Author: Uinseann Mould

I did not solve the problems I wanted to solve. OSC would not work with my version of PD, it did not work with any other version I used, I tried reinstalling, troubleshooting and even tried to do this on my sister’s MacBook, to no avail.

Even if PD had recognised the OSC objects, I couldn’t even connect to the network. Every time I tried to it would save that the target machine actively refused it. This usually isn’t a problem for me as I just search for the solution but this was not the case this time. I could not find one piece of documentation that talked about this issue in context to pure data or processing, anything I tried to find on the OSC issue was outdated and didn’t even use the same syntax that is used in PD now and on top of this I am running out of time in my own personal life because of the economic crisis going on meaning that I have to freeze to eat, I now I have a workshop I have to plan for which originally was not my job but now it is mine alone( and this is not even including my own shifts at my actual job).

Because I can’t use OSC anymore, at least not on my machine, because it seems to not want to work at all and the only way I think I could fix it is by doing a clean boot which is something I don’t want to do as I’ve got a lot of dated programs that I’m not sure I could download again.

Without the communication between programs, Processing is redundant and while I did learn a lot and will definitely upload it with my Pure Data patch, the most I can do is change for toggling of the semicircles to being based on key presses instead of mouse clicks. It’s unfortunate because the visual aspect of this piece wasn’t necessarily meant to be very robust, it was just meant to serve as a controller for the soundscape and is now little more than a pretty picture.

How can I change my PD patch to show off a bit of what I’ve learnt through the designing sounds textbook while still keeping the randomised nature of the piece? I’m going to achieve this by using the random object and smart use of the key and select objects.

I was quite quick at remodelling the patch at the start, but at this stage, everything wasn’t working and it was just bugging out. It took me about 2 nights to realise that I just had a duplicate patch open that was messing with everything.

Above is the patch completed with all the working hotkeys. After my little hiccup, it actually wasn’t too hard to do what I wanted to do and also solve the conflicts or oversights.

The last thing I did before my write-ups was create an offshoot of my Processing sketch and change all the mousePressed() functions into keyPressed().

Before going down this pathway I had looked into Open Sound Control(OSC) as a means to have Pure Data communicate with Processing. I knew that due to Pure Data and Processing having different languages that it would be tedious to make(in the sense that I have to be really careful with wording in both programs), but out of everything that I’ve learnt over the weeks, this seems to be the easiest part.

From my understanding, the way in which OSC works is that you have to define the messages sent and received in the programs themselves as OSC is just a means of encoding information in a format that every program has support to understand. This means that I could theoretically have Processing send numerous amounts of messages to PD and then I could unpack and use all the different messages to trigger bangs that could be sent to different parts of the patch. The plan is for the randomisation to then be sorted on Processing’s side.
This idea of having only 1-way communication is similar to this image from the Floss manual.

However, as I have been researching Pure Data and OSC, a lot of the tutorials and information on the topic are outdated or in reference to PD Extended, an outdated offshoot of PD Vanilla. Even in the diagram above it has forward slashes which are redundant in newer versions of PD Vanilla.

Another big issue I’m having is that despite having the newest version of PD as well as the externals to use OSC properly, my Pure Data isn’t recognising a lot of the objects like sendOSC or receiveOSC.

Despite being very busy to the point where I’m planning a week-long workshop for high school students in my community, I’m actually decently ahead in my own timeline for this project and I do believe I’ll be up to iron out all the kinks in the end.

What I need to learn for Processing is the language, how to make visual stuff and how to track the mouse position. Luckily upon opening the program, I can learn all these things quickly. I used the tutorial on the Processing website on top of the example sketches to formalise what I’m learning because I haven’t coded in java before.

Here’s the first version of the visual part of my project.

With an ellipse around the mouse to better understand what is being affected by the mouse

I’ve tried to implement toggling of the circles if the mouse is in the area of the circles and is pressed, but it’s not working at all

The circles now toggle and I’ve gotten rid of some bloat from the sketch; however, I can’t seem to solve the problem of how the circles change to the next ellipse’s colour, but I have a feeling that I would have to make each circle its own unique object(using the class keyword maybe?).

Now that I have a working GUI, I think I can leave this section to learn about OSC and how I’m going to implement it into pure data.

All I need to learn and create for the Pure data portion are the noise generators.

I decided to create 2 different noise generators for each of the 4 layers so that there is a 1 in 2 chance for the same generator to be picked again on each start-up. This means that there would be 128 different permutations of the program, I think that’s fine for variance. Those noise generators are:

Water;

Wind;

Bird calls;

Countryside footsteps;

A police siren;

A pedestrian crossing;

Car engines;

“Urban footsteps”?

