Well, I'm back at the Electronic Music Unit after twelve 'glorious' months of Music Education. Now I'm a fully qualified music teacher. I will fail your future children, so be nice to me.
First things first, I'm going to update this page design with something more... formal. As fitting as it is in regards to my Honours topic, I'm sick to death of it and I actually enjoy HTML coding.
Secondly, I'm going to find out whether I even need to use this blog any more. I think it works well for documentation purposes, even though all the file links die in 6 months (apparently).
For those interested, My Honours topic is something along the lines of Context Driven Game Audio Design. Basically I'm looking at how and why certain sounds 'fit' better than other sounds for a given situation in video games.
For my project I endeavoured to make a multiple-sample break-beat cutting program. The idea was that you could use it in a live situation to cut up and mix samples into a real-time composition.
"MultiCutter" ultimately did not reach its final form, however the concepts were all realised in some sense. The program uses GUI controls to change values of BBCut algorithms, which in turn change the playback of samples being cut up by each respective BBCut. I feel that the biggest success of my program was the efficiency of the GUI creation and BBCut instancing. The GUI slider, label and number box are only written once in the code, and are then reproduced for each control parameter for BBCut. This process is then repeated for each BBCut, meaning the slider, label and numbox are again duplicated and assigned a parameter. The BBCut is also only written once, but it is duplicated for each sample. This easily expandable environment would have eventually been used to change GUI elements and number of BBCuts depending on the number of samples chosen by the user.
I had two fatal bugs in my program; the worst involved trying to get each instance of BBCut to read an individual sample source. Each instance was assigned a number (under argument xPos) which I attempted to replace with a letter so that BBCut could use it as a buffer address. Replacing the instace number with a letter was easy (I left that bit of code in comments) however it simply would not work in BBCut as a buffer address. This meant that each instance of BBCut would not load a different sound file. Replacing the buffer address with a constant (“g”) results in each BBCut playing the same sample. This is how I have left the patch, as it proves that each instance of BBCut is individually controllable by its respective GUI.
The other bug involved using a master tempo slider to control every instance of BBCut. Looking back, I would have needed to redesign the code so that the TempoClock was outside of the instace-creating do-loop. MultiCutter + Sound Samples : 370kB
The segment of Nosferatu I chose to create sound and music for has five scenes: The village talking about the vampire, the first Knock chase scene, Ellen sewing, the second Knock chase scene, and the Orlock and Ellen scene. I based the sound design around the fusing of two approaches; old-sounding piano and modern sound design.
SFX
THE PIANO To make the piano sound old I recorded a slightly out-of-tune piano and cut off the low and high frequencies using an equaliser, which resulted in a tinny sound likened to an old recording on a vinyl LP. I also recorded the pops and scratches from the start of a real LP so that it sounds like somebody putting on a record at the beginning of the clip. I originally had the LP crackling throughout the entire clip, however it became too distracting.
THE CHOIR The choir recording originally consisted of my own voice layered several times for each part, however the soprano part ultimately sounded too much like a male falsetto. I asked fellow Music Tech student Poppi Doser to sing the soprano and alto parts, which certainly helped with the falsetto issue. I was very grateful that she could reach the high B flat, as it is the pivotal part of the end of the scene. I filled out the ‘rest’ of the choir using a lot of reverb.
SCENE 1 : VILLAGERS DISCUSSING THE VAMPIRE The sound effects in the first scene are fairly subliminal. For a later scene I recorded the sound of a table dragging across a hard floor, and for this scene I used some of the accidental sounds from the room, reversed to add an unnatural feel. There is also heavy stereo delay applied to the sound, which gave it a more swirling feel.
SCENE 2 : KNOCK CHASE PART 1 The obvious sound required for this scene is the footsteps. After much experimentation I decided that creating the footstep sounds for a crowd of 70 people is beyond the scope of this project, and the sounds I did create overwhelmed the piano. The sound I used is very much underneath the piano sound, except where there is only a single person running. For a couple of the Knock running scenes where he is the only person in frame I matched up footstep recordings one foot at a time.
SCENE 3 : ELLEN SEWING When Ellen looks up from her sewing I introduced a raspy scraping sound (a table being dragged along a floor) combined with a deep growl (my own voice), which would later be connected to Orlock. I used the sound here to insinuate that Ellen was somehow connected to Orlock, however this would not become apparent until the final Ellen and Orlock scene.
SCENE 4 : KNOCK CHASE PART 2 Like the previous chase scene there were many footsteps to account for, which ultimately took a back seat to the music. As this chase is in the wilderness, I superimposed a recording of outdoor ambience including bird sounds. The very end of this scene uses a transitional sound effect I created at my work by hitting a large bin with a hammer. This sound was run through numerous reverbs until it sounded like a giant wave of sound. For the scene I only used the build-up of the sound, then cut it off suddenly to emphasise the sudden chance of scene.
