorial example ;)

I edited together all my video into a long file about 2.8 GB long. 

(and I only have 2GB ram). Then I load this file into the 

jit.qt.movie and stop it (it plays when I just load it), then I just 

route a metronome to a counter that frame advances the movie at video 

rate (25fps in the UK)... then I can keep track of where I am in the 

Because I am running the video from a counter object, I can then send 

a frame number to jump somewhere, or to loop and it works deadly 

accurately... I create a list of cues and then send a number to the 

list which then bangs out the frame number, etc, etc...

This seems always better than the jit.qt.movie object running at its 

I would like to know if there is a better general approach than this 

for rhythmic accuracy... the nice VJ patch in the jitter recipes 

pauses ever so slightly when it changes clip, but I cannot tolerate 

even that, so the metro-counter combination works the best. Then 

again as usual I am just hacking things together so I'm open to 

Maybe this is too simple, but I thought I'd mention it.

On 19 Dec 2006, at 16:21, Carl Rohumaa wrote:

> I have a patch where I am triggering four videos with a little sub- 

> patch that uses the tempo object as a common time base. This works 

> in conjunction with some counters and select objects. The patch 

> allows me to reverse engineer simple generative compositions and 

> then play the individual elements (usually four) on real 

> instruments recorded as video. The video part of the patch is quite 

> simple and uses messages received from the trigger patch.

> Loop-lengths of all the videos are different and relate to the re- 

> triggering of musical parts reflecting the settings of the trigger 

> My problem is that although the patch works and triggers the videos 

> it is sometimes very slightly out of time. It will go back into 

> time after a revolution of each loop has completed. Most of the 

> time all is well and the rhythm of each part is synced. All it 

> takes is a little glitch in the re-triggering of the video to make 

> I am sure this is not down to the trigger patch but to do with the 

> re-starting of the videos. I have used the loadram message but this 

> My media is light-weight at the moment, each video is a 320x240 

> medium quality photo jpeg. I would post the patch but as the media 

> is so specific it would make no sense without it.

> Basically I need better accuracy of the video re-triggering so that 

> my patch plays back music with NO timing errors. The patch runs for 

> up to about a minute sometimes without problems. I have 2gb of ram 

> Sorry for not being able to be concise, but some things need a bit 

> of detail. Your thoughts are always welcome

I had a very similar problem, I'm not an expert on video so this is  
probably a not very elegant solution, but it works solidly. loadram  
did not work well for me either, btw.  I also see you are already  
using counters... are you using them to run the video? if not, read  
on...

I do not use the jit.qt.movie loop point messages as these only seem  
to work excelently in the tutorial example  ;)

I edited together all my video into a long file about 2.8 GB long.  
(and I only have 2GB ram). Then I load this file into the  
jit.qt.movie and stop it (it plays when I just load it), then I just  
route a metronome to a counter that frame advances the movie at video  
rate (25fps in the UK)... then I can keep track of where I am in the  
frame count.

Because I am running the video from a counter object, I can then send  
a frame number to jump somewhere, or to loop and it works deadly  
accurately... I create a list of cues and then send a number to the  
list which then bangs out the frame number, etc, etc...

This seems always better than the jit.qt.movie object running at its  
own internal speed.

I would like to know if there is a better general approach than this  
for rhythmic accuracy... the nice VJ patch in the jitter recipes  
pauses ever so slightly when it changes clip, but I cannot tolerate  
even that, so the metro-counter combination works the best. Then  
again as usual I am just hacking things together so I'm open to  
suggestions.

