DOCSLIB.ORG

Auto-Adaptive Resonance Equalization Using Dilated Residual Networks

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Auto-Adaptive Resonance Equalization Using Dilated Residual Networks AUTO-ADAPTIVE RESONANCE EQUALIZATION USING DILATED RESIDUAL NETWORKS Maarten Grachten Emmanuel Deruty Alexandre Tanguy Contractor for Sony CSL Paris, France Sony CSL Paris, France Yascore, Paris, France ABSTRACT The use of machine learning, in particular neural net- works, to solve audio production related tasks is recent. In music and audio production, attenuation of spectral res- Automatic mixing tasks that have been addressed in this onances is an important step towards a technically correct way include automatic reverbation [6], dynamic range result. In this paper we present a two-component system compression [18], and demixing/remixing of tracks [17]. to automate the task of resonance equalization. The first To our knowledge, there is no documented example of the component is a dynamic equalizer that automatically de- use of neural networks for automatic equalization. tects resonances, to be attenuated by a user-specified fac- A specific form of equalization used both in mixing and tor. The second component is a deep neural network that mastering is the attenuation of resonating or salient fre- predicts the optimal attenuation factor based on the win- quencies, i.e. frequencies that are substantially louder than dowed audio. The network is trained and validated on em- their neighbors [2]. Salient frequencies may originate from pirical data gathered from a listening experiment. We test different phenomena, such as the acoustic resonances of a two distinct network architectures for the predictive model physical instrument or an acoustic space. They are consid- and find that an agnostic network architecture operating ered a deficiency in the sense that they may mask the con- directly on the audio signal is on a par with a network tent of other frequency regions. One particular difficulty architecture that relies on hand-designed features. Both in resonance attenuation (RA) is finding the right amount architectures significantly improve a baseline approach to of attenuation. For example, too much attenuation may un- predicting human-preferred resonance attenuation factors. mask noise that would otherwise remain unheard, or flatten the spectrum to the point of garbling the original audio. 1. INTRODUCTION AND RELATED WORK The subject of this paper is the automation of the RA process using machine learning. We limit our study to Equalization is part of the audio mixing and mastering neural networks as the state of the art machine learning process. It is a redistribution of the energy of the signal technique. Our method fully automates the RA process. It in different frequency bands. The process has been tra- includes 1) a 0.5s windowed RA process that can be con- ditionally performed by skilled sound engineers or mu- trolled with a single parameter—the resonance attenuation sicians who determine the proper equalization given the factor (RAF), 2) a deep neural network that predicts the at- characteristics of the input audio. Recently methods have tenuation factor from the input audio, making the process been developed for semi-automatic and automatic equal- auto-adaptive [21]. ization. These methods include automatic detection of fre- For the training and validation we conduct a listen- quency resonances [1], equalization derived from expert ing experiment determining optimal RAFs for a set of practices [7], and conformation to a target spectrum [15]. tracks, as chosen by sound engineers. We compare a neu- Equalization profiles may also be derived from semantic ral network architecture that operates directly on the au- descriptors [5]. Appropriate equalization settings can be dio signal to a more traditional approach that includes a found through different means, for example by compar- feature-extraction stage yielding a set of features com- ing the input source to previously equalized content [20], monly used in MIR. Results show that both approaches or by formulating equalization as an optimization problem perform equally well, and significantly outperform a base- where inter-track masking is used as the cost function [10]. line. Some automated equalization functionalities are featured The paper is organized as follows. Section 2 describes in commercial products 1 2 . the RA process. The listening experiment is described in Section 3. The design, training, and evaluation of the pre- 1 www.izotope.com/en/products/mix/neutron 2 www.soundtheory.com/home dictive models is presented in Section 4, and conclusions are presented in Section 5. c Maarten Grachten, Emmanuel Deruty, Alexandre Tan- guy. Licensed under a Creative Commons Attribution 4.0 International 2. RESONANCE EQUALIZATION License (CC BY 4.0). Attribution: Maarten Grachten, Emmanuel Deruty, Alexandre Tanguy. “Auto-adaptive Resonance Equalization us- Traditionally RA has been a manual task where a sound ing Dilated Residual Networks”, 20th International Society