SCENE 5 : ELLEN AND ORLOCK For this scene I went all-out with the sound design, as I wanted to change the feeling from action to suspense. The first sound that can be noticed is the Orlock sound mentioned previously, using a table dragging sound mixed with a low growl. I use this throughout the scene whenever Orlock is on screen, making it louder as the scene progresses. The table dragging sound is also used as a suspense device, as I mixed together many different recordings of the table being dragged at different severities so that the sound builds up over the entire scene. The main benefit from this sound is its sub-harmonic quality, which sounds particularly good through a sub-woofer. I also used a recording I made at work of a bunch of industrial-sized bins being towed by a little car. The rumble that these bins make can be heard from the fourth floor of the building, so it was particularly useful for creating suspense towards the end of the scene. It can be heard easily during the credits. Another sound I used was screams that were recorded in the stairwell of the Schulz building. The natural reverberation that occurs in the 12-storey room meant all I had to do was reverse the sound, so you can hear the reverb before the sound itself. I used this most prominently as the lead-in to the choir section.
MUSIC
SCENE 1 : VILLAGERS DISCUSSING THE VAMPIRE For this scene I created a heavy, unmetered piano part that uses chromatic lines to mimic speech. This is to supplement the superimposed text that occurs when the villagers are discussing the vampire. The left hand is playing an ostinato between D, G# and D. This tritone hides any true tonality, and when combined with the chromatic ‘vocal’ lines in the right hand it creates a very uneasy feeling.
SCENE 2 : KNOCK CHASE PART 1 The first chase scene uses a low repeated D pounded in a constant rhythm, with a clichéd horror melody played in octaves with the right hand.
There is also a small three-note motif played, which is later established as the Knock theme. The end of this scene finishes with a repeating high D octave, ultimately landing on a low cluster chord to coincide with Knock jumping off of a roof. The high D octave is in fact a segue to the next scene.
SCENE 3 : ELLEN SEWING The Ellen Sewing scene begins with a haunting melody starting on D. The intervals minor 2nd and minor 3rd (also a diminished 4th from the root) were chosen to express that despite the scene being a fairly relaxing moment, there was still tension regarding events occurring elsewhere.
SCENE 4 : KNOCK CHASE PART 2 The previous scene quickly leads to the introduction of the second Knock chase scene theme. A simple six-finger pattern is played on five notes.
This the fades into the same pattern three octaves lower, forming the start of the chase scene. The 6-note pattern moves across several chords, generally only changing two notes at a time. The Knock theme is also played over the top of the pattern in the second half of the chase scene.
SCENE 5 : ORLOCK AND ELLEN The final scene is a much more heavily composed piece, and the changes in the composition generally mimic changes in the clip. I included voice for the climax as the piano simply wasn’t able to bring the scene to life. While doing so certainly takes a step away from the ‘old piano’ concept, I feel it was a necessary step to properly realise the scene, and still falls under the banner of ‘modern sound design’. Following is the complete score for the Orlock and Ellen scene, sans the sound effects.
Finally fixed the template, turns out Ripway.com (where my pictures are stored) was incorrectly registering download bandwidth and maxing out my limit every day. I sent an email to ripway support and within an hour they had fixed the problem. Not bad for a free service!
Today the remainder of the 3rd Year students presented to the Music Technology horde. I was first, presenting my failed MaxMSP patch from Semester 1 2007, Fantastical Metal. The patch seemed to work well (in it's own way), but I would have liked to present my SuperCollider patch from last semester so as to round off a 'My Failures' theme.
Luke's SuperCollider patch was aesthetically pleasing to listen to, especially the combined organ sounds. The end was a little abrupt, and I think it may benefit from a nice thumpy kick drum throughout the piece.
Will presented his Max patch from last year, which we didn't really get to listen to. His environment sounds for a video game were well done, however they seemed to pale in comparison to the Half-Life 2 SFX.
Dave's patch obviously took a lot of effort to create, and the inclusion of drum beats and random samples helped to 'keep it from being boring'. I like the idea that the patch sounds fairly different each time it is run.
Last was Jake, who played us his SuperCollider patch from last semester. This was a 5.1 channel industrial ambient piece, and was certainly intense and very enjoyable. Jake obviously has a good understanding of the 5.1 sound field, with a 'spawning' effect perceivable throughout.