>
> I have a patch where I am triggering four videos with a little sub- 
> patch that uses the tempo object as a common time base. This works  
> in conjunction with some counters and select objects. The patch  
> allows me to reverse engineer simple generative compositions and  
> then play the individual elements (usually four) on real  
> instruments recorded as video. The video part of the patch is quite  
> simple and uses messages received from the trigger patch.
>
> Loop-lengths of all the videos are different and relate to the re- 
> triggering of musical parts reflecting the settings of the trigger  
> sub-patch.
>
> My problem is that although the patch works and triggers the videos  
> it is sometimes very slightly out of time. It will go back into  
> time after a revolution of each loop has completed. Most of the  
> time all is well and the rhythm of each part is synced. All it  
> takes is a little glitch in the re-triggering of the video to make  
> the next revolution slightly out.
>
> I am sure this is not down to the trigger patch but to do with the  
> re-starting of the videos. I have used the loadram message but this  
> dose not help.
>
> My media is light-weight at the moment, each video is a 320x240  
> medium quality photo jpeg. I would post the patch but as the media  
> is so specific it would make no sense without it.
> Basically I need better accuracy of the video re-triggering so that  
> my patch plays back music with NO timing errors. The patch runs for  
> up to about a minute sometimes without problems. I have 2gb of ram  
> and a 2.16mHz intel cpu macbookpro.
> Sorry for not being able to be concise, but some things need a bit  
> of detail. Your thoughts are always welcome
> Carl
> --
> Some links to my stuff n that :-) http://zappa.tvu.ac.uk/~crohumaa/ 
> CV/cv.php


If you post or attach a patch it will probably be possible to give you a direct answer that solves your problem. To me it sounds like only one hint should be enough but there are a lot of possibilities depending on the way you built your patch.

Ok maybe this will help you understand what I am doing. This is just the timing part of the patch. It is all commented. I am sure the timing is exact enough for music. Well I hope the tempo object is in errr tempo right?

If there is a better way I would love to know. Not sure how I would drive my video from a counter as julio says he is doing in this senario. I am already using counters to give me the exact number of beats in the bar that I need. I understand what julio is doing but am confused as to whether this is possible in my situation. What will be the best way to trigger my video in this situation. Currently I am just sending a [time 0] message.

If there is a better way I would love to know. Not sure how I would drive my video from a counter as julio says he is doing in this senario. I am already using counters to give me the exact number of beats in the bar that I need. I understand what julio is doing but am confused as to whether this is possible in my situation. What will be the best way to trigger my video in this situation. Currently I am just sending a [time 0] message.
Am I due another complete re-think? :)

Quote: CarlR wrote on Wed, 20 December 2006 13:43

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

> Ok maybe this will help you understand what I am doing. This is just the timing part of the patch. It is all commented. I am sure the timing is exact enough for music. Well I hope the tempo object is in errr tempo right?

This is a neat patch and I don't see anything wrong with it that may cause bad timing.

As you said in your first post: "I am sure this is not down to the trigger patch but to do with the re-starting of the videos. I have used the loadram message but this dose not help." The video part is not in your example though.

The very best thing would be to narrow down your problem to a basic patch that allows me (or us) to follow a few steps and then see the problem happen.

Quote: CarlR wrote on Wed, 20 December 2006 13:43
----------------------------------------------------
> Ok maybe this will help you understand what I am doing. This is just the timing part of the patch. It is all commented. I am sure the timing is exact enough for music. Well I hope the tempo object is in errr tempo right?

This is a neat patch and I don't see anything wrong with it that may cause bad timing. 

As you said in your first post: "I am sure this is not down to the trigger patch but to do with the re-starting of the videos. I have used the loadram message but this dose not help." The video part is not in your example though. 

The very best thing would be to narrow down your problem to a basic patch that allows me (or us) to follow a few steps and then see the problem happen. 

cool patch, but... I suspect the prob is the jit.qt.mov...

On 20 Dec 2006, at 12:43, Carl Rohumaa wrote:

> Ok maybe this will help you understand what I am doing. This is 

> just the timing part of the patch. It is all commented. I am sure 

> the timing is exact enough for music. Well I hope the tempo object 

> If there is a better way I would love to know. Not sure how I would 

> drive my video from a counter as julio says he is doing in this 

> senario. I am already using counters to give me the exact number of 

> beats in the bar that I need. I understand what julio is doing but 

> am confused as to whether this is possible in my situation. What 

> will be the best way to trigger my video in this situation. 

