Tag Archive Binauralix

Byadmin

Spectral Select: An acousmatic 3D audio study

 

 

Abstract:
Spectral Select explores the spectral content of one sample and the amplitude curve of a second sample and unites them in a new musical context. The meditative character of the output created by iteration is both contrasted and structured by louder amplitude peaks.
In a revised version, Spectral Select was spatialized in Ambisonics HOA-5 format.

Supervisor: Prof. Dr. Marlon Schumacher

A study by: Anselm Weber
Winter semester 2021/22
University of Music, Karlsruhe

 


About the study:
In which forms of expression is the connection between frequency and amplitude expressed ? Are both areas intrinsically connected and if so, what could be approaches to redesigning this order?
Such questions have occupied me for some time. That’s why the attempt to redesign them is the core topic of Spectral Select.
I was inspired by AudioSculpt from IRCAM, which we got to know in our course: “Symbolic Sound Processing and Analysis/Synthesis” together with Prof. Dr. Marlon Schumacher and Brandon L. Snyder and which we partially rebuilt.
Spectral Edit works on a similar principle, but instead of having a user work out interesting areas within a spectrum of a sample, it was decided to use a second audio sample. This additional sample (from now on referred to as “amplitude sound” in the course of this article) determines how the first sample (from now on referred to as “spectral sound”) is to be processed by OM-Sox.
To achieve this, two loops are used:
First, individual amplitude peaks are analyzed out of the amplitude sound in the first “peakloop”. This analysis is then used in the heart of the patch, the “choosefreq” loop, to select interesting sub-ranges from the spectral sample. Loud peaks filter narrower bands from higher frequency ranges and form a contrast to weaker peaks, which filter somewhat broader bands from lower frequency ranges.

peakloop – Analysis
choosefreq Loop – Audio Processing


How small the respective iteration steps are affects both the length and the resolution of the overall output. Depending on the sample material, a large number of short grains or fewer but longer subsections can be created. However, both of these parameters can be selected freely and independently of each other.
In the enclosed piece, for example, a relatively high resolution (i.e. an increased number of iteration steps) was chosen in combination with a longer duration of the cut sample. This creates a rather meditative character, whereby no two sections will be 100% identical, as there are constantly minimal changes under the peak amplitudes of the amplitude sound.
The still relatively raw result of this algorithm is the first version of my acousmatic study.

Acousmatic study version 1


The subsequent revision step was primarily aimed at working out the differences between the individual iteration steps more precisely. For this purpose, a series of effects were used, which in turn behave differently depending on the peak amplitude of the amplitude sound. To make this possible, the series of effects was integrated directly into the peak loop.

Acousmatic study version 2


In the third and final revision step, the audio was spatialized to 8 channels.
The individual channels sound into each other and change their position in a clockwise direction. This means that the basic character of the piece remains the same, but it is now also possible to follow the “working through” of the choosefreq loop spatially. To maintain this spatiality, the output was then converted to binaural stereo for the upload using Binauralix.

Acoustic study version 3 – Binaural

 

Spectral Select – Ambisonics

In the course of a further revision, Spectral Select was re-spatialized using the spatialization class “Hoa-Trajectory” from OM-Prisma and converted to the Ambisonics format.
To ensure that this step fits in well conceptually and sonically with the previous edits, the amplitude sound should also play an important role in the spatial position.
The possibilities for spatializing sounds with the help of Open Music and OM Prism are numerous. In the end, it was decided to work with Hoa-Trajectory. Here, the sound source is not bound to a fixed position in space and can be described with a trajectory that is scaled to the total duration of the audio input.

Spatialization with HOA.TRAEJECTORY

 

 

The trajectory is created depending on the amplitude analysis in the previous step.
A simple, three-dimensional circular movement, which spirals downwards, is perturbed with a more complex, two-dimensional curve. The Y-values of the more complex curve correspond to the analyzed amplitude values of the amplitude sound.
Depending on the scaling of the amplitude curve, this results in more or less pronounced deviations in the circular motion. Higher amplitude values therefore ensure more extensive movements in space.

