excerpts from the sound correspondence

     Subject: new ideas for sound
     Date: Tue, 18 Jul 2000 12:58:53 +1000
     From: "Rainer Linz"
     To: Gordana Novakovic ,
     cc: Zoran Milkovic

     Hi again
     I have been going back over some of my sound ideas for infonoise -
     some were from 18 months ago I now realise - and while I am still
     interested in some aspects I think it might be better to
     reconsider everything in light of what has happened since.

     In view of the new technical possibilities with java, I need
     feedback about whether the ideas are useful or whether there is a
     better way to think about them.

     Zoran I'm not sure if I ever sent you my original sound proposal.
     When I first made it, I did not know that the old java had
     such poor support for sound. In short, my ideas would have been
     difficult without using a native environment, but now with recent
     advances in java, at least there is enough flexibility to create
     something more interesting. This of course means using the new jmf
     v2.0 which I understand only in a general sense.

     Firstly I am very keen to use good quality sound samples, 16 bit
     at 44.1kHz. This is always my first question, since anything else
     produces a scratchy sound. It can be an advantage sometimes, but
     I feel strongly the sound should be made with good quality samples.

     My other, major question is how to make the sound work with the
     rest of the processes in the installation. For instance, since the
     projected texts will be moving around the moebius strip, I would
     very much like to have the sound moving around also, though not
     necessarily in the same way. If there IS some way to move the
     sound towards the visitor - in the same way that the projected
     texts will move - then I would be very keen to do it. (any
     suggestions?)

     And there are interdependencies. I understand that the motion
     sensors will control the movements of the texts around the strip.
     The sound control will also use the same sensors... So it would
     help me to understand how the projections will detect the visitor.

     I am very keen to use sound in the physical space which will mean
     having multiple loudspeakers, and ways of controlling the panning
     position and volume through the software.

     My original idea was that each of the 6 computers would have their
     left loudspeaker positioned in or under the moebius strip, and the
     right loudspeakers would be quite far away from it, maybe mounted
     on a wall. In this way, panning a (mono) sound left and right
     would move it closer or farther away.

     I still like this idea but am worried now about the speakers
     themselves. I thought it could work with powered, multimedia-style
     computer loudspeakers. But since I worked with them on a garden
     installation piece last year, I am less impressed with their
     capability.

     Now I am thinking that four of the computers could have their
     loudspeakers mounted inside the strip as before ( = 8 speakers).
     The other two computers would be connected to four larger speakers
     in the corners of the installation space. What we would need to be
     able to do here is route a mono source to any of the 12
     loudspeakers.

     I don't know how feasible this is, however, since it would rely on
     network processes. It would mean implementing a kind of sound
     mixer that could route a variety of sounds to different
     locations...

     Anyway, I would like to get the process moving with these general
     points, I will write more in the next few days but hopefully you
     will have a chance to think about it and maybe give some feedback.