While I could go through all the patches, the first noise generator I decided on was the pedestrian crossing as it was one of the tutorials that were on the moodle page (the designing sound one).

The basic version is something like this

You use the toggle to activate the metro object that is sending a bang every 100 bpm. Then the float object outputs the stored float which also triggers the “+ 1” object(binary arithmetic object), adding 1 to the float and then sending that back to the float object that will then output the float from the binary arithmetic object to the mod object. The mod object divides the float input into 2 and outputs the reminder (0s or 1s in this case). Those numbers are converted into a signal that controls when the oscillator at 2500hz is heard(so the amplitude), creating the classic beeping sound from a pedestrian crossing.

This small part would still have to be changed a lot, like balancing it with the other sounds, giving it a ramp (using the line~ object) so that it doesn’t click at all (I hope) and also making sure it turns off when you turn off the toggle.

All of these patches are ones made after going through all the practicals in the Designing Sound book, which are all designed to work stand-alone, something I have to solve.

When trying to learn what every does in the PD patches, I struggled a great deal with understanding the order of maths operations and what operation is appropriate for different situations, these 2 pet peeves really exemplify themselves in the bird call and footstep patches.

At this point, the patches are far from finished but after I’ve learnt all the necessary tools I’ll revisit the patches for one final iteration.

In this blog, I’ll be talking about my final idea through my initial statement for the project, why it is my final idea and what I need to do to achieve it.

My Initial statement/plan for the project is as follows –

“Making a pseudorandomised interactive “audio jungle”. This will be achieved by creating a patch in Pure Data to generate the sounds and then using Touch OSC to let it communicate with Processing 4 for the visual/interactive aspect of the project.

The plan is for there to be around 6-8 layers of sound in which each layer will all around 2-5 different possible noise generators that will either be synthesised sound that mimics the sounds of nature or the sounds of urban living.

Every time the program in Processing is run, 1 noise generator will be picked for each layer, in which certain parameters in that noise generator are randomised within a range.

The Processing portion of the project will act as the controller and the visual aspect of the piece. When the patch is run, each layer of sound would be able to be turned on by moving the cursor into its section of the window. All the sections overlap in the middle and make something akin to a ven-diagram so that every sound would be played when the cursor was in the middle of the screen.

There would be around 12-30 different noise generators with 6 layers or 16 to 40 generators with 8 layers. If having that many digital outputs/inputs communicate with Processing is not viable, having the parameters of 8 different noise generators randomise would be a good alternative. “

I’ve decided to fuse the randomise sound project with a soundscape patch similar to what Milo showed in class, using PD as the audio engine and Processing 4 as the GUI. I picked this over the other favourited idea as the Simon Says game was way too complex for me to understand when I tried to learn a few things about how to go about it and it sounded like a brilliant idea for the gallery space instead. I opted to use Processing 4 and OSC and the intermediary between the 2 programs as it seemed like a useful thing to understand for future projects.

The things I need to decide for this project are what sounds I will include in the soundscape, how to use pure data in a more detailed way, how to use processing and by extension the Java coding language(I believe) and how to get them to communicate with OSC.

What I could foresee happening with this project is that I might be overestimating the abilities of OSC especially after briefly researching it and realising about messages between Processing and Pure Data are limited in a way that it would be easy to make a lot of mistakes.
Also, the number of noise generators and the number of controllable variables are subject to change as I do think there are far too many in the initial statement and it’s too complex to be enjoyable to interact with.

Now I know I’m going down the option C route, I thought it prudent to make clear my goals for this project, brainstorm more ideas for what I could do and then decide what frameworks for this project I like the most.
My main goal is to learn skills that I will need for element 2, which is making an installation for a gallery. My practice centres around creating novel things and the process of making them, meaning that the piece would have some level of physicality and probably a digital backing through a BELA board and coding. In this project, I will try to learn and understand Pure Data or at least a program I can use for making sound art. A secondary goal will be to incorporate interactivity into the piece to try to understand what is enjoyable to use for a listener.

Ideas I have based on these goals:

A step sequencer that uses pre-made samples and easy hotkey controls to control the composition of the piece;

A fully fleshed-out hybrid of the noise toaster and a sequencer with again easy hotkey controls for live usage without a mouse;

Simon Says type of game that is played by rebuilding either the sound that is played or the waveform.

A spectrogram player that could take an image and turn into sound.

A composition that is slowly discovered through the movement of a mouse.

A randomised sound piece.

A more complex and in-depth version of painting waveforms that then play as you’re doing it.