for Music In- engineer determines the resonating frequencies by ear or formation Retrieval Conference, Delft, The Netherlands, 2019. using a graphical tool, in order to reduce the energy of the 405 Proceedings of the 20th ISMIR Conference, Delft, Netherlands, November 4-8, 2019 s Corrected PCM Signal a PCM Signal the experimental design of the listening test it proved un- practical to ask subjects to set a varying RAF. Therefore we r inverse DFT b DFT chose relatively homogeneous music fragments (excluding transitions between different sections of songs) and asked q Corrected c Power Spectrum p Power Spectrum subjects for a single attenuation factor for the whole frag- ment, ensuring relatively homogeneous sound fragments. n d Equal Loud- o Spectrum Log-to-linear/freq- ness Weighting Correction interpolation 3.1 Participants e ELC-weighted m Power Spec- l Attenuation A group of 15 subjects was recruited for the experiment, Power Spectrum trum Correction Factor around Paris (France). All subjects are recognized pro- k ∗ fessionals in the industry. Nine subjects specialize in stu- f Smoothing j − dio recording (Classic/Jazz/Pop/Rock/Movie Music, audio g post-production), three are experts in live music, and three Smoothed h Positive i Resonance Power Spectrum Difference are composer/music producers. The average age was 32 (min: 24, max: 42). The subjects were recruited and paid Figure 1. Resonance equalization block diagram; White as if they were working on a commercial project. and gray blocks represent data and processes respectively; The green block depicts the single user-controlled param- 3.2 Data eter; The symbols , ∗ and − represent elementwise vec- A set of 150 audio tracks was used for the listening exper- tor/vector multiplication, elementwise scalar/vector multi- iment. The tracks are excerpts from longer pieces, with a plication, and unary negation respectively. mean duration of 46 seconds and a standard deviation of 16 seconds. All tracks were processed using Nugen AMB R128 4 so that they were aligned to the same median loud- signal in those frequencies by an appropriate amount [16]. ness. The set comprised contemporary pop and rock mu- In this section, we describe a procedure that identifies res- sic, as well as film scores. Of this set, 131 tracks were onating frequencies autonomously, and reduces the energy unique recordings, while the remaining 19 tracks were in those frequencies by a factor that is controlled by the variants of some of the unique 131 recordings, with dif- user. The procedure works on overlapping audio windows ferences in mixing. None of the tracks were previously that must be large enough to allow for spectral analysis at mastered. a high frequency resolution. Figure 1 displays a block diagram of the RA process, 3.3 Procedure where each element is denoted by a letter. We will use these letters to refer to the corresponding elements in the The listening experiment took place in a recording studio, diagram. First the audio signal is used to compute a where participants listened to the audio tracks individually, power spectrum weighted by Equal-Loudness Contours using studio monitors, at a measured loudness of 80 dBC– (ELC) [12] at a fixed monitoring level of 80 phon (Fig- a typical listening loudness during audio production. The ure 1, element d) to reflect the perceptual salience of the participants were presented with a web interface in which signal energy at different frequency bands. The value of they could listen to each track with different degrees of 80 phon is chosen in relation to the procedure detailed in RA, ranging from 0 (no attenuation) to 1 in 17 steps. They Section 3.3. The ELC-weighted power spectrum (e) con- could select their preferred degree of RA, or alternatively sists of 400 log-scaled frequency bands. decline to select any version, indicating that none of the Resonances (i) are determined by smoothing ELC- versions sounded acceptable. The tracks were separated weighted power spectrum 3 (e) to obtain (g) and comput- from each other by 10 seconds of pink noise surrounded ing the elementwise differences (e) minus (g), setting nega- by a short silence to give the participants a fixed reference. tive elements to zero (h). The negative of the resonances is Sessions of 50 tracks were alternated with breaks. then scaled by the user defined RAF (l), transformed back to a linear scale and converted back to the shape of the 3.4 Results and discussion original spectrum using interpolation (h). The result (o) is a vector of scaling factors (one per DFT bin). Multiplying Basic statistics of the results per subject are given in Ta- the original spectrum (c) with the scaling factors gives the ble 1. Subject 13 stands out because of the number of corrected spectrum (q) which—by the inverse DFT (r)— missing ratings (21 versus a median of 1 over all subjects). yields the corrected audio signal (s). Subjects 1 and 15 have abnormally high rates of 0.0 rat- ings (72 and 58 respectively, versus a median of 16 over all subjects). Finally, Subject 7 stands out in terms of median 3.