     best wishes
     Rainer

     
Subject: Re: new ideas for sound Date: Wed, 19 Jul 2000 13:06:34 GMT From: "Zoran Milkovic" To: "Rainer Linz" Hi Rainer, I have carefully examined your last mail, and I will try now to explain you how I think the whole thing should work. I do not know yet how it will be done in detail, but the principle and the organization - to speak only for the software - is rather clear to me. First, the installation will be made on a Local Area Network. It means that we will have all the bandwidth we need to achieve a high data transfer rate between the computers in the installation, whatever their number is. That means we can use high quality audio samples, and high resolution images - without the limitations of the web project. Second, I plan to organize the software on these computers as "data processors", i.e. they will process the information available from the LAN and produce adequate output by rendering the input data. For example, let's imagine a headline and a sound, and that they are linked. This "package" has some attributes such as position, level, zoom, timing, ... and whatever. This "package" (the right word would be "object") is in fact a file on the LAN server, containing all the data and references. These data change in time (several times per second), and these changes are driven by the sensors: the software on the server scans the sensors, detects changes, tries to establish dependencies (motion estimation), and generates or changes the content of the "packages" (objects), updating them. Now, the other computers on the network only cover a segment of the "space" of the installation. They keep examinig the "packages" (reading the files), and when they detect that they are "covering" the space the "package" enters, they start "rendering" that object, taking account of its attributes. Timing is of great importance, because sound (and animation) is a temporal process and error in synchronization will produce inacceptable results. Imagine playing the same sample on two loudspeakers at different level to produce a panning effect, and one of your loudspeakers starts playing the sound with a delay. You start hearing two sounds instead of one. And in our installation the sound could have started in a distant segment, so we need a precise time reference to be able to produce the right sample with the right timing. This is just to explain why we need to define all the attributes we will ever need. Now let's get back to your letter. Producing a stereo sound is quite clear to me: you take a (mono) sample and send it to two speakers at different levels, so anyone listening to it feels that the sound source is somewhere between the two speakers. But this is only true if the listener and the speakers make a triangle. If you are in line with the speakers, I think you will have no sense of the position of the sound (I may be wrong). Anyway, with the concept of "objects" I tried to explain, this will be of no concern. The space of the installation (as far as the sound is concerned) need not to be limited to the moebius strip, it can easily be expanded to the whole room, by adding loudspeakers at some distance away from the strip. We are not limited to 6 computers - that figure was given just to fill a blank. We will use the number of computers we feel will produce adequate sound for the installation. The "video" part is more limiting: a larger number of video projectors would be too expensive. Sound production is much cheaper (don't get me wrong - I am talking of hardware!). I met Gordana a couple of days ago, and I will meet her again this afternoon. She will then see this letter - so if she has some remarks, I will write it to you in my next mail. Regards, Zoran
Subject: Re: new ideas for sound Date: Thu, 20 Jul 2000 09:21:07 +1000 From: "Rainer Linz" To: Zoran Milkovic CC: Gordana Novakovic Hi Zoran thanks for your thoughts. Yes I thought we would be working with a LAN, it seems the sensible thing under these circumstances, with regard to data rates, etc. Only I don't know much about LANs and what limitations they would have under our plan. > Second, I plan to organize the software on these computers > as "data processors", i.e. they will process the information > available from the LAN and produce adequate output by > rendering the input data. This also makes sense, I guess it means that the audio files (or samples) will be loaded from the hard disk. DO you see any problems with latency in this regard? What I have read about the jmf soundplayers indicates a lot of time is needed (comparatively) to prefetch, realize etc the players. This means there could be a long latency between the time that the sensor is activated, and when the sound starts playing. > This "package" (the right word would be "object") is in fact > a file on the LAN server, containing all the data and references. I guess the specifics of each new object are created in real time in response to the sensor triggers, and not just called up from an array of ready made objects stored somewhere. > you take a (mono) sample and send it to two speakers at different > levels, so anyone listening to it feels that the sound source is > somewhere between the two speakers. But this is only true if the > listener and the speakers make a triangle. yes that's right, but I think we have a problem using the term "stereo". When I wrote about using the sound spatially, I didn't mean to use the speakers as stereo pairs, but rather as single point sources from which the sound can originate. So a mono sound panned totally left comes out of the left speaker. As we pan it towards the right it's not really a stereo sound, but simply the it starts to come from a new physical source (R speaker), which is positioned somewhere else. What happens in between this L-R travel is not really stereo, but some strange kind of movement effect that will depend on the positions of the speakers and the listener as you say. I'm glad you raised the possibility of playing a single sound on two computers at the same time. This would enable the panning effect but rather than thinking of it as a stereo condition, do you think we might for instance, move a sound between six loudspeakers, one after the other? This would give the kind of movement flexibility that I tried to suggest. What kind of parameters and variables do you think will be held in the packages or objects that you have mentioned? I am trying to think how closely we should link the sound and the headline images, and to what extent they could still work independently, ie in different combinations. best wishes Rainer