David "Dirty Harry" Harris was back in the driver's seat this week, and steered the bumbling cohort of disenchanted musicians (also known as Mus-Tech Students) towards the insanity of David Lynch. I have seen very little of David Lynch's films, but apparently Eraserhead was a good place to start. Well, so I was told, however I found myself developing a rather shocking headache throughout the movie. I know it's kind-of the point, but I found all of the scenes in which the conversations travel at snail's pace so incredibly, incredibly boring. Silence is only funny for so long, then it just gets old.
From a sound-design aspect, the film was very interesting, but the tedium of the on-screen action (using the term loosely), particularly in the first half hour, was often too overwhelming to maintain my interest in any aspect of the film.
David Harris. "Music Technology Forum - Semester 2, Week 8 - My Favourite Things III." Lecture presented at the Electronic Music Unit, University of Adelaide, South Australia, 18/09/2008.
This week I had to bail on forum due to an over-active teaching practicum, a disease which I apparently contracted when I first started University. According to other student's blogs, it was student presentations. When I found out what Doug's MaxMSP project last semester was (a chord progression generator), I had a brainwave of what to do for my own presentation - my failed projects. I'll start with my Auto-Piano MaxMSP thingy I made semester 1 last year, then finish off with my craptabulous Rainforest generator I created last semester in SuperCollusion.
Can't wait!
1. Stephen Whittington. "Music Technology Forum - Week 7 - 2nd and 3rd Year Presentations." Lecture presented at the Electronic Music Unit, University of Adelaide, South Australia, 11/09/2008.
This week we presented the work that we had done thus far on our Film Sound project. I was the only one who didn't have something synced to the movie already, mainly because I didn't realise we needed to this early on. I played the musical ideas I have had on the piano, and informed the class of my plans to make the music sound as old as the film, meaning bad quality and inconsistencies. To be honest I believe this will be more difficult than just recording straight piano, as I am having to record to an aesthetic. The new Zoom portable recorders will no doubt come in handy when sourcing an old piano.
John and Dave are taking a more traditional approach to the sound design, using symphonic instruments and/or pounding bass-driven beats. Jake's approach is quite compelling, using mostly ambient sounds interspersed with pulsing rhythms. Luke has approached the project with a theme of 'aggravation', whereby the audience is made to feel uncomfortable at convenient parts of the film. I really hope this works out, as I would love the see (and feel) the results.
Luke Harrald. "Audio Arts 3: Semester 2, Week 6. Student Presentations." Lecture presented at the Electronic Music Unit, University of Adelaide, South Australia, 2/09/2008.
This week several lucky students had the chance to present their collected Indian Aesthetic sounds that they spent all yesterday evening working on, and the rest of the class had to Guess The Emotion™. Most notable was Freddie's, where he explained how studies have proven that infant humans are able to communicate emotions through different types of crying. While he was playing the well-researched sounds, I, as most of the other people in the room, attempted to 'understand' what the baby wanted, based purely on the way it was crying. It was not until after all the samples were played that Freddie informed us that we were supposed to be writing how the cry made US feel, not what we thought the baby was feeling. I love a good twist at the end of a boring movie!
This week Stephen Whittington (1) played us a selection of works from Negative Land. Since 1980, Negativland have been creating records, CDs, video, books, radio and live performance using 'appropriated' sound, image and text. Mixing original materials and original music with things taken from corporately owned mass culture and the world around them, Negativland re-arranges these found bits and pieces to make them say and suggest things that they never intended to.(2) The outcome of this seems far too 80s for my liking, and also has a little too much of a 'cool to hate' vibe. While I can tell a lot of effort went into the creation of their new DVD Our Favourite Things, ultimately I did not enjoy it. Perhaps this is due to my inability to accept pop art as a respectable creative field, but in the end I just did not enjoy the sound or the visuals or the combination of the two.
1. Stephen Whittington. "Music Technology Forum - Week 5 - My Favourite Things II." Lecture presented at the Electronic Music Unit, University of Adelaide, South Australia, 28/08/2008.
Here are some of my interpretations of Indian Emotions in a musical/auditory context. HASYA (HAPPINESS/JOY/COMIC) Major tonality Consonance Bouncy rhythm Laughter
ADBHUTA (WONDER/REVERENCE) Augmented chords Elongated notes High pitched notes
VEERA (COURAGE/HEROIC) Resolution Major tonality Loud Trumpets/horns
KARUNA (COMPASSION/LOVE/EROTIC) Slow tempo Saxophone Sparse instrumentation Kiss sound (etc.)
This week some of the 1st Year students presented whatever they wanted.
The first couple students played their Musique Concrete piece from last semester, which was all kind of the same.