Load more

Recommended publications
  • Learning to Build Natural Audio Production Interfaces
    arts Article Learning to Build Natural Audio Production Interfaces Bryan Pardo 1,*, Mark Cartwright 2 , Prem Seetharaman 1 and Bongjun Kim 1 1 Department of Computer Science, McCormick School of Engineering, Northwestern University, Evanston, IL 60208, USA 2 Department of Music and Performing Arts Professions, Steinhardt School of Culture, Education, and Human Development, New York University, New York, NY 10003, USA * Correspondence: [email protected] Received: 11 July 2019; Accepted: 20 August 2019; Published: 29 August 2019 Abstract: Improving audio production tools provides a great opportunity for meaningful enhancement of creative activities due to the disconnect between existing tools and the conceptual frameworks within which many people work. In our work, we focus on bridging the gap between the intentions of both amateur and professional musicians and the audio manipulation tools available through software. Rather than force nonintuitive interactions, or remove control altogether, we reframe the controls to work within the interaction paradigms identified by research done on how audio engineers and musicians communicate auditory concepts to each other: evaluative feedback, natural language, vocal imitation, and exploration. In this article, we provide an overview of our research on building audio production tools, such as mixers and equalizers, to support these kinds of interactions. We describe the learning algorithms, design approaches, and software that support these interaction paradigms in the context of music and audio production. We also discuss the strengths and weaknesses of the interaction approach we describe in comparison with existing control paradigms. Keywords: music; audio; creativity support; machine learning; human computer interaction 1. Introduction In recent years, the roles of producer, engineer, composer, and performer have merged for many forms of music (Moorefield 2010).
    [Show full text]
  • DMA Recording Project Guidelines (Fall 2011) Page 1 of 2
    DMA – TMUS 8329 Major Project (4–6 cr.)* Guidelines for Recording Project In addition to the normal requirements of a recital, it is expected that the student will become involved in all aspects of the recording, in effect acting as producer from start to finish. Before undertaking the Recording Project: the student must submit a prospectus that is reviewed and approved by the faculty advisory committee and the Associate Dean of Graduate Studies. A copy of the prospectus must be kept in the student’s file. For a recording to fulfill the requirement, it must adhere to the following criteria. 1. Be comparable in length to a recital. 2. The student is responsible for coordinating all matters pertaining to the recording including: contracting of musicians, studio manager and recording engineer; CD printing or duplication, graphic art design layout, and recording label (optional). 3. The recording must be unique in some way that sets it apart from other recordings. Examples can include but are not limited to recordings that feature: original compositions and/or arrangements, collections of works that are less known and/or have not been readily available in recordings, and so on. This aspect of the project should reflect the student’s creativity and research skills. 4. The submitted recording must be of professional quality. 5. Post-recording work is an essential aspect of the project. The student must oversee the editing, mixing, and (optionally) the mastering of the recording. Refer to the definitions of these processes at the bottom of this document. a. Editing (if necessary) b. Mixing c.
