The importance of randomness / different timings

[syzygy]

Has anyone else noticed that visuals wit h random element or at least with lots of differenent things going on in them can give the illusion that they are matched to the beat much better than visuals that follow a regular rhythm?

In our experiments, we found that loops where things happen irregularly seem to give people the impression that they are matched to the beat even if they are not. Loops where things happen every regularly need to be more carefully matched to the beats of the music in order for them to look great.

For example, if you make a loop where coloured circles flash up onto the screen irregularly and play it alongside some music, the circles often seem to match with beats. If you make the timing of the circles more regular, it seems more obvious that the visuals and the music are not in synch.

My theory is based on the fact that the human brain is adept at spotting correspondances between things. Things that match are much more noticeable than things that don't match. To go back to the circles example i sued earlier, I think our brains notice the circles that coincide with beats much more than the ones that don't.

I think making the visuals regular breaks the illusion because the brain then has a regular pattern to follow in the visuals as well as in the music and it can tell immediately that they are out of synch.

Has anyone else played around with this?

SyZyGy

[ristuuk]

its the old scenario of turning the sound down on the tv and putting the record player on.

[rhythmicimaging]

I find image bursts work well especially as i usually perform over hardcore/hardhouse sets. as all the image bursts still include frames of movement it gives the illusion (probably because its too fast for the brain too work out whats going on) that it matches the tunes beats as they are so fast or hard.

Thats also when i use programs like midivid i attach single bitmaps to individual keys so you can show them in time to the beats. As long as you are near the sound system and dont have a wall between you and the main room as i did, as this gives you the problem of a slight delay and your images not appearing to be beat matched.

[Rovastar]

*sigh* what a backwards attitude to visuals. Surely playing random stuff does not help the visuals/VJing cause.

The best way is to learn/use the beat/sound movement in more intellegent ways to trick the viewer more into intellegent music reactions (automatic or otherwise)

Sorry folks but I do beleive that viewpoints like 'it is alright to play anything as long as it is quick enough as every so often something goes in time' is negative and defeatist. How are you going to advance with this attitude. :confused: :confused: :confused:

[rhythmicimaging]

why not use simple techniques?

randomness can equal uniqueness

variety is the spice of life

[elbows]

Rovastar Im not quite sure that was the point, rather that peeps were just looking at the way the human mind works with these things, well I found it interesting anyway. I dont think it was meant to mean "hey lets just sit back and project anything with no thought or sync"

Made me think of Cornelius because his videos to last album seemed like every single part of the image was very tightly synced, but maybe it was just illusion - I'll have to buy his DVD and analyse it :D

[syzygy]

Originally posted by Rovastar
*sigh* what a backwards attitude to visuals. Surely playing random stuff does not help the visuals/VJing cause.


I didn't start this thread to claim that we can get away with sticking some randomly timed visuals on and leave them to it. If I gave that impression then I didn't express myself clearly enough.

When we perform, we put a lot of effort in. There are two of us triggering loops using Midivid and a midi keyboard and adding effects using VCDs, an MX1 and a WJ-AVE5 in combination. Anyone that see's us in full flow would agree that we certainly don't take a lazy approach to visuals.

I posted this thread to discuss an effect that I find interesting - the fact that loops with lots going on or with random timing intervals give an illusion of synching to music more than clip with less going on or very regular timing intervals.

I wasn't suggesting that this is a replacement for tapping out rythms with a midi keyboard or even using sound reactive technology to synch stuff to music - just that there is an interesting effect there that can be harnessed IN ADDITION to whatever other techniques are being used.

Have I expressed it better this time?

SyZyGy

[brain]

hi syzygy,

yes i think you`re right concerning the observation. one should know the mechanics of perception to cleverly use them. even i'm sometimes pleased and amazed what coincidences may happen :)

but i've seen people rely too much on what you've described - which may become boring or annoying very soon. even stoned clubbers might figure out the "random flicker technique" after a while :)

people relying heaviliy on this mostly run into problems when speed or atmosphere in a track changes or when there's a dj change. suddenly what seemed so cool looks cheap or not adequate (=bad).

and there's one more point: too much short, cut loops tend to stress the viewer (at least me :) ), so he doesn't look any longer. sometimes i start to feel like these poor animals in a zoo that turn crazy/agessive because of dumb repetition of sight and movement over and over. maybe to much loops even add to the retarding effect of some drugs used :eek:

but anyway, as i said you're right. i was just addin some thougts...

brain

[eXhale]

I think it makes sense that randomness (=chaos) is able to best match the music in the punter's head, our brain always tend to create patterns and it's much easier to create patterns in synch with the music out of a chaotic structure (that's my guess anyway). It could mean we're useless but luckily there is more to VJing than beat-matching :)

[syzygy]

I'd agree with what brain says. Overusing this effect is definitely a bad idea. Leaving the same loop running through a change in the music would definitely be a no-no for us (the promotors are paying us to do the visuals in keeping with the music, after all :) )

What I'm talking about is much shorter periods of time, maybe 16, 4 or even just one bar of music. It seems that the human brain can pick out correspondences between music and visuals even in this short time.