 

 


It is interesting to note that OM-Prisma also takes Doppler effects into account. As a result, it is also audible that at higher amplitude values, more extreme distances to the listening position are covered in the same time. This step therefore has a direct influence on the timbre of the entire piece.
Depending on the scaling of the trajectory, fast movements can be strongly overemphasized, but artifacts can also occur (if the distance is too great).
To get a better impression, 2 different runs of the algorithm with different distances to the listener follow.

 

Version with extreme Doppler effects which can result in artifacts – binaural stereo

Versionwith closer distance and more moderate Dopp ler effects – Binaural Stereo

 

In contrast to the previous sound examples, the spectral sound and amplitude sound have been replaced in this example. This is a longer sound file for analyzing the amplitudes and a less distorted drone as a spectral sound.
The idea behind this project is to experiment with different sound files anyway.
Therefore, the old algorithm has been reworked to offer more flexibility with different sound files:

Revised scalable version of the old algorithm for selecting from the spectral sound

In addition, a randomized selection is now made from the spectral sound on the time axis. As a result, any shaping context should come from the magnitude of the amplitude sound and any timbre should be extracted from the spectral sound.

 

ByVeronika Reutz

Composing in 8 channels with OpenMusic

In this article I present my ideas, creative processes and technical data for the patch programmed for the class “Symbolic Sound Processing and Analysis/Synthesis” with Prof. Marlon Schumacher. The idea of this text is to show the technical solutions for my creative ideas and to share the knowledge gained to help the reader with their ideas. The purpose of this patch is to take sounds from everyday life and transform them into your own composition using several processes within Open Music.

Responsible: Veronika Reutz Drobnić, winter semester 21/22

Introduction, Iteration 1

The initial idea of the piece was to transform everyday sounds, for example the sound of a kettle, into a different, processed sound by implementing technical solutions in Open Music. This patch processes and merges several files into one composition. There are three iterations of the patch that I worked on during the semester. I will describe them in chronological order.

The original idea for the patch came from musique concréte. I wanted to make a 2-minute piece from concrete sounds (not synthesized in Open Music, but recorded). This patch consists of three subpatches that are connected to the maquette in the main patch.

The main patch

Read More

Pages: 1 2 3

ByLukas Körfer

Speaking Objects

Abstract

In this project, an audio-only augmented reality sound installation was created as part of the course „Studienprojekte Musikprogrammierung“ (“Study Projects Music Programming”) at the Karlsruhe University of Music. It is important for the following text to distinguish the terminology from virtual reality (VR for short), in which the user is completely immersed in the virtual world. Augmented reality (AR for short) is the extension of reality through the technical addition of information.

 

Motivation

On the one hand, this sound installation should meet a certain artistic standard, on the other hand, my personal goal was to bring AR and especially auditory AR closer to the participants and to get them excited about this new technology. Unfortunately, augmented reality is very often only understood as the visual representation of information, as is the case with navigation systems or smartphone applications, for example. However, in my opinion, it is important to sensitize people more and more to the auditory extension of reality. I am convinced that this technology also has enormous potential and that there is a lot of catching up to do in terms of public awareness compared to visual augmented reality. There are already numerous areas of application in which the benefits of auditory AR have been demonstrated. These range from areas in which many applications of visual AR can already be found, such as education, increasing productivity or purely for entertainment purposes, to specialist areas such as medicine. Ten years ago, for example, there were already attempts to use auditory AR to enhance the sense of hearing for people with visual impairments. By sonifying real objects, it was possible to create a purely auditory orientation aid.

 

Methodology

In this project, participants should be able to move freely in a room in which objects are positioned and although these do not produce sounds in reality, the participants should be able to perceive sounds through headphones. In this sense, it is an extension of reality (“augmented reality”), as information is added to reality in auditory form using technical means. Essentially, the areas for implementation extend on the one hand to the positioning of the person (motion capture) and binauralization and on the other hand in the artistic sense to the design of the sound scene by positioning and synthesizing the sounds.