    [Show full text]
  • “Knowing Is Seeing”: the Digital Audio Workstation and the Visualization of Sound
    “KNOWING IS SEEING”: THE DIGITAL AUDIO WORKSTATION AND THE VISUALIZATION OF SOUND IAN MACCHIUSI A DISSERTATION SUBMITTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY GRADUATE PROGRAM IN MUSIC YORK UNIVERSITY TORONTO, ONTARIO September 2017 © Ian Macchiusi, 2017 ii Abstract The computer’s visual representation of sound has revolutionized the creation of music through the interface of the Digital Audio Workstation software (DAW). With the rise of DAW- based composition in popular music styles, many artists’ sole experience of musical creation is through the computer screen. I assert that the particular sonic visualizations of the DAW propagate certain assumptions about music, influencing aesthetics and adding new visually- based parameters to the creative process. I believe many of these new parameters are greatly indebted to the visual structures, interactional dictates and standardizations (such as the office metaphor depicted by operating systems such as Apple’s OS and Microsoft’s Windows) of the Graphical User Interface (GUI). Whether manipulating text, video or audio, a user’s interaction with the GUI is usually structured in the same manner—clicking on windows, icons and menus with a mouse-driven cursor. Focussing on the dialogs from the Reddit communities of Making hip-hop and EDM production, DAW user manuals, as well as interface design guidebooks, this dissertation will address the ways these visualizations and methods of working affect the workflow, composition style and musical conceptions of DAW-based producers. iii Dedication To Ba, Dadas and Mary, for all your love and support. iv Table of Contents Abstract ..................................................................................................................
    [Show full text]
  • Thesis Table of Contents
    A new sound mixing framework for enhanced emotive sound design within contemporary moving-picture audio production and post-production. Volume 1 of 3 Neil Martin Georges Hillman PhD University of York Theatre, Film and Television September 2017 2 Abstract This study comprises of an investigation into the relationship between the creative process of mixing moving-picture soundtracks and the emotions elicited by the final film. As research shows that listeners are able to infer a speaker’s emotion from auditory cues, independently from the meaning of the words uttered, it is possible that moving-picture soundtracks may be designed in such a way as to intentionally influence the emotional state and attitude of its listening-viewers, independently from the story and visuals of the film. This study sets out to determine whether certain aspects of audience emotions can be enhanced through specific ways of mix-balancing the soundtrack of a moving-picture production, primarily to intensify the viewing experience. Central to this thesis is the proposal that within a film soundtrack there are four distinct ‘sound areas’, described as the Narrative, Abstract, Temporal and Spatial; and these form a useful framework for both the consideration and the creation of emotional sound design. This research work evaluates to what extent the exploration of the Narrative, Abstract, Temporal and Spatial sound areas offers a new and useful framework for academics to better understand, and more easily communicate, emotive sound design theory and analysis; whilst providing practitioners with a framework to explore a new sound design approach within the bounds of contemporary workflow and methodology, to encourage an enhanced emotional engagement by the audience to the soundtrack.
    [Show full text]
  • Understanding Audio Production Practices of People with Vision Impairments
    Understanding Audio Production Practices of People with Vision Impairments Abir Saha Anne Marie Piper Northwestern University University of California, Irvine Evanston, IL, USA Irvine, CA, USA [email protected] [email protected] ABSTRACT of audio content creation, including music, podcasts, audio drama, The advent of digital audio workstations and other digital audio radio shows, sound art and so on. In modern times, audio content tools has brought a critical shift in the audio industry by empower- creation has increasingly become computer-supported – digital in- ing amateur and professional audio content creators with the nec- struments are used to replicate sounds of physical instruments (e.g., essary means to produce high quality audio content. Yet, we know guitars, drums, etc.) with high-fdelity. Likewise, editing, mixing, little about the accessibility of widely used audio production tools and mastering tasks are also mediated through the use of digital for people with vision impairments. Through interviews with 18 audio workstations (DAWs) and efects plugins (e.g., compression, audio professionals and hobbyists with vision impairments, we fnd equalization, and reverb). This computer-aided work practice is that accessible audio production involves: piecing together accessi- supported by a number of commercially developed DAWs, such 1 2 3 ble and efcient workfows through a combination of mainstream as Pro Tools , Logic Pro and REAPER . In addition to these com- and custom tools; achieving professional competency through a mercial eforts, academic researchers have also invested signifcant steep learning curve in which domain knowledge and accessibility attention towards developing new digital tools to support audio are inseparable; and facilitating learning and creating access by production tasks (e.g., automated editing and mixing) [29, 57, 61].
