Wonderfuls are a local Brisbane band who recently headlined The Bearded Lady in Brisbane, celebrating the launch of their latest record. I was recently asked to mix their performance, featuring three of the group's members in an atypical performance, for radio play on 4ZZZ. Follow my process here...
Robert Vagg and Danny McGirr are the two masterminds behind Brisbane group Wonderfuls, making and releasing music under this moniker for almost a decade with Robert as the group's vocalist, and Danny on guitar / synth. Though recently they have added the "cloud-bursting" tenor sax of Tom Watkins and the "lithely imaginative" drumming of Donat Tahiraj.
Together they explore the darkest depths of hopelessness and isolation, playing what has been described as "sad downer pop: glazed with a kinda dreamy autumnal patina this is discomforting and broken" (Prescott, 2012). Instrumental intermissions create atmospheric whitewashing between song clusters, ultimately turning despair into a form of deeply moving fine art. Check out their Triple J Unearthed page for a better idea of their eclectic sound!
For their recent half-hour performance, I was given three mono tracks direct from the venue's desk:
Vox
Guitar
Sax (DI)
I also received a stereo recording of the room taken from a Zoom H6 with the X-Y mic attachment.
The first step after importing these tracks to my ProTools session, was to listen to each of the tracks and remove any unwanted pops, clicks, or other unfavourable noises. Throughout the vocal track I also had to pay close attention for any bad language, as the final mix may be played on other radio shows around the country. If the bad language featured during the lyrics, I would apply a 1kHz sine wave over the vowel sound in the word. Though on this particular occasion the swearing occurred in between songs, whilst talking with the crowd. Consequently simply adjusting the clip gain during the word sufficed.
VOX
Whilst going through the vocal track, I also noticed a few occasions where Robert had moved the mic away from his mouth. To compensate I carefully adjusted the clip gain, sometimes having to separate the track and appropriately crossfade, to avoid any pops from the waveform being manipulated.
I noticed that he sang with quite a deep voice, and at times it could be difficult to understand. By adding an EQ band with a reasonably tight Q and increasing its gain excessively, I was able to sweep through the frequency range to find the point that provided some intelligibility, in this case at 4.3kHz. I also noticed during the sweep that a 10.82kHz boost helped to cut through the mix, adding some sparkle to his voice. Finally I cut the lower frequencies by adding a HPF at around 80Hz.
I quickly realised that automating his vocal level for the duration of the 35 minute performance would be an arduous task. As a solution I discovered a useful plugin created by Waves: Vocal Rider. This wonderful piece of software writes automation data for its own level fader directly into your DAW, without adding any colouration (Waves, 2019). The white fader featured in the centre is the volume control for the incoming signal. It will move automatically to create a consistent level (equivalent to 'riding the fader') with respect to the Target level set at the top (Waves Audio, 2017).
The plugin is capable of differentiating vocals from background noise, by adjusting the Vocal Sensitivity knob. This is, in essence, a threshold that tells Vocal Rider to ignore sounds below a certain level; a Ride/Idle activity LED lights when an above‑threshold vocal has been detected (White, 2010). In a SoundOnSound review, Paul writes "Vocal Rider also includes Waves' own breath‑detection algorithm, so that the plug‑in doesn't turn up all the breath sounds."
After the vocal sensitivity control is set such that vocals are recognised and spill is ignored, the next step is to adjust the Music Sensitivity knob. This feature enables Vocal Rider to react to the backing mix level, so if, for example, the song gets louder in the chorus, the vocal will also be lifted so that it doesn't get buried (White, 2010). For this a sensitivity LED monitors the response to the side‑chain input.
Another important control is the double‑headed Range slider, which limits the amount by which the vocal can be turned up or down. Finally, on the other side is the Output gain fader and its associated clip LED, enabling you to adjust the output level to your satisfaction.
By following this process I was able to automate the vocal level to a precise level of consistency, cutting down significant time during the mixing process. After writing the automation to my track, there were only a few occasions where I decided to manually adjust the automation curve to compliment the singer's intent and suite the feel of the song at the time.
After adding a de-esser into the processing chain to remove unwanted sibilance, I found that the vocal was still sounding a little dull, lacking in the higher frequencies, and not shining through the mix as I would have liked. I decided to teach myself a new production technique and try a new approach to correcting this.
Whilst purchasing the Vocal Rider, I also spotted a software replica of the Studer J37 Tape Machine, something of great significance if you know your recording history (Sgt Pepper's et al.). The J37 tape saturation plugin is supposed to be a precision model of this historic machine, faithfully recreating its inimitable sonic signature (Waves, 2019).
In addition to this, the J37 plugin also features the option of three distinctive tape formulas, developed by EMI for exclusive use in its studios (Waves & Abbey Road, 2019):
888: Considered “lo-fi” and “grainier” compared to the other formulas, with more distortion between 1 kHz and 8 kHz.
811: Offers better high frequency response and slightly less distortion than the 888 formula
815: Delivers flatter high frequency response and less distortion than the 811. Recommended when minimal coloration is desired.
For the purposes of slightly colouring Robert's voice, I selected 811.
The next user-option refers to the speed at which the tape moves across the heads; 7.5 ips or 15 ips (inches per second). The user manual (2019) describes 7.5 to deliver "somewhat tighter low frequencies at the price of reduced high frequency". Consequently, I chose 15, which is recommended for material that requires more sensitive handling of high frequency content.
Bias is something that I had often read engineers would "wind up" or adjust, but never had I truly grasped the concept. With a little investigation I learned that it is in fact a special signal that is applied during the recording phase. The first tape recorders simply applied the raw audio signal to the electromagnet in the head. It worked, but it produced a lot of distortion on low-frequency sounds (Sweetwater, 2007). As a solution, a bias signal refers to the inaudible ultrasonic high frequency tone (usually a sine wave ranging from 40 - 100 kHz) that is added to the audio signal to improve the fidelity of analog magnetic recordings. Tape has a nonlinear response at low signal strengths, so the bias moves the signal being recorded up into the linear portion of the tape's magnetization curve. This movement means that the tape reproduces the sound recorded on it more faithfully (Brain, 2008).
The plugin features options for adjusting the Wow and Flutter. Wow refers to frequency modulation caused by speed irregularities of the tape machine motor. Where as Flutter refers to amplitude modulation caused by movement of the tape in relation to the machine’s recording heads (Waves & Abbey Road, 2019). Since I wanted to keep the vocal in great condition, I didn't touch either of these parameters.
Finally, I used a touch of saturation to introduce some more harmonic content, and ever-so-slightly added some slap delay with a high-pass filter preventing all of the low frequency content from being repeated and potentially muddying up the mix. The vocal was also sent tastefully via an aux send to a reverb (discussed in more detail for the guitar).
It seemed the vocal was sitting where I wanted it to at times throughout the recording, but I thought it could do with some parallel compression (otherwise known as New York compression) to deal with the large dynamic range of the live performance. This technique involved sending the vocal track to an auxiliary return channel featuring heavy compression. On this track (labelled 'NYC' above) I also included an EQ prior to the compressor, to remove the low-frequency content, but predominantly to place a shelf boost at around 4kHz to really bring out the presence of his voice. By mixing the vocal and NYC tracks together tastefully, I found a sweet spot which enabled the dynamic vocal track to take charge during intense moments of the Wonderful's performance, whilst allowing the auxiliary track to come through during sotto voce moments and maintain a consistent vocal level.
GUITAR
Initially, the biggest challenge I recognised with mixing this performance, was creating