Subject: Re: new ideas for sound Date: Thu, 20 Jul 2000 18:10:35 +0200 From: "Darko Novakovic" To: "rainer linz" hi Rainer, thanks for keeping me informed. i do enjoy your exchange. still have nothing to add. met zoran yesterday, and have some clearer idea. still, i cannot imagine the sound. hope to meet Zoran soon again, and make some contribution myself. best, Gordana
Subject: Re: new ideas for sound Date: Fri, 21 Jul 2000 08:52:52 +1000 From: rainer linz To: Darko Novakovic Hi gordana yes, we are still working out some technical groundplans, I am trying to understand how the sound could work with the image, etc, and what programming limitations there may be and so on. I want to suggest to Zoran that we could download audio material to use for the piece. Like every night we could download more audio material from certain sites and use that in the following day. if you look at http://abc.net.au/am/ this is a news service with audio, for instance. anyway, just a thought. best Rainer
Subject: Re: new ideas for sound Date: Mon, 31 Jul 2000 10:40:29 +1000 From: "Rainer Linz" To: Zoran Milkovic Hi zoran so I am still thinking about the best way to render the sounds for the installation, and have been thinking in general about how best to move the sounds around, that is, whether this is a viable option. I need you to tell me whether any ideas are impractical, or if there are better ways to do things, or if we need another approach... At this stage I think there are still a number of options for how the sounds can move. In my last email i tried to imagine a ring of loudspeakers with a single sound panning through all of them. This, I now suppose, will mean synchronising the start of a sound on all the channels, and adjusting the volume or pan to give the impression of continuous movement. I can't say if this is difficult or not. The main thing is, there are still other options if it can't be done. The key would be to arrange the speakers in the gallery space in such a way as to give the impression of movement. Take for instance, a row of four loudspeakers, arranged 1L, 2L, 1R, 2R. Computer 1 can pan from 1R to 1L, computer 2 can pan from 2L to 2R. A person standing in front of speaker 1R could hear a sound moving from 2R to 2L. Technically the panning should be quite simple, but the impression of movement will be more complex, since a sound panned in the centre of 2L & 2R would appear to come from the same position as 1R. There are many ways we can work with the space in this manner. It's just a question of what is feasible. have you had any more thoughts> cheers Rainer
Subject: Re: new ideas for sound Date: Mon, 31 Jul 2000 13:14:51 GMT From: "Zoran Milkovic" To: "Rainer Linz" Hi Rainer, I have been spending a lot of time with Gordana, and there are a lot of things we agreed upon, and should talk to you about. No real changes to the concept, but just clarifications. But we'll get to it later. Now I will answer the questions you are asking me again and again - sorry for not answering them before. > What is still unclear to me from your outline, is whether > the sound files will exist on the LAN server, or whether > they will be stored on the local computers' hard drives. > What is the speed of data transfer in a LAN? You should look to the LAN as just another hard disk on your computer. Yes, it is a little slower than newer hard disks, but you can acheive a transfer rate of several hundred kilobytes per second, so it isn't so slow after all... If you have at least two computers, you can try to connect them, all you need is two network cards - an ISA card costs here 20DM (Aus$). Win311 (WFW) and W9x have all the support you need. I can explain you the procedure to make it work. Interested? > In my last email i tried to imagine a ring of loudspeakers > with a single sound panning through all of them. This, I now > suppose, will mean synchronising the start of a sound on all > the channels, and adjusting the volume or pan to give the > impression of continuous movement. I can't say if this is > difficult or not. No problem at all. JMF has mechanisms for absolute synchonization of sounds, including all the preparations to eliminate latency. There is also Java3D, with a lot of interesting features... > Take for instance, a row of four loudspeakers, arranged > 1L, 2L, 1R, 2R. Computer 1 can pan from 1R to 1L, computer > 2 can pan from 2L to 2R. A person standing in front of > speaker 1R could hear a sound moving from 2R to 2L. > There are many ways we can work with the space in this > manner. It's just a question of what is feasible. Anything can be done. Still, the most simple thing is to play the same sound(s) from several computers, and to change volume levels on all the channels to obtain the movement in space we want to acheive. I will write you soon, with more info. Best, Zoran
Subject: Re: new ideas for sound Date: Tue, 01 Aug 2000 08:35:58 +1000 From: rainer linz To: Zoran Milkovic Thanks Zoran I will look into the LAN cards, i know they are quite inexpensive these days... Anyway, I am very keen to hear about your discussions with Gordana. Please keep me informed. I have also had some further ideas about the sound, namely the kinds of sound samples I will prepare to use. In short, these are MIDI overflow sounds created on a synthesiser by connecting the MIDI THRU to the MIDI IN. This creates a feedback loop which goes into overload, creating some fairly complex and interesting (I think) sounds, that would also be appropriate for this piece. I will also send more details tomorrow. best wishes Rainer