> Currently I am just sending a [time 0] message.

cool patch, but...  I suspect the prob is the jit.qt.mov...
metro-counter is your friend  ;)

>
> Ok maybe this will help you understand what I am doing. This is  
> just the timing part of the patch. It is all commented. I am sure  
> the timing is exact enough for music. Well I hope the tempo object  
> is in errr tempo right?
>
> If there is a better way I would love to know. Not sure how I would  
> drive my video from a counter as julio says he is doing in this  
> senario. I am already using counters to give me the exact number of  
> beats in the bar that I need. I understand what julio is doing but  
> am confused as to whether this is possible in my situation. What  
> will be the best way to trigger my video in this situation.  
> Currently I am just sending a [time 0] message.
> Am I due another complete re-think? :)
>
>
> max v2;
> #N vpatcher 130 44 862 731;
> #N comlet sel 1;
> #P inlet 608 372 15 0;
> #N comlet sel 1;
> #P inlet 576 372 15 0;
> #N comlet sel 1;
> #P inlet 540 372 15 0;
> #N comlet sel 1;
> #P inlet 507 372 15 0;
> #N comlet bpm;
> #P inlet 188 156 15 0;
> #N comlet START STOP;
> #P inlet 236 156 15 0;
> #N comlet RESET COUNTERS;
> #P inlet 137 156 15 0;
> #P window setfont "Sans Serif" 9.;
> #P window linecount 1;
> #P message 363 156 20 196617 16;
> #P message 306 156 14 196617 1;
> #P message 258 156 20 196617 80;
> #P message 534 280 20 196617 51;
> #P message 505 280 20 196617 43;
> #P message 478 280 20 196617 35;
> #P message 451 280 20 196617 27;
> #P newex 409 114 48 196617 loadbang;
> #N comlet BANGMOV-4;
> #P outlet 361 540 15 0;
> #N comlet BANGMOV-3;
> #P outlet 271 540 15 0;
> #N comlet BANGMOV-2;
> #P outlet 180 540 15 0;
> #N comlet BANGMOV-1;
> #P outlet 87 540 15 0;
> #P button 137 176 15 0;
> #P message 137 213 14 196617 0;
> #P number 258 183 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
> #P newex 361 405 46 196617 select 3;
> #P newex 271 405 46 196617 select 2;
> #P newex 180 405 46 196617 select 1;
> #P number 363 183 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
> #P number 306 183 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
> #P button 361 443 71 12;
> #P button 271 443 71 9;
> #P button 180 443 71 13;
> #P button 87 443 71 6;
> #P number 417 331 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
> #P number 327 331 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
> #P number 236 331 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
> #P number 143 331 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
> #N counter;
> #X flags 0 0;
> #P newobj 361 356 66 196617 counter;
> #N counter;
> #X flags 0 0;
> #P newobj 271 356 66 196617 counter;
> #N counter;
> #X flags 0 0;
> #P newobj 180 356 66 196617 counter;
> #P toggle 236 183 15 0;
> #P newex 87 405 46 196617 select 0;
> #N counter;
> #X flags 0 0;
> #P newobj 87 356 66 196617 counter;
> #P newex 236 213 53 196617 tempo 80;
> #P comment 44 177 80 196617 COUNTER RESET;
> #P window linecount 5;
> #P comment 575 282 100 196617
> 16th chunks. In this case 7 , 9 , 11 , and 13 whole beats.;
> #P window linecount 7;
> #P comment 501 408 151 196617 The select objects offset the timing  
> so that each video is triggered one sixteenth apart in the counters  
> loop length. This is done to reproduce the behavior of the  
> generative nature of the original composition;
> #P window linecount 6;
> #P comment 295 211 100 196617 Timing needs to be tempo based as the  
> videos it triggers are recordings of actual performances done to a  
> click;
> #P window linecount 2;
> #P comment 93 582 281 196617 These out-lets bang out to the four  
> videos triggering the start of each;
> #P window setfont "Sans Serif" 12.;
> #P window linecount 4;
> #P comment 11 40 393 196620 This patch provides the timing I need  
> to trigger my videos. It creates poly-rhythmic events that  
> represent the different bar lengths from the original compositions.  
> In this example the poly rhythms roll over each other like this 7/4  
> 9/4 11/4 and 13/4;
> #P fasten 6 0 7 0 241 303 92 303;
> #P connect 7 0 8 0;
> #P connect 8 0 17 0;
> #P connect 17 0 29 0;
> #P fasten 27 0 7 2 142 287 120 287;
> #P fasten 44 0 8 1 512 396 128 396;
> #P connect 41 0 28 0;
> #P connect 28 0 27 0;
> #P fasten 34 0 13 0 456 308 148 308;
> #P connect 13 0 7 4;
> #P fasten 6 0 10 0 241 303 185 303;
> #P connect 10 0 23 0;
> #P connect 23 0 18 0;
> #P connect 18 0 30 0;
> #P fasten 27 0 10 2 142 320 213 320;
> #P fasten 45 0 23 1 545 396 221 396;
> #P connect 42 0 9 0;
> #P connect 9 0 6 0;
> #P fasten 35 0 14 0 483 308 241 308;
> #P connect 14 0 10 4;
> #P fasten 26 0 6 1 263 203 255 203;
> #P fasten 43 0 6 1 193 180 255 180;
> #P fasten 33 0 38 0 414 140 263 140;
> #P connect 38 0 26 0;
> #P fasten 21 0 6 2 311 203 269 203;
> #P fasten 6 0 11 0 241 303 276 303;
> #P connect 11 0 24 0;
> #P connect 24 0 19 0;
> #P connect 19 0 31 0;
> #P fasten 22 0 6 3 368 203 283 203;
> #P fasten 27 0 11 2 142 320 304 320;
> #P fasten 33 0 39 0 414 140 311 140;
> #P connect 39 0 21 0;
> #P fasten 46 0 24 1 581 396 312 396;
> #P fasten 36 0 15 0 510 308 332 308;
> #P connect 15 0 11 4;
> #P fasten 6 0 12 0 241 303 366 303;
> #P connect 12 0 25 0;
> #P connect 25 0 20 0;
> #P connect 20 0 32 0;
> #P fasten 33 0 40 0 414 140 368 140;
> #P connect 40 0 22 0;
> #P fasten 27 0 12 2 142 320 394 320;
> #P fasten 47 0 25 1 613 396 402 396;
> #P fasten 37 0 16 0 539 308 422 308;
> #P connect 16 0 12 4;
> #P fasten 33 0 34 0 414 270 456 270;
> #P fasten 33 0 35 0 414 270 483 270;
> #P fasten 33 0 36 0 414 270 510 270;
> #P fasten 33 0 37 0 414 211 539 211;
> #P pop;
>
> --
> Some links to my stuff n that :-) http://zappa.tvu.ac.uk/~crohumaa/ 
> CV/cv.php