    [Show full text]
  • Recording: Processing Audio and the Modern Recording Studio
    Chapter 7. Meeting 7, Recording: Processing Audio and the Modern Recording Studio 7.1. Announcements • Quiz next Thursday • Numerous listenings assignments for next week 7.2. Processing Audio • Contemporary processors take many physical forms: effects units, stomp-boxes © source unknown. All rights reserved. This content is excluded from our Creative Commons license. For more information, see http://ocw.mit.edu/fairuse. 160 Photo courtesy of kernelslacker on Flickr. 161 Photo courtesy of michael morel on Flickr. Courtesy of George Massenburg Labs. Used with permission. 162 Original photo courtesy of eyeliam on Flickr; edited by Wikipedia User:Shoulder-synth. 163 • As software, most are implemented as plug-ins 164 © Avid Technology, Inc. All rights reserved. This content is excluded from our Creative Commons license. For more information, see http://ocw.mit.edu/fairuse. 165 © MOTU, Inc. All rights reserved. This content is excluded from our Creative Commons license. For more information, see http://ocw.mit.edu/fairuse. 7.3. Distortion • Pushing a signal beyond its dynamic range squares the waveform • Making round signals more square adds extra harmonics [demo/processorsDistortion.pd] 166 • Examples • Overdrive • Fuzz • Crunch 7.4. Dynamics Processors • Transform the amplitude of a signal in real-time • Amplitudes can be pushed down above or below a threshold to decrease or increase dynamic range • Examples 167 • Compressors and Limiters • Expanders and Gates 7.5. Dynamics Processors: Compression • Reduces a signal’s dynamic range • Makes the quiet sounds louder • Helps a track maintain its position in the mix • Two steps • Reduce dynamic range: turn amplitudes down if a above a specific level (the threshold) • Increase amplitude of entire signal so that new peaks are where the old were 168 To be compressed Uncompressed Peak Threshold Sound Energy Time Compression occurs Previous 0 VU Threshold Sound Energy Time 0 VU Sound Energy Boost overall level Time Figure by MIT OpenCourseWare.
    [Show full text]
  • Approaches in Intelligent Music Production
    arts Article Approaches in Intelligent Music Production David Moffat * and Mark B. Sandler School of Electronic Engineering and Computer Science, Queen Mary University of London, London E1 4NS, UK * Correspondence: [email protected]; Tel.: +44-(0)20-7882-5555 Received: 18 July 2019; Accepted: 10 September 2019; Published: 25 September 2019 Abstract: Music production technology has made few advancements over the past few decades. State-of-the-art approaches are based on traditional studio paradigms with new developments primarily focusing on digital modelling of analog equipment. Intelligent music production (IMP) is the approach of introducing some level of artificial intelligence into the space of music production, which has the ability to change the field considerably. There are a multitude of methods that intelligent systems can employ to analyse, interact with, and modify audio. Some systems interact and collaborate with human mix engineers, while others are purely black box autonomous systems, which are uninterpretable and challenging to work with. This article outlines a number of key decisions that need to be considered while producing an intelligent music production system, and identifies some of the assumptions and constraints of each of the various approaches. One of the key aspects to consider in any IMP system is how an individual will interact with the system, and to what extent they can consistently use any IMP tools. The other key aspects are how the target or goal of the system is created and defined, and the manner in which the system directly interacts with audio. The potential for IMP systems to produce new and interesting approaches for analysing and manipulating audio, both for the intended application and creative misappropriation, is considerable.