Subject: Gordana is back in London Date: Thu, 10 Aug 2000 11:53:54 GMT From: "Zoran Milkovic" To: "Rainer Linz" CC: gordana.novakovic@virgin.net Hi Rainer, Now that Gordana has gone back to England, I'm returning to the e- form of communication. I must say that eye to eye contact is better. So, if everything goes well, we might see each other in Paris in December. But there may be some obstacles, we'll see... Now, about the things you wrote in your mail: > Anyway, I am very keen to hear about your discussions with > Gordana. Please keep me informed. I'm sure that Gordana will write you as soon as she gets back to her computer in London. She didn't open her mail while she was here. So, I suppose she'll have a lot of it. We have been discussing the whole project, and we agreed that the emphasis should not be on the news, but rather on the invariance of the form of news, so that when you're looking at the news, you don't notice that you may be looking at some rather old news. It's always the same. (I am oversimplifying the whole thing). > In short, these are MIDI overflow sounds created on a > synthesiser by connecting the MIDI THRU to the MIDI IN. I'm not sure that MIDI overloading gives the same results on diferent MIDI equipment. I'm not an expert on that field, I have no experience of my own, but for as much as I know, MIDI equipment tend to add some information if the MIDI flow, that may create loss of some of the original message. But this is a random process, and I suppose it gives quite unpredictable effects. If there is any chance to do it, I would like you to send me a sample of this kind of sound. Make a .WAV file, and then convert it to .MP3 (if you don't have the tools, I'll see what I can do). One megabyte of MP3 may be a whole minute of sound, with low compression loss of quality. In any case, please keep it well below 1MBy, otherwise I will not be able to get it - there is a limit in my account. best wishes, Zoran ps: I will translate to english some of the texts I gave to Gordana, and send it to you soon.
Subject: Re: Gordana is back in London Date: Fri, 11 Aug 2000 08:56:15 +1000 From: "Rainer Linz" To: Zoran Milkovic HI Zoran thanks for your mail. Yes, I am looking forward to solving some of the basic questions so that I can concentrate on the detail and structure of the sound. What I wrote about the midi overflow method of creating the sounds, I should have made it clear that these sounds are created in the studio on midi equipment, and then recorded into WAV format to be played as ordinary samples. So the method of creating these sounds is very "hands on", I will choose the ones I like and then prepare them for use as samples. I had thought of sending an MP3 version of something for you, I have to get hold of an MP3 encoder (I have some shareware progs here, but have not tried them yet, not all of them record at 44.1 kHz). Anyway I will look into this and see what I can do. And I hope we will be able to meet in Paris later this year, I will find out more today about how likely it is. As you may know, S is discussing the possibility of a performance, if it happens then I will be invited to assist, if not then I won't be there... I have also been reading up on the jmf. I have the 1.2 version here and have compiled some of the example programs. a video player applet and so on. cheers Rainer
Subject: sounds Date: Mon, 14 Aug 2000 19:41:48 +1000 From: "Rainer Linz" To: Zoran Milkovic CC: Gordana Novakovic Hi here are a couple of the midi overflow sounds I wrote about earlier. Sorry to send them as 44.1kHz samples, I have not yet been able to get an MP3 encoder. Do you have any suggestions about the best one to use? With these short samples, if you loop-play them you will get an idea of what the continuous sound is like. With the longer samples however the rhythms are more subtle, and change a little over time. The sounds themselves are quite complex and there is a strong repetitive component which has to do with the length of the synthesiser's buffer, and other random things can happen over time as well. Hmmm... there are many different types of sounds that can be drawn from the synthesiser in this way, but they all have similar behaviours. cheers Rainer
Subject: Re: back Date: Tue, 29 Aug 2000 09:47:50 +0100 From: "Gordana Novakovic" To: "Rainer Linz" hi Rainer, glad to hear you had a good time (and are back). i like very much your sounds by themselves, and the concept. completely "in tune" with the work in all, i am very happy with it. i was thinking of some layer of "atmospheric" sound, if you know what i mean. not maybe human voices (or maybe yes, not sure) - something of the kind the orchestra produce before performance with authentic murmur of audience, while getting in tune, you know what i mean, sorry for clumsy explanation. (just small idea). Zoran sent me some txts, but i still cannot concentrate. best, gordana
Subject: Re: back Date: Wed, 30 Aug 2000 09:37:29 +1000 From: "Rainer Linz" To: Gordana Novakovic CC: Zoran Milkovic Hi Gordana I'm glad you liked the sounds I sent, as long as you know they are only very short versions of longer sound samples... My idea is to play a number of them at the same time, to make an even more complex sound texture, that is also moving in detail around the space. In a way, the sounds remind me also of signals, something like morse code or satellite transmissions, sounds that appear to have some meaning or intelligence but can't be decoded. I had also thought of using voices, like I wrote to you some time ago, but I don't think you were very interested in using voices at that time. To me, voices can be used with the headline texts that will be projected onto the moebius. This does not mean that the voices must speak the same headline texts which are projected, but rather that many short, cryptic texts in the manner of headlines can be used to present a different meaning. Anyway, you could imagine a room full of people, all of them whispering, to give a sense of this kind of sound. If the whispers are also moving around the space, it will have the effect of a conversation that you can only hear a few words from... In this way, the sounds like I have already presented, and the voices/whispers, would complete the installation sound. The mail from mc5 is also very encouraging, in my view. It is another example of how word spreads around, and also that there are people with a genuine interest in the piece. In a way, I can accept that people have problems with budgets, maybe they do or don't have the money to install such a work, what I try to read between the lines is a genuine interest and a desire to do something with it. I don't know anything about mc5, but at least they ask the right question first (ie, the budget!!!). what do you think? rainer