actually, Carl, I can send you my patch directly if you like, then 

actually, Carl, I can send you my patch directly if you like, then  
you can see what I did...
let me know?

Here is a version of my patch with the video stuff plumbed in. Spacebar for fullscreen and x for the counter reset. The counter reset is so the videos can be brought back into line after they have loaded. The patch dose settle down and play ok after a couple of loops but dosent feel totally solid.

to test this correctly you will need four separate movies. the movies soundtracks need to have a musical event on each downbeat @ 80BPM, anything will do. movie 1 is 7 beats long 2 is 9 beats 3 is 11 beats and 4 is 13 beats long. Load them in backwards i.e. movie 4 first. What you should get musically is events playing in 16ths (perfectly) as a combination of all the videos. Events should play back sequentially not simultaneously

I hope this makes sense. This is my first video patch!

Yes Julio please send me your counter metro example, I wonder if it can work in my context.

Here is a version of my patch with the video stuff plumbed in. Spacebar for fullscreen and x for the counter reset. The counter reset is  so the videos can be brought back into line after they have loaded. The patch dose settle down and play ok after a couple of loops but dosent feel totally solid. 

I hope this makes sense. This is my first video patch!
Yes Julio please send me your counter metro example, I wonder if it can work in my context.

http://www.youtube.com/watch?v=fcqF9hDNoJk&mode=related&search=

Unfortunately, none of this shortcuts works on my computer

does someone have try this with success on mac?

Unfortunately, none of this shortcuts works on my computer
Powerbook 1.67Ghz, OSX.4.8, Max 4.6.2

this is not the thread you are looking for m8

Re: Accuracy triggering videos? :(
hijacked!


or are you saying that I can get better Accuracy triggering videos with these shortcuts? ;>