    [Show full text]
  • Digital Transmit Audio Processing with Presonus Studio One Artist
    Digital transmit audio processing with Presonus Studio One Artist ... Setup guide for a software based digital audio workstation (DAW) as an alternative to conventional hardware based audio processing techniques. Appendix A provides additional info for items high lighted in yellow … How this scheme works: A. Digital audio output from the Rode NT-USB microphone is routed to the computer via a type A/B USB cable. B. Digital audio is then processed within the computer by the software based system Presonus Studio One Artist. C. Digital transmit audio output from Studio One Artist is then routed via Type A/B USB cable to the Kenwood TS- 590 via a USB cable. Finally, it is converted from digital to analog and transmitted. D. 90 watts RF output from TS-590S transceiver. E. 800 watts RF output from Ameritron AL80B linear. F. 800 watts RF output from Palstar AT1500BAL to horizontal loop. G. Receive audio from the Kenwood 590S is routed via USB cable (C.) back to the computer and outputted to a pair of near field monitor speakers. Setup: Hardware: Kenwood TS-590S PC running Window’s 10 with internal Realtek sound card. Necessary interconnection cabling. Presonus Studio One Artist Digital Audio Workstation (DAW) software: Your best bet is to purchase Presonus AudioBox USB audio interface as described at http://www.sweetwater.com/store/detail/AudioBoxUSB . This interface includes a free copy of Studio One Artist. Note: Alternatively, you can use the AudioBox USB interface together with conventional analog cables for transceivers that are not equipped with USB connectivity. Kenwood 590 settings: 61 - 115200 62 - 115200 63 - USB 64 - 1 65 - 2 68 - off 69 - on (assuming you prefer to use VOX) 70 - 20 71 - 3 79 - 205 (for ptt via front panel "PFA' button) Note: Other Kenwood 590 settings include menu 25/27 at 100, menu 26/28 at 2900.
    [Show full text]
  • Advanced Automatic Mixing Tools for Music Perez Gonzalez, Enrique
    Advanced automatic mixing tools for music Perez Gonzalez, Enrique The copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without the prior written consent of the author For additional information about this publication click this link. https://qmro.qmul.ac.uk/jspui/handle/123456789/614 Information about this research object was correct at the time of download; we occasionally make corrections to records, please therefore check the published record when citing. For more information contact [email protected] Advanced Automatic Mixing Tools for Music Submitted by Enrique Perez Gonzalez For the Ph.D. degree of Queen Mary University Of London Mile End Road London E1 4NS September 30, 2010 2 I certify that this thesis, and the research to which it refers, are the product of our own work, and that any ideas or quotations from the work of other people, published or otherwise, are fully acknowledged in accordance with the standard referencing practices of the discipline. I acknowledge the helpful guidance and support of our supervisor, Dr. Johua Daniel Reiss. 3 Abstract This thesis presents research on several independent systems that when combined together can generate an automatic sound mix out of an unknown set of multi‐channel inputs. The research explores the possibility of reproducing the mixing decisions of a skilled audio engineer with minimal or no human interaction. The research is restricted to non‐time varying mixes for large room acoustics. This research has applications in dynamic sound music concerts, remote mixing, recording and postproduction as well as live mixing for interactive scenes.
    [Show full text]
  • Dynamic Range Compression and the Semantic Descriptor Aggressive
    applied sciences Article Dynamic Range Compression and the Semantic Descriptor Aggressive Austin Moore Centre for Audio and Psychoacoustic Engineering, School of Computing and Engineering, University of Huddersfield, Huddersfield HD1 3DH, UK; [email protected] Received: 29 January 2020; Accepted: 12 March 2020; Published: 30 March 2020 Featured Application: The current study will be of interest to designers of professional audio software and hardware devices as it will allow them to design their tools to increase or diminish the sonic character discussed in the paper. In addition, it will benefit professional audio engineers due to its potential to speed up their workflow. Abstract: In popular music productions, the lead vocal is often the main focus of the mix and engineers will work to impart creative colouration onto this source. This paper conducts listening experiments to test if there is a correlation between perceived distortion and the descriptor “aggressive”, which is often used to describe the sonic signature of Universal Audio 1176, a much-used dynamic range compressor in professional music production. The results from this study show compression settings that impart audible distortion are perceived as aggressive by the listener, and there is a strong correlation between the subjective listener scores for distorted and aggressive. Additionally, it was shown there is a strong correlation between compression settings rated with high aggressive scores and the audio feature roughness. Keywords: dynamic range compression; music production; semantic audio; audio mixing; 1176 compressor; FET compression; listening experiment 1. Introduction 1.1. Background In addition to general dynamic range control, it is common for music producers to use dynamic range compression (DRC) for colouration and non-linear signal processing techniques, specifically to impart distortion onto program material.