Subject: Date: Wed, 30 Aug 2000 09:18:53 +0100 From: "Gordana Novakovic" To: "Rainer Linz" CC: "Zoran Milkovic" hi, thanks for this letter. it finaly gave me the idea how you see the sound in total. i know this is very demanding concept specially for you, regarding the nature of the sound compared to image. for me, it was much easier - images do not mix and cumulate in space, first of all. i was wondering all the time how you will master this and other problems. so, congratulations! it looks great to me! i did not want to expect anything in particular, and now i am happy for this. i can imagine this sound, and find it very sofisticated, but powerful at the same time, rich, with refined elements...can't wait to hear it, simply. voices - i did not dislike them, they were not on my mind, only. now, the work has evolved, sure i am happy for it, and the whole thing is much clearer to me. anyhow, i felt the need for some layer of "human" origin, to connect it to the painted images. whispers - i like that. very much. i feel so bad now that i will be away for 2 weeks, but i hope Zoran and you will keep the thing alive in the meantime. tomorrow - holiday, like it or not! mc5 - i did not know anything about myself as well, but also felt very encouraged. however, i would like it much better to get real support at some "smaller" (?) event, than to be in trouble in a "big" one. i replied {sent them textual description and cost details) and we will see. now it is for sure - it will BE somwhere. maybe we should try tm - they can not complain of small budget ) well, all my best till 14th, gordana
Subject: Re: happy holidays Date: Thu, 31 Aug 2000 12:42:20 +1000 From: "Rainer Linz" To: Gordana Novakovic Yes it's very difficult to give an exact description of the sound, since I am very interested always in HOW the sound behaves, not just in WHAT the sound is... So that a movement around the space (which has yet to be determined) constitutes a kind of behaviour.... Anyway, do you have some suggestions about how to collect texts for the whispers? I feel that we should use texts in different languages, and we should collect them from around the world not just in Australia, so it would be good if some of our friends could collect and send me a selection of newspaper headlines according to some "rule" that may be useful - for example all headlines in the first week of september, or some such. Rainer
Subject: demo Date: Fri, 01 Sep 2000 10:03:16 +1000 From: "Rainer Linz" To: Zoran Milkovic Hi again Zoran I have heard from Gordana that she is going on holidays today, some people are lucky... Anyway I have had a thought about how we might go about a presentation... I have decided in the end that I will get some networking hardware, that way I can network my computers (a 486 and a PII/450) and use this to demonstrate the movement of the sound around the space as we discussed. Do you think it's possible to make a demo program which would use two computers in this way, with 4 loudspeakers (2 x stereo soundcards) and the sensors, to demonstrate how the sound will move according to the sensor information? This would make a great demonstration, without having to network six computers as for the installation... If this could be done, I would feel much happier about acquitting the project, and justifying the spending of the grant. as always I look forward to hear from you, and how you are going etc. best wishes Rainer
Subject: Re: demo Date: Fri, 01 Sep 2000 12:09:05 GMT From: "Zoran Milkovic" To: "Rainer Linz" Hi Rainer, Sure I think it is possible to make the demo with two networked computers, with a sound card each. I will even try to make it using Java. Just in case, I will also make a Pascal program to demonstrate all the details, if I get problems with Java (accessing hardware is not straightforward). I have asked Gordana to check the text I have written about the project, but she had no time to do it. So I will send you this draft in my next mail. We had some problems with a hard disk at work, and the text is on it. I think I have a backup somewhere, and I will try to find it. I will write you on monday. best regards, zoran
Subject: Re: demo Date: Sun, 03 Sep 2000 07:54:08 +1000 From: "Rainer Linz" To: Zoran Milkovic Hi Zoran I am looking forward to reading the text you have written - please send it in whatever form you have... meanwhile, I have finally got some network cards, but of course it is not as simple as it could be. In the meantime I am looking at a PC card type network interface, (PCMCIA) for my laptop. I have decided on a 10/100 Mb system, people are suggesting I should also get a network hub to connect everything together. I have the cards but no cables yet, mainly because I was advised to get a hub, and I am still thinking about it... As I mentioned, I tried to get your TOMPG program working without success, so I have dowloaded a program called CDEX which should do the same thing. Anyway, these things are still happening, and I will send some longer sounds soon. best wishes rainer
Subject: Re: demo Date: Mon, 04 Sep 2000 09:34:43 GMT From: "Zoran Milkovic" To: "Rainer Linz" Hi Rainer, here's the text I told you about: The principle of viewer motion detection is based on statistical information processing. These informations are in fact the states of Passive InfraRed detectors (PIRs) placed under the central object (a Moebius-shaped strip). Let's consider a group of three detectors, the middle one being activated. By monitoring the state of the left and right detector one may conclude in which direction the viewer is moving. This is, of course, not always true: the neighbouring detector might be activated by another event, not necessarily by the movement of the viewer that has activated the middle one. But, statistically looking at the whole installation space, the gathered informations will MOSTLY be accurate. For this to be true, there must be a large number of PIR detectors, with a short reaction time. The software for motion analysis makes a time-based acquisition of the detectors state, and tries to establish a