While I did play around with a few other ideas, these were the standouts that I could actually imagine fleshing out and following through with. There are 2 prominent ideas that I really like, the randomised piece and the Simon Says type game.
Out of all the ideas they were the only ones that fulfilled the first goal of learning the most. I’ve used pure data in the past and understand how to do basic things in it, and a lot of my ideas were just showing that knowledge or making the patches I know more user-friendly. These 2 ideas I pointed out are both things I have no idea where to even start so I know there is a lot to learn in making them.
One thing I have to think about now is if I should use a different program instead of pure data or in conjunction with it as PD isn’t really that robust graphically compared to other programs and one of these ideas has a visual portion to it. On top of that, I know my art usually has a visual aspect to it but also isn’t conventional in how the audio and the visual interact, making me think that it could be a better idea to use or learn an intermediary program so that I can use that program in conjunction with any other program I use (for example being able to create sound pieces using FFmpeg and the command line).

In class this week, the concept of Gesture and Texture, theorised by Dennis Smalley came up. Texture means something akin to the “normal” sound something makes, like the sound it would make in a steady drone. Gesture is then how that sound is articulated via action or interaction with an object. Specific gestures can also be linked to specific sounds like one expects to hear swooshing sounds when someone waves around a sword.

When I thought of applying this concept to my work, I realised that I could use it to understand if the gestures in my work are congruent with the sonic reality which is integral to something like a soundscape.

While I can apply this framework of thinking to multi-channel pieces, gesture doesn’t really exist as strongly in the digital world as fabricated noise is mostly out of context of the actions taken to create that noise and thus there would be no clear gesture. Arguably, trying to create gestures in fully synthesised pieces isn’t wanted as it makes more sense to push the boundaries of a medium than work towards mimicking a different one.

Interesting concepts aside, I have finally decided to go down the Expanded Studio Practice for 21st-Century Sound Artists route for a number of reasons.

As mentioned before, I wanted to have a heavy focus on a narrative but it would be a waste outside of an exhibition space as no one would be able to receive it properly other than my tutor. I thought was strange before but now I know that element 2, I’ve realised that if I want to do a multi-channel piece, it would make sense to do it for the gallery space where I could show other people my work.

The second reason was learning or judging what was more conducive to my learning. While multi-channel is something I’ve never done before and would be useful to have experience in it if I wanted to be that type of artist in the future, if I am a sound artist proper rather than say a sound engineer or producer, coding will as always come up in my work. 

Even before I was making sound art, most of my pieces were held back by the fact that I don’t know how to code as my ideas would have some type of iterative process built in or something that lends itself to coding like a set of parameters of randomising parameters that then affect other parameters. 

This focus on learning is a product of reading up on literature for my sound paper and one of the ideas that I have for it talks a lot about learning and if a university course lends itself to that.

I know that if I want to push the way I work, I would need to learn to code, so I’m decided to make a sound piece in pure data and maybe a different program too. 

My ideas for the project haven’t really developed all too much since the last blog. Right now I’m stuck on what the context of my work is, as the purpose of the work is clear to me but I have no idea what the form will be like. If I still keep interactivity at the forefront of my mind the piece I make should have some easy way to control the sound generated in a way that doesn’t take a lot of creativity from the listener to do.

Before I decide on what pathway I want to focus on, I wanted to take some time to look at a few sound artists that will hopefully inform my decision as well as what I want to do for this project.

Natasha Barrett is a British contemporary sound artist and composer that focuses on acousmatic/electronic multi-channel compositions for installations and concerts. She says in her about page that she is trying to discover “new music” and “the way it can touch the listen”, an idea that is shown greatly in a 2021 piece by her Growth.

Growth is a spawn of the corona lockdowns and a part of her album Heterotopia. Based on the articles I have read about the album, it pushes heavily into the musique concrete direction to show, at least in Growth, the isolation and forced ingenuity of composers during the lockdowns. The piece shows this within its three different acts. The first is an erratic and unsettling fast-paced section that could be seen to mimic the start of the pandemic and that panic. The middle section is a calmer and more serene section that still has that unsettling underpinning that feels like it’s referring to how the piece came about, being trapped and forced to make art in replacement of social interaction. The last is a hopeful and wondrous section that might reference coming out of the pandemic or at least out of lockdown through the usage of the “light at the end of the tunnel” trope.

Growth is a piece that strives to be a new form of music and in my opinion, this is shown in one of the biggest problems with this piece, the lack of proper context from the artist. Despite being a researcher, I am hard-pressed to find anything about Growth or Heterotopia outside of the small descriptions on Bandcamp, meaning that I can’t understand her process fully. Although it is a decently interesting piece, it becomes just another covid piece like all the other ones which ironically is akin to popular music, similar to everything else from the time.

https://natashabarrett.bandcamp.com/album/heterotopia

That condemnation of her work is quite harsh and I think Natasha Barrett’s skills and practice are best shown in her installations. Looking through her website, it seems that she makes a lot of historical installations that are usually based on the history of the space. 