Figure 1

The motion capture in this project is realized with the Polhemus G4 system. The direction and position of a micro-sensor, which is attached to a pair of glasses worn by the participant, is determined by a magnetic field generated by two transmitters. A hub, which is connected to the micro-sensor via a cable, sends the motion capture data to a USB dongle connected to a laptop. This data is sent to another laptop, on which the binauralization takes place and which is ultimately connected to the wireless headphones.

Figure 2 shows two of the six objects in one variant each (angles of 45° and 90°). The next illustration (Fig. 3) shows the over-glasses (protective glasses that can also be worn over glasses) that are used in the sound installation. These goggles have a wide nose bridge to which the micro-sensor is attached with a micro-mount from Polhemus.

Figure 2

 

Figure 3

As previously explained, various decisions have to be made before the artistic aspect of the sound installation can be realized. This involves the positioning of the objects / sound sources and the sounds themselves.

Figure 4

 

Figure 5

Figure 4 shows a sketched top view of the complete structure. The six blue-colored circles mark the positions of the objects in the room and, of course, the sound sources of the scene in Binauralix, which can be seen in Figure 5. The direction and angle of the sources can be taken from the colorless areas (in Fig. 4), at either 45° or 90° angles, around the sound sources.

The completely wireless position detection and data transmission enables the participants to immerse themselves fully in this experience of the interactive reality-expanding sound world. The sound synthesis was carried out using the SuperCollider software. The sounds were mainly created through various tapping and clicking noises recorded by the SoundIn object, and finally changes and alienation of the sounds through amplitude and frequency modulation and various filters. By routing the sounds to a total of 6 output channels and “s.record(numChannels:6)”, I was able to create a two-minute multi-channel audio file in SuperCollider. When playing the file in Binauralix, the first channel is automatically mapped to source one, the second channel to source 2 and so on.

 

Technical implementation

The technical challenge for the implementation of the project initially consisted of receiving and reformatting the data from the sensor so that it could be used in Binauralix. The initial problem was that Binauralix is only available for MacOS and the software for the Polhemus G4 system is only available for Windows and Linux. As I had a MacBook and a laptop with Ubuntu Linux as my operating system at the time, I installed the Polhemus software for Linux.

After building and installing the Polhemus G4 software on Linux, the five applications “G4DevCfg”, “CreateSrcCfg”, “g4term”, “g4display” and “g4export” were available. For my project, all devices used must first be connected and configured with “G4DevCfg”. The terminal application “g4export” can be used to transmit the sensor data via UDP by specifying the previously created source configuration file, the local IP address of the receiver device and a port. The source configuration file is a file in which the position and orientation of the transmitter are defined by a “virtual frame of reference” and settings can be made for the entry hemisphere into the magnetic field, floor compensation and source calibration file. To run the application, the transmitters and the hub must be switched on at this point, the USB dongle must be connected to the laptop and the sensor to the hub, and the hub must be connected to the USB dongle. If the MacBook is now in the same network as the Linux laptop, the data can be received by specifying the previously used port. This is done with my sound installation in a self-created MaxMSP patch.

Figure 6

In this application, the appropriate port must first be selected on the left-hand side. As soon as the connection is established and the messages arrive, you can view them in raw form under the selection field. The six values that can be seen at the top in the middle of the application are the values for position and orientation that have been separated from the raw message. Final settings for the correct calibration can now be made in the action field below. There is also the option to mirror the axes individually or to change the Yaw value if unexpected problems should arise when setting up the sound installation. Once the values have been formatted into messages that can be used by Binauralix (visible at the bottom right of the application), they are sent to Binauralix.

The following videos provide a view of the scene in Binauralix and an auditory impression as the listener — driven by the sensor data — moves through the scene.

 

 

Past performances of the sound installation

The sound installation as a contribution to the EFFEKTE lecture series of the Wissenschaftsbüro-Karlsruhe

 

 

test
The sound installation as the subject of a workshop for the Kulturakademie at the HfM-Karlsruhe