Another guy presented the music he created for a video game. It was a little clichéd, but there was good use of Taiko-style drumming. It would have greatly benefited from a better sound bank, as it was a little General MIDI sounding.
Another one played a recording he mixed and mastered. This was quite impressive, particularly with the recording of violin and sax, which I know from experience are difficult to get sounding good.
The following guy presented his Musique Concrete, which used found sounds from his house. The result was quite good, particularly the 'thunder' from the shed door.
David Harris presented all the things that make him warm inside- Terra Rapture (his own composition) and a Schubert work for strings. His own piece made great use of sul ponticello, which requires the player to bow the string very close to the bridge. The sound is scratchy and airy, and is very effective when done by a quartet. I enjoyed Terra Rapture more then the Schubert, as is was more unpredictable and less contrived. This class made me pine for the Forums of old, where DH would play us music from his eclectic collection of LPs.
All I wanted is to get some array-built SCSliders to control some things in a Pbind. Sounds easy enough, but none of the examples used SCsliders. It was either EZSlider or SCRangeSlider. Considering how whole-heartedly the EZSlider was disapproved of, I found it irritating that there was no explanation of how to implement the SCSlider into an array-built interface. I tried for ages just to get the example patches to work with SCsliders, but it wouldn't work. The SCSlider Help just uses the copypaste method. I'm sure there is some tiny syntax thing I have to do to make it work, but if I haven't been shown it and it isn't in the Help file then I can't know what it is.
I ended up having to use the EZSlider, which I (eventually) got working for controlling the speed of the Pbind. Then when I tried to control another part of the Pbind (\octave), the speed was reset to a value not even in my code. And now when I change the speed, the octave gets reset. And none of this actually appears on the interface, you can only hear it.
I have to say that MaxMSP sh!ts on SuperCollider for GUI. If making a crap-looking interface takes this long, I don't even want to think about making a good-looking one. Needless to say, I spent no time at all on the prettiness.
The Matrix has been one of the most influential action movies ever made, most notably with its pioneering use of 'bullet-time'. I have probably watched it over 50 times in my life, and the famed lobby scene will always be one of my favourite moments in any film.
The lobby scene begins with the pile driver sound that started in the previous scene with Agent Smith and Morpheus. The presence of this sound in the previous scene seems to pre-empt the impending character Neo entering the building. The sound is linked to Neo's feet, giving a great weight to his footsteps. This is then joined by a fast hihat rhythm, increasing the tension of the scene as soon as you see the high-level security. There is a strange 'zip' sound as Neo puts his bag on the x-ray conveyor belt, which does not seem to fit the on-screen action.
When Neo opens his coat to expose his personal artillery to a security guard, a large whooshing sound enhances the significance of such a simple action. At this moment a pile driver hit echoes and the hihat stops, mimicking the shock of the security guard. Word count reached. Holy shit!
Oh how I miss MaxMSP. I just want to add a picture for the background! Whatever. This week I attempted to create a GUI that could be used for sound file playback, looping and surround panning. I decided that rather than copypasting the same code several times for each instance, I placed one set of code into a do function and let it populate the window itself. This also allowed for quick customisation of the interface- you can choose how many instances you want with just a single number change. The window automatically resizes to fit the number of instances. This also led to a whole bag of troubles regarding GUI object output, but whatever.
I also ripped off Dave's background code to replace the metal one, but I couldn't get a good mix of colours, and the metal looks so cool anyway.
Considering how much effort is required, the outcome is quite lame. Surrounder.RTF
Haines, Christian. "Creative Computing - Week 1 - Semester 2, 2008: Graphical User Interface (1)." Lecture presented at Tutorial room 408, level 4, Schultz building, University of Adelaide, 31st of July 2008.
Film sound is not an interest of mine, partly due to the 'Sans Video' we had to do in 1st year *shudder*. I understand it takes a great amount of experience to successfully produce the sound for a movie, and I am curious as to where this experience should be coming from for our projects. I have not done the Sound and Media elective or studied composition, so I am hinging on the weekly classes having some digestible/regurgitatible content. I am also bemused about what we exactly have to do, as there is no project documentation yet. It is strange that we have to choose the clip this early on, when we don't know the project requirements. For example, I would like to produce the sound for clip 3, however if we have produce all the foley with found sounds AND produce our own musical compositions then this clip would be ridiculously complex to produce.
Regardless, an idea I have had is to produce the sound as if it were recorded with the archaic sound recording equipment available around the time of the movie. I feel that this 'lo-fi' approach would better suit the film than some grandiose Bourne Identity style production.
Again, I don't know if this would be allowed, so it's really just all speculation.
Also, I haven't ever seen Nosferatu before.