While not one of those history-inspired installations, Subliminal Throwback is a site-specific 8-channel installation that was part of a Norwegian research project that was investigating how sound art could inspire awareness of outdoor sound environments. 

Installed in an amphitheatre in Oslo, Subliminal Throwback is a piece that is meant to exist with the natural soundscape (in a way), shown through the difference in composition between the studio version and the on-site version.

Something that really interests me with the mechanics of the installation itself is the custom-designed beam-forming speaker that directs sound to purposely reflect off space itself. I find this brilliant because I know from personal experience that pieces like this are interesting to the average person and while the average listener probably would not necessarily appreciate the technicality or meaning of the work, an impact is definitely made with the novelty of the idea of directional sound.

https://natashabarrett.bandcamp.com/album/subliminal-throwback

The other artist I really want to talk about is Yuki Suzuki. Suzuki is a sound artist that is most notable for his installations. These installations focus on an active interaction between the listener and the installation itself, either using the listener as the primary noise generator with the installation transforming that noise or forcing the listener to go out of their way to hear the piece (like in Sound of The Earth Chapter 2). 

At the heart of Suzuki’s work is the relationship between the subject and the object as well as the necessity to create fun and palatable sound art pieces.

That fun and the palatable sentiment is palpable in his installation The Welcome Chorus. This installation was made for Margate NOW in 2019 and centred around the ‘Kentish experience’ which is why it initially interests me as someone who grew up in Kent. 

The crux of the piece is the horns, the AI with its lyrics and the giant microphone or ‘conductor sculpture’. An AI was trained to sing lyrics about the aftermentioned experience through the 12 horns and while the installation was up, listeners would be able to speak or sing their own lyrics through the microphone. This would then play back to the listener as well as further train the AI.

I was able to see this during my foundation year in the arts and what I believe is one of my first interactions with a sound art piece. While I knew nothing of the practice, if it wasn’t obvious already, the thing that stuck with me was how fun it was the interact with and how it was interesting for those who weren’t interested in sound art at all.

https://www.yurisuzuki.com/projects/the-welcome-chorus

As I stated before, I want to bring a level of interactivity to my work, but as someone who has an unhealthy obsession with technological listening, the mechanics of the art piece I have talked about greatly interest me as well. This does make me think that option C would be better than option A; however, I have a stronger idea for Specialising and Exhibiting and is something I’m finding hard to ignore.

In this module, I picked the Spatialisation for Installation & Performance and Expanded Studio Practice for Twenty-First Century Sound Artists options. I’ve always found great interest in multitrack pieces as installations as they lend themselves to that format (unlike Sound for Screen strangely). Option C was my other choice due to its technical nature as well as its usefulness to me and my practice in the present and the future. 

I am leaning more towards option A as my pathway for this assignment as I am focused on creating narratives and stories. That aspect of sound art can be hard to express in a pure data patch. Artists like Yuki Suzuki have inspired me to make my work interactive, with my initial ideas for a multitrack piece focusing on interactivity and narrative. 

A few ideas that come to mind are:

A dialogue between two characters in which the sweet spot for listening travels around the eight-speaker setup, meaning that the listener would be playing the role of one of the characters in that dialogue;

A soundscape where the listener would be stationary in the middle of the ring of eight, listening to a peaceful scene that slowly becomes uncomfortable to listen to with the introduction of out-of-character sounds slowly creeping up on the subject.

Even with these initial ideas (which I quite like), a problem for myself is creating a multi-channel piece that focuses on narrative when I know I won’t necessarily learn that much. I usually would be fine with this; however, I find it awkward to make such a narrative-heavy piece and interactive piece that would only be heard by the tutor marking my work. No regular person would be able to listen to it in all its glory as most people don’t own an 8-speaker set-up. On top of that, I go to university to learn, not just express myself creatively so it might be better, in the long run, to make a pure data patch instead. 

I understand that one’s art shouldn’t be constrained by the fact that it’s for a university degree though reality cannot be ignored. If not stated explicitly otherwise, making a piece with content that is meaning meaningful can almost be seen as silly and ironic that something that has substance is graded.

 As someone who has been debating if “true art “exists and if art can be this “true art” if it’s made for reasons other than the reaction, it’s trying to elicit, trying to figure out what I want out of my university art pieces has been an interesting journey and one which I’m still figuring out.