relationship in the order in which the detectors are activated and to adapt the "movement" of the objects (both visual and aural) in the installation space. There is no direct link between the state of the detectors and the motion of objects - the relationship is established by applying the results of the analysis to the parameters of the movement of objects, the destruction (disappearance) of old objects and the creation of new ones. This indirect correspondence between the statistical (average) motion of the viewer and the motion of objects on the scene provides an additional indetermination, thus disabling the viewer to willingly affect the events, and adds "noise" to the interaction. The installation software will be written in JAVA programming language (Java2, Standard Edition), with additional packages Java3D and JavaMedia Framework. The application will run on PCs with Windows9x operating system. Some of the routines for PIR detectors state acquisition will be written in another programming language (such as C), and linked with the main application through the Java Native Interface. The computers are networked, so all the informations necessary for the operation of the installation are located on one of them, the one in charge for the detectors state acquisition, and the rest of the computers access it through the network to find the data necessary for their operation. Best regards, Zoran
Subject: Re: demo Date: Tue, 05 Sep 2000 13:12:27 +1000 From: "Rainer Linz" To: Zoran Milkovic Hi again thanks for your mail. Your description of how the motion sensors work is quite logical. Unfortunately we cannot assume that only one person will be in the space at any time, so that any "prediction" of motion paths must be statistical in the sense that you have described. > For this to be true, there must be a > large number of PIR detectors, with a short reaction time. So, how many detectors do you think we'll need? And is there a special type of detector which has a short reaction time? I know that some PIRs can be adjusted in their "on" times, but then if a person just stands in front of them, they continuously switch on and off... Some kind of routine that would sense a person repeatedly triggering the same detector could be useful also. Although under what circumstances that might work... Is it possible that with say, 12 sensors, there would be 12 headlines and 12 sounds moving about? best wishes rainer
Subject: motion detection Date: Wed, 06 Sep 2000 10:53:46 GMT From: "Zoran Milkovic" To: "Rainer Linz" Hi Rainer, The reason why I mentionned "large number of sensors" is just to justify the statistical approach of motion detection. The "short reaction time" is to obtain lots of data to process. There are a lot of scenarios in which the motion detection process will give erroneous results, but that's the best we can do without having to put a camera and process its picture to detect everyone's motion (that would be impossible to do with today's technology!). We are not making an absolute detection system, but an installation with all its imperfections. The goal is not to demonstrate technological achievments, but to make an interesting piece of art. Meanwhile, I haven't done anything yet with the demo, we are now in the process of changing the hardware at work: I have now a PIII (500) with 128MBy, and am expecting another one in the next couple of days. Please tell me the deadlines for the demo. It will be then much easier to me to get to work. all the best, zoran
Subject: Re: motion detection Date: Thu, 07 Sep 2000 07:44:55 +1000 From: "Rainer Linz" To: Zoran Milkovic Hi Zoran I understand that motion detection is a complex problem, and I would never suggest a camera system in million years. It seems to me that even if the system makes a good guess about what is happening (ie what the visitors are doing and where they are going) the response needs to be consistent but also complex and interesting. As you say we are not demonstrating technology, we are making an artwork. How many headlines do you think can be projected at one time onto the moebius strip? Does this relate directly to the number of sensors, or the number of computers, or is the upper limit determined in some other way like bandwidth for instance? So, I could imagine say, 12 sensors but only maximum six headlines, and six independent streams of sound... Or with 12 sensors, should I be thinking about (max) 12 simultaneous streams of sound? Ultimately this is a bandwidth problem, and I am trying to get an idea of how much to expect... Just aside, I am also looking forward to using quite long samples in this piece, rather than short (1-second) samples that are continuously looped. The streaming aspect of java media is a good opportunity to use some longer sounds. Now that you have a PIII computer everything should be working much faster ha ha. best wishes Rainer
Subject: mp3 Date: Tue, 19 Sep 2000 09:25:57 +1000 From: "Rainer Linz" To: Gordana Novakovic CC: z_milkovic@hotmail.com Hi again here is a little excerpt of some of the midi sounds I have been working with... This one is almost like a folksong... I don't know if this particular sample is suitable for the installation, but again, it has the characteristics that I mentioned before - quite a complex (and subtle) rhytmic interplay with a quasi random element as well.. This longer sample shows the variety of the sound, rather than a short sample looping continuously as I sent before. cheers Rainer
Subject: Re: mp3 Date: Tue, 19 Sep 2000 21:38:02 +0100 From: "Gordana Novakovic" To: "rainer linz" hi!! thanks for mp3, i do like it. the rhythm, or those interruptions just ones you get used to it, those subtle changes, and the melody - yes, that is it. i liked "small" ones as well. it is now clear that there is a great possibility, or variety of it. i will try to think more. maybe i could send you some of my national music, if you wish? (just if you feel like, this is not even a suggestion). i am still not clear whether you want to follow the layers of the visual, or on the contrary - not, developing something else. it is all up to you, sure. but it will help me in thinking of it and making eventual comments. you might have told me. i am not sure now . i liked very much the sounds on the first piece on your cd. but, as i said, will think more. i am still "camping" in the new flat, waiting for the furniture, almost unpacked so far, but not all in normality. soon, i do hope. as for i, i am not surprised you are confused with the terms, i went a number of times to their website to catch the point. so, finally (IF!) our "infnoise" is accepted for 30 min presentation, and my paper in theory category is, thank god, rejected in last circle. i have got their info on lodging, so i will fwd it you as well. how is your work with S advancing? any news on your tarvelling to paris? in the meantime, i was thinking a bit of the way of presenting our work, but later on it. best, Gordana