Something else I've just thought of, the nights we work at are Drum and bass, Hip-hop, asian-beats, funky breaks, more drum and bass, and a mixture of all of the above. Perhaps the effect I am talking about is most suited to breakbeat music...

SyZyGy

[Rovastar]

Sorry about that maybe I did misunderstand a little.

I thought you were saying that random stuff is as good as getting the beat as the audience don't notice. Thus rendering what I do redundant that is why I was a little...hostile.

But still any signifcant randomness (things appearing at random time intervals) of obvious images doesn't help in my opinion - for the future devolpment of visuals. e.g. educate the public that all/most we do is sound reactive (manual or automatic)
Randomness confuses the public so they do not know what is what.

[syzygy]

Originally posted by Rovastar

I thought you were saying that random stuff is as good as getting the beat as the audience don't notice. Thus rendering what I do redundant that is why I was a little...hostile.

[/B]

Ahh - I see why you are anti the idea now. Rest assured that I don't think what you are doing is in any way redundant (In fact, I can't wait to see your stuff - will you be showing it off at AVIT?)

I'm still struggling to fully express what I'm talking about here. I'm not suggesting randomness to the point that the audience thinks "hey - that's random". In that case, willing suspension of disbelief would be broken and people would be thinking about how the visuals were put together rather than experiencing them fully.

What I am talking about is using irregularity on certain elements in the mix rather than regularity.

For example, we use a midi-controlled computer to trigger and scratch clips to the beat. We do this over other layers provided by VCDs. What I'm saying is that the stuff on VCD tends to work best if it has irregular timings rather than regular timings. The irrgularity in the VCDs is in no way a replacement for the live clip triggering that we do, but using irregular timings in other layers seems to make them work better than ones with regular timings.

If irregularity ever confused anyone, it would be being poorly used. Having said that, I think regularity can be even more confusing if it is not properly beat matched - the eye is then seeing one pattern while the ear hears another - a recipe for problems (unless confusion is the effect you are after of course)

I think I've worked out where I went wrong in my first post now. I shouldn't be talking about randomness, as that indicates a disregard for how the visuals are timed. The word I should have used all along was irregularity. Carefully employed irregularity.

Another thing springs to mind, a lot of music relies on synchopation (sp?) , where beats are offset from the regular beat of the time signature. The music is more effective in setting up a 'groove' because it is irregular (but not random ;) ) I wonder if similar ideas could apply to visuals...

SyZyGy

[fluchtpunkt]

Originally posted by syzygy

The music is more effective in setting up a 'groove' because it is irregular (but not random ;) )

...analysis has shown, that most famous musicians introduce such iregularities into their play. so there seems to be something about it :)

Originally posted by syzygy

Something else I've just thought of, the nights we work at are Drum and bass, Hip-hop, asian-beats, funky breaks, more drum and bass, and a mixture of all of the above. Perhaps the effect I am talking about is most suited to breakbeat music...


i know what you mean. i record most of my mixes, but always without audio. after the party i always review the tapes with some source of music. i've noticed that mixes from techhouse parties for example never look as good as i remembered them. they don't look bad - they just won't 'fit' with music. mixes from dnb/breaks parties on the other hand work together perfectly with new audio & seem to miracuously be synced (& i often make use of this 'human perception glitch' by (re)remixing. often the bpm stays the same all night, so when remixing you'll always stay synced to the source - giving the illusion that the (re)remixed material too is live).

[wellREDman]

on the Cornelius front, i wouldn't be at all surprised if everything on the dvd was perfectly synched, a freind of mine was a stage hand at glasto, when they played and he told me that the drummer(and thus the whole band) was playing to a click track, supplied and controlled by their video guy

[elbows]

Nice one red, youve solved the mystery :D

Heres the vid of theirs that seemed tightest to me, I really like it:

http://www.matadorrecords.com/rams/cornelius/i_hate_hate.ram

[syzygy]

Originally posted by *****
Syzygy.

im not sure if its what your refering to but theres a clip technique call MOCK MATCHING.

G

Thanks ***** - that's exactly the sort of thing I'm talking about.

Now I know the right word for the experiments we've been doing :)

Are the 80-120- 103- 122-97 bpm values specific ones that you know to work, or just examples of what you are talking about?

(i.e. is slowest - fast - medium - fastest - slow an ordering that has been found to be particularly useful)

Any idea if there is an underlying theory behind the timings to use or whether trial and error is the best approach?

We've been doing exactly what you mention with premier - slicing a clip up and adjusting the timinings of different sections of it.

Thanks again for sharing this info ***** - it's good to know that we're barking up the right tree with this :)

SyZyGy

[unjulation]

i'm teling you my computer is starting to get better at mixing then i am