Harrald, Luke. “Audio Arts – Week 1 – Semester 2, 2008.” Lecture presented at the Audio Lab, Level 4, Schultz building, University of Adelaide, 29th of July 2008.
Just like old times. Our first forum for this last-for-some-but-not-for-me semester somewhat centred around the role music plays in our society. Personally, music is generally used as a tool for boredom alleviation or experience augmentation. For example, today I spent 2 hours catching buses and trains from Modbury to Gawler, during which I listened to some 80's classics, such as:
Hall & Oates: Out Of Touch Wang Chung: Dance Hall Days Michael Jackson: Billie Jean Go West: Call Me INXS: Kiss the Dirt Yes: Owner of a Lonely Heart Aneka: Japanese Boy Talk Talk: Life's What You Make It Joe Jackson: Stepping Out The Fixx: One Thing Leads to Another
I know what you're thinking: "What are you, gay?", but no, these songs were in fact a part of an hour-long MP3 which I ripped from the Grand Theft Auto: Vice City game disc. You may know it as "Flash FM", and it even features all the adverts and DJ banter. In this instance I was listening to music to alleviate boredom (and it should be noted that this music could not be used for much else). As for 'experience augmentation', this often occurs when in a moving vehicle (eg. RATM), walking (recently Earth, Wind and Fire for some reason) or dancing (Parliament or funk in general). Word count w00t!
1. Stephen Whittington. Music Technology Forum, Semester 2, Week 1. "Technology Theory & Culture." Lecture presented at the Electronic Music Unit, University of Adelaide, South Australia, 31/07/2008.
I suppose late is better than never, except at university. My major project for Creative Computing is centred around a concept of environment and weather, in particular rain in a rainforest.
In the main composition, the rain that builds up and then subsides was the original concept for the program, and is the main controller of thematic material, namely drums and bass (both natural inhabitants of a rainforest).
The animal sounds were downloaded from www.a1freesoundeffects.com and http://www.sound-effect.com/ for free. The low quality of the sounds meant I had to change the sample rate (from 8000Hz to 44100Hz) and use Noise Removal in Audacity. Once the files were played in a SynthDef with a reverb, the low quality was barely noticeable. I wanted to use Gverb on every animal sample, however this would crash the server after about 5 animals were initiated. I considered using FreeVerb, which would allow me to pan many sounds within a stereo field without panning the reverb, however the implementation took far too long and I eventually cut my losses and recorded the patch in sections. Due to such CPU/memory issues, I did not implement an automatic play-through code, as the server would inevitably crash.
Oh, and all soundfiles were originally mono (even the thunder).
The drum SynthDef was created by accident while trying to build a flute/pan pipe instrument. Utilising a nested sequence of Pseqs (based on a weekly task), I created a form of drum sequencer, a part of which includes a random order sequencer and a random beat sequencer. I particularly like the effect I created for this drum sound, as I am running a ClipNoise into a klank, meaning the amount of resonance in the drum depends on how hard it is hit (ie how loud the ClipNoise is). Combine this with the drum sequencer which has attenuable amplitudes, and the the result is fairly cool.
By far the most taxing task was trying to create a realistic rain sound. While my result is not yet indistinguishable from the real thing, I feel I brought it fairly close. Dust, Dust and more Dust. Ironically.
My score is as good as they ever are, basically just showing a timeline of events.
If I had more time/better organisational skills, I would have created a cricket and frog synth, as this seems to be the main thing missing in the mix. You can imagine my state of mind after working on it for far too long in one sitting, all while listening to rainforest ambience - relaxing sounds are not the best stay-awake technique. Oh how I wish I still drank coffee.
A big shout out to my main man SuperCollider Help, who pulled me out of a dark time in my coding life and carried me to safety. Did I say shout out? I meant middle finger. David Cottle's book was very useful, as were the weekly tasks, so an actual shout out to them.
My composition is best heard with speakers capable of producing low frequencies well. Or with a subwoofer. Doof doof squawk. Composition: MP3 4.53MB
My SuperCollider code is in this order: Ocean breeze (background noise), Ambientator (ambient synth), Rain, Conga (feat. Drum Synth, Reverb Synth & Drummer), Bass (feat. Bass Synth & Bassist) and finally Animals (about 9 sections, includes Thunder, the king of all animals). Rainforest Code: RTF 27KB
You do not have to download the animal sounds to use my patch, unless you actually want to play them. Don't, they aren't very interesting. You can hear them in the above composition anyway. Animal Audio: ZIP 2.3MB
My patch in layman's terms. John Layman. Hmm, if he exists he should write one of those 'for Dummies' books. Program Note: PDF 38KB
This'd be me if I smoked.