Subject: Re: mp3 Date: Wed, 20 Sep 2000 20:46:27 +1000 From: "Rainer Linz" To: Gordana Novakovic Hi gordana Im glad the mp3 arrived OK... I'm not used to sending such big attachments. I also sent a copy to zoran which bounced back... I would like to work out some kind of movement scheme for the sounds in the installation, but am still undecided exactly how. The sounds should approach the visitor, but they can also disappear again more easily than the texts (which must be read), in other words a more intricate movement scheme would be possible. I am thinking that when a visitor triggers a detector, a sound will start from the opposite side of the sculpture and move around to the "front" where the visitor is standing. From there it might move around the two or three adjacent loudspeakers, then fade away again to the other side of the sculpture. I imagine it being quite dynamic in movement like this. Because the small computer-type loudspeakers have very poor quality in sound, my original idea was to have a parallel sound system with 4 bigger loudspeakers (one in each corner of the room). The work with S has reached a pause and I haven't heard any more news. Please keep sending me your developing thoughts about the presentation of the work best wishes Rainer
Subject: Re: try again Date: Thu, 21 Sep 2000 12:51:56 GMT From: "Zoran Milkovic" To: "Rainer Linz" Hi Rainer, I like the sample you sent. Never mind the longer one, this one gave me the idea of what it will be. If you ever need to send me a longer sample, you can always split it into parts and send it in pieces. Or we can share an address on the net (all we have to do is exchange passwords, and we will all three have access to some disk space (there are locations that give up to 20MBy of storage space, I can find their addresses if necessary). Best wishes, Zoran
Subject: Date: Sat, 23 Sep 2000 12:19:22 +0100 From: "Gordana Novakovic" To: "Rainer Linz" CC: "Zoran Milkovic" hi Rainer, as i have said, i would like to start from the begining, for myself mostly. the main thing to define now seems to be movement and it's nature, i think. the image/projection is divided into 3 segments: 1- dreamscape. it has very slow movement along the sculpture, almost unnoticable. only in a while you can notice that it has changed viewing from the very same point. 2- ouroboros (wire shaped snake). it's movement is the result of algorithm, so it is a bit strange, and constant. naturaly, since it represents eternity, etc. i think you can use it also for generating some level of sound. it is very simple (and old) program for geology. if you wish, Zoran can (i hope) send it to you, i have lost it in all my movings. you might use it somehow. 3- headlines. they are floating in some random movement, and "confusing" using the old computer "game of life". this structure might also be of some use for you. i like very much your idea of more complex way of moving the sound through the space. it will add more dynamics to the event. and yes, the advantage that sound is more, hm, flexible then the image should be taken. 4 loudspeakers - of course, it is already in our technical description, as far as i remember.
Subject: Re: try again Date: Mon, 25 Sep 2000 09:19:34 +1000 From: "Rainer Linz" To: Zoran Milkovic Hi zoran thanks for the message. I am still working on the sound samples and have been discussing some possible movement schemes with gordana. I mentioned that the sound movement could be more active than the movement of the headlines. I suggested a scheme where, for instance, when a visitor triggers a sensor, a sample will start playing on the opposite side (back) of the moebius sculpture, and move quickly to the front where the visitor is standing (? as we discussed). From there it could move around the two adjacent loudspeakers in a sort of "dance" in front of the visitor. this was just an idea. regarding the movement of the images, gordana replied > 2- ouroboros (wire shaped snake). its movement is the result of > algorithm, so it is a bit strange, and constant. naturaly, since it > represnts eternety, etc. i think you can > use it also for generating some level of sound. it is very > simple (and old) program for geology. if you wish, Zoran can (i > hope) send it to you, i have lost it in all my > movings. you > might use it somehow. > I think this is the shape that is moving around the moebius in the web applet? It makes sense to me to relate this in some way to the sound as well. As for sharing web space, I thought we had an address for this already, but have not been using it... I haven't tried any of the web hosting services like this, I understand they all have advertising etc... hope things are going well cheers Rainer
