Practical issues in recording for the Web

Alan Anderson, School of Arts and Social Sciences, School of Commerce and Management, Southern Cross University, Lismore, 2480. Email: alan.anderson@scu.edu.au

Introduction

Expanding on issues raised in Anderson (2007b)[HREF 1], this poster aims to demonstrate how a fundamental knowledge of professional audio production techniques can help educators to produce high quality audio content for the Web. The quality of equipment used is one determining factor. Equally important is learning how to best use the technology in various recording and postproduction situations. Beginning with sound capture, for instance, it is important understand that microphones have a polar pattern (pickup pattern). Specific applications are discussed in Anderson (2007b) whilst the following diagrams of microphone pick-up patterns (figures 1, 2 & 3) serve as a point of reference for this poster presentation.

Microphone pickup patterns

Three commonly used patterns are shown in Figures 1, 2 and 3 (diagrams adapted from Microphone Madness, 2007)[HREF 2].

Working too close to a microphone will cause significant changes in both level and tone as the singer/speaker moves his or her head (White, 2002, p. 35). Other issues such as using a pop shield and signal processing technology to reduce sibilance and enhance a speaker’s voice are mentioned in Anderson (2007b) [HREF 1] and will be demonstrated or discussed further in this presentation.

Preamps - the next essential link in the signal path

A knowledge of signal path and gain structure issues helps to correctly configure audio software and hardware for various applications. It also helps to troubleshoot if components in the signal path (from recording to playback) fail or need reconfiguring from time to time for different recording applications. After sound is picked-up by a microphone, the next link in the signal chain is the preamp. The “preamplifier (preamp) precedes another amplifier, and its function is to prepare an electronic signal for further amplification or processing” [HREF 3]. In order to amplify the output signal from most professional microphones, it is necessary to use either a suitable soundcard, an external preamp interface or a mixing console with its own built-in preamps, attenuators and phantom power for condenser (capacitor) microphones. In these contexts, the “function of a preamp is to amplify a low level signal (possibly at high impedance) to line-level”. Examples of portable external preamp interfaces include M-AUDIO’s™ Firewire 410 and Digidesign’s MBox II (USB) bundled with ProTools LE software [HREF 3]. If a number of musical instruments and vocals must be recorded simultaneously, the preamp interface ought to have enough input channels to allocate one or more to each sound source (i.e. line in from an electric instrument, vocal microphone and/or close-miked instrument). Consequently, a preamp such as a Mackie™ Onyx 800R 8-Channel Microphone Preamp or a larger mixing console with more inputs and outputs (i/o) would be necessary.

Virtual instruments and virtual signal routing

Creating multiple tracks for virtual instruments entirely in the software domain is simple by comparison. Garage Band, Logic Pro, CuBase (to name but a few programs) enable one to create music by loop-based composition and/or record live acoustic instruments and vocals on multiple tracks. Logic Pro™is touted by Apple as “the industry-leading application for music creation and audio production,” whilst Garage Band is essentially an entry-level program that incorporates special features for creating Podcasts [HREF 4]. Having created a Podcast with background music, GarageBand enables one to put the talk above the tunes with a dynamic “ducking” effect that automatically reduces music volume while the announcer (DJ) is speaking. Radio stations have long been configuring their mixing consoles and outboard hardware for ducking so this is not a new concept. In the modern software domain, using programs such as Protools or LogicPro, dynamic ducking can be achieved by virtual signal routing of a compressor plugin on the music track with side chain input from the speech track. The advantage that Garage Band brings to the new user is that ducking is engaged automatically by use of a simple drop down menu – thus negating the need to do any virtual or physical re-routing of the signal path. A Speech Enhancer can be engaged just as easily, to reduce distracting background noise and simulate a professional microphone using compression techniques to improve the user’s voice [HREF 4].

Not everyone aiming to record audio content for the Web will be interested in creating complex audio content involving music and voice. Therefore, with the various needs of users in mind, Figure 4 outlines options ranging from a simple record-by-phone service to the use of professional studios equipped for production of commercial CD release quality music and voice. Options are ranked from least expensive (1) to most expensive (6).

Fig 4. Audio recording for Webcast options ranked from least expensive (1) to most expensive (6)

Fix it in the mix

Awareness of what can or cannot be corrected or smoothed over by audio signal processing technology is useful for deciding whether a particular recorded performance or ‘take’ is of adequate quality to keep before moving on to record the next track. Roe (2006) notes it is important not to become complacent or careless adopting a ‘fix it in the mix’ mentality because contrary to popular belief, audio processing cannot always remove all extraneous noises on the fly or in postproduction [HREF 6]. Unwanted sounds that occupy the same frequency range as other wanted sounds can be particularly troublesome - even with the benefit of powerful multi-band compressors and equalization (EQ) tools that can isolate and filter out selected frequencies. The general consensus among audio engineers is “if you can get it right at the recording stage life becomes so much easier at mixdown” (Roe, 2006) [HREF 6]. In a live (real-time) VoIP situation, there is no opportunity to stop and do another ‘take,’ so care must be taken to do and say things right the first time.usWeb papers contain two types of references.