    [Show full text]
  • Touchmix Application Guide for Musicians and Bands
    Application Guide for Musicians & Bands TOUCHMIX APPLICATION GUIDE FOR MUSICIANS AND BANDS Do you need TouchMix? • Are you a performing musician? • Would you like your live shows to sound like a professionally mixed concert? • Would you like to have multi-track recordings of your shows? • Do you and your band-mates struggle with hearing yourselves on stage? • Do you have problems with feedback during performances? • Are you a better musician than a sound engineer? If you are answer yes to most of these questions, you’ve come to the right place because the QSC TouchMix series mixers have a solution for you. WHAT’S IN THIS GUIDE There are two sections to this guide. The first section will explain how to perform some tasks that are common to almost any live performance, whether you’re a solo performer or a ten-piece band. Topics include mixer selection, setups and presets, features. The second part of the guide demonstrates several real-world applications of TouchMix mixers, ranging from solo performer gigs to club concerts. Which TouchMix is right for you? In the entire history of audio mixing consoles, nobody has ever been sorry they got a mixer with a few more inputs and outputs then they needed. So it’s probably a good idea to consider your current needs and then give yourself a little room to grow. The TouchMix digital mixer series—Top left: TouchMix-16; Top right: TouchMix-8; Bottom: TouchMix-30 Pro 2 TouchMix Compact Digital Mixers PRE-FLIGHT: SETTING UP YOUR TOUCHMIX AT HOME TO SAVE TIME AT THE GIG.
    [Show full text]
  • An Autonomous Method for Multi-Track Dynamic Range Compression Templates for Dafx-08, Finland, France
    Proc. of the 15th Int. Conference on Digital Audio Effects (DAFx-12), York, UK, September 17-21, 2012 AN AUTONOMOUS METHOD FOR MULTI-TRACK DYNAMIC RANGE COMPRESSION TEMPLATES FOR DAFX-08, FINLAND, FRANCE Jacob A. Maddams, Saoirse Finn, Joshua D. Reiss Centre for Digital Music, Queen Mary University of London London, UK [email protected] ABSTRACT DRC has many applications when it comes to mixing multi- track audio. Typical examples include controlling the transient Dynamic range compression is a nonlinear audio effect that re- attack of percussive instruments such as drums, raising the over- duces the dynamic range of a signal and is frequently used as part all loudness of a sound source by applying compression with of the process of mixing multi-track audio recordings. A system make-up gain and providing a more consistent signal level [7]. for automatically setting the parameters of multiple dynamic The purpose of this paper is to develop and evaluate a system range compressors (one acting on each track of the multi-track which, using high level, perceptual audio features, automatically mix) is described. The perceptual signal features loudness and and cross-adaptively applies DRC to individual tracks within a loudness range are used to cross-adaptively control each com- multi-track mix. pressor. The system is fully autonomous and includes six differ- ent modes of operation. These were compared and evaluated against a mix in which compressor settings were chosen by an 2. DYNAMIC RANGE COMPRESSION expert audio mix engineer. Clear preferences were established for the different modes of operation, and it was found that the auton- DRC is a nonlinear audio effect that narrows the dynamic range omous system was capable of producing audio mixes of approx- (i.e.
    [Show full text]