Subject: Re: Date: Mon, 25 Sep 2000 09:19:38 +1000 From: "Rainer Linz" To: Gordana Novakovic Hi gordana It seems I am still discovering things about this installation! Many thanks for sending your thoughts about the movement. What you have written about the movement of the images is familiar to me from the web-based piece, but for instance, I thought the dreamscape would be "painted" onto the moebius sculpture in the gallery, and not projected onto it. Now it is clear to me. I like the idea very much, of having a slowly-moving dreamscape projected on the strip. It makes for a very strong connection with the web piece, and will also lend a type of dynamism to the installation. The ouroborous is the wire-frame image that is moving around the strip. So this will be present in the gallery as well. Great. And as for the headlines, I wonder if their positions should be governed by the sensors trigger after all, rather than a random movement as you suggest. Its easy to imagine all of these movements (and the sound as well) happening in a quasi random way, but I feel it still needs to be sourced in some way from the sensors. the responses do not have to be strictly one-to-one. So instead of saying that the headline (or the sound) will move directly toward the sensor that has been triggered, it will rather move "around" the sensor in a slightly unpredictable way, based on the statistical processing that zoran wrote about. This way it seems to have a life o\f its own while still responding to the presence of the visitors. I'm not explaining it very well, it's also late at night here. I guess what I am trying to indicate is that the piece can have a type of behaviour, rather than being a moving but repetitive (and so static) object... So I am trying to imagine how it might behave... Yes, the four larger speakers were in the original specs, I mentioned them again to make sure they were still there... So physically there will be two louspeaker systems, one with 8 or so smaller loudspeakers around the moebius and 4 larger speakers in the corners of the gallery space. The method of playing sounds over these loudspeakers is the same - two computers will drive the larger speakers, four computers will drive the smaller ones. All are networked which means it should be possible even to move the sounds from around the strip to around the room.... Anyway there are more possibilities, important for me is to find the correct combination of possibilities so that everything can make sense to the visitor. hope all is well Rainer
Subject: headlines behaviour Date: Mon, 25 Sep 2000 10:43:01 +0100 From: "Gordana Novakovic" To: "rainer linz" CC: "Zoran Milkovic" Hi Rainer, well, sometimes i am dying for 2 hours of "live exchange" with you. we will avoid situations, more bad explanations of mine :( that i asume you know, but...) anyway, lets go on. i am very happy with all your comments. in spite of all, seems we are in the same "tune". i am glad that your suggestions are exactly as it is supposed to be, if i did not misunderstand you. about behaviour of headlines. this segment of projection (whole "image" will be also "illusion" - projected changable combination of 3 levels simultaneously, digested by networked PCs) will actually have its autonomous behaviour according to the rules of the software (old computer game "of life"). it means that they will have one way of "playing" among themselves (moving allong in some sort of game, not at all linear stream). the sensors, while triggered, will interrupt the movement in its area and the "play", movements in all directions by the rules of software, will turn into chaotic mess, always different, making always new combimnation of headlines, thus creating noise. so, all in all, it will be vagely readable. it will be more an idea of headlines, or flashes. alike having a glance at tv sets in the shop window while walking the street, i suppose (who knows how it will REALLY look like?). that is why i think we can make prints of all headlines in random line for visitors to take on the exit, perhaps) please let us continue this "vivisection" of the work. it will help me to clarify all. and i think it gives you firmer frame for your work. i will send this to Zoran as well, so that he can correct me on "things" - hm, confusion in belgrade, this morning my home (container) is supposed to arrive at this place...worried and happy, waiting, waiting.... and you? best, Gordana
Subject: Re: headlines behaviour Date: Wed, 27 Sep 2000 10:03:32 +1000 From: "Rainer Linz" To: Gordana Novakovic HI gordana thanks for your reply. Yes, the autonomous behaviour of the headlines, interrupted by the arrival of visitors, is very compatible with the idea for sound. This type of behaviour can also be used by the sound, in terms of movement in space. In a basic sense the same parameters would control the sound and the images, which is the kind of connection I am always pleased to see. Some direct comparisons would be: - selection of which headlines to project - > selection of which sounds to play - number of headlines projected at any time = (minimum to maximum) - > number of sounds playing at any time = (min to max number) - weighted random ( = "life" rules) movements in space - > similar movements of sound in space. - autonomous headline behaviour interrupted by visitors - > autonomous sound behaviour interrupted by visitors. there are other similarities, that is to say direct comparisons, that can be made between the sound and image behaviour. Maybe we could draw up a list of those parameters that can be shared, and in this way devise an overall strategy for how to manage these "resources". Also I think it is important to be able to (automatically) turn down the volume of the sound, when no visitors are present. This is first of all to save the gallery staff in the space from having to hear it every hour of every day (eventually they will turn it down - or off - themselves). I have been following the news of the elections in belgrade, it sounds like more of the same..? Maybe you were lucky to get your belonging out, at this time... Anyway I hope everything arrived OK, There is something cleansing about moving all of one's belongings to another place but I don't think I could do it! best wishes Rainer
Subject: snake Date: Fri, 29 Sep 2000 14:35:05 +0100 From: "Gordana Novakovic" To: "Rainer Linz" hi again. i was thinking of starting with the "snake". yes, you are right. this the sample that Zoran sent you long ago (wire). for me, this is also very important part. it is constant, out of interaction, autonomous. it has it's own independent "life". i would be very happy if you felt like in a way detecting, or depicting the sound from snakes's schema. for me, with all those atributes of ouroboros, it represents rational, cerebral, let's say - mathematical way of thinking. it might be, somehow a bit transformed, or not, the sound for installation in idle mood, perhaps. what do you say? i got a message from Zoran, saying that he is following the exchange with "comprehension and without remarks, so far" ;) cheers, gordana

 

© 2002 infonoise