Audio compression (data)

It is common practice in professional audio circles to aim for the highest possible audio quality when recording. But recording at a higher quality level inevitably creates a large file as the analogue data is recorded into uncompressed digital format such as WAV or AIFF. The digital source file must consequently be encoded to compress the file into the appropriate sampling rate and bit depth to distribute audio over the Internet (Steward, 2007) [HREF 7] (Gerth, 2001) [HREF 8]. Gerth notes that an "encoding tool is used to select a codec based on the bandwidth target (connection speed) and the audio content (music or voice) to compress the original sound file and create a new audio file." If a large dynamic range of sound is required,16 bit resolution is recommended, whereas, if all sounds are about the same volume, 8 bit may be suitable (ibid). Sampling rate refers to the number of samples of sound taken per second whereas 'bit depth' defines the dynamic resolution of the audio. The more bits used to represent the sampled value, the more accurate the sample will be with respect to the original analog sound wave (ibid). For the purpose of comparison, it is worth noting that CD quality audio is 44,100 (44.10 kHz) samples per second, 16 bit, 2 channel (stereo), whereas telephone quality audio is 1 channel (mono) 8 bit at 11,025 samples per second (11.025 kHz) [HREF 7] [HREF 8]. One minute of CD quality sound requires around 10.5 MB of disk space whereas one minute of telephone audio quality would require as little as 0.6MB. Smaller file sizes facilitate faster download times from Web sites but the most economical options for keeping file sizes to a minimum invariably involve some trade-off in sound quality [HREF 8] (Beggs & Thede, 2007)[HREF 9]. Classical music, for instance, is characteristically wide in its range of timbres and dynamic range, and more likely to be adversely affected by lower bit resolution than speech (see Figures 5 & 6) [HREF 8] [HREF 10]. Likewise, full-bodied vocal requires 16 bit resolution [HREF 8].

Fig 5. Storage requirements for one minute of audio (Gerth, 2001) [HREF 8]

Fig 6. Streaming audio (PC Magazine, 1999 in Gerth, 2001) [HREF 8]

Clearly, the appropriateness of particular formats and compression settings must therefore be weighed carefully in relation to the nature of the audio content and the purpose of including it in the Web site or Webcast.

Recommendations for good quality audio production

• It is important to strive for the best possible recording of the best possible performance, rather than become over-reliant on corrective editing to fix or enhancing a mediocre performance in the mix.
• Always to start with the best quality master copy obtainable in an uncompressed format.
• Equipment must be of appropriate quality and specifications for the particular recording or postproduction application
• A basic knowledge of signal flow and gain structure is essential to optimize configuration and level settings for hardware and software, and to avoid unwanted distortion.
• Appropriate microphone technique and knowledge of pick up patterns is important to optimize recording of wanted sounds and rejection of unwanted sound (including extraneous noises from mishandling conductive components of microphones, stands and cables).
• Correct handling and storage of equipment is important to avoid malfunctions of audio equipment.
• Levels must be turned down before making or breaking cable connections while equipment is active.
• Equipment such as microphones and speakers in particular should be correctly position to avoid generating unwanted feedback loops.
• Audio effects and dynamics processing must be used judiciously to maintain clarity and appropriate levels for distribution and playback over a variety of mediums and systems.
• Advanced editing and signal processing techniques can be used to eliminate or significantly reduce pops, hisses or abrupt finishes.
• Automated ducker technology should be used to create professional segues from voice to music or voice-over lower level music backing tracks (especially in live Webcast situations whereby it is not easy to control volume faders on several tracks at once).
• In efforts to reduce file size, care must be taken to use appropriate media formats and avoid excessive compression that can result in a loss of depth, dynamics and tonal colour).

Conclusion

Expanding on issues raised in Anderson (2007b) [HREF 1], this poster presentation is designed to demonstrate how a fundamental knowledge of professional audio production techniques can help educators to produce high quality audio content for the Web. These skills are directly applicable to Voice over Internet Protocol (VoIP) applications and hybrid VoIP and Web-based learning environments such as Elluminate.

References

White, P. 2002, Recording and Production Techniques, Sanctuary Publishing Limited, London, UK.

Hypertext References

HREF 1
http://ausweb.scu.edu.au/aw07/papers/refereed/anderson3/paper.html

HREF 2
http://www.microphonemadness.com/faq.htm#whatelse

HREF 3]
http://www.tweakheadz.com/microphone_preamps.htm

HREF 4
http://www.apple.com/ilife/garageband/

HREF 5
http://freesound.iua.upf.edu/tagsViewSingle.php?id=36_6_117_27

HREF 6
http://www.mixbuss.com/article.php?id=32

HREF 7
http://ezinearticles.com/?Adding-Sound-To-Your-Web-Site---The-Good,-The-Bad-And-The-Ugly&id=22528

HREF 8
http://64.233.179.104/scholar?hl=en&lr=&safe=off&q=cache:1MloFpiKq8kJ:www.coe.uh.edu/courses/cuin7330/class7/audio-on-the-web.pdf+poor+audio+flash+audio+on+classical+music

HREF 9
http://www.intranetjournal.com/articles/200102/psm1a.html

HREF 10
http://musicalassumptions.blogspot.com/2007/05/high-fidelity-death-of-classical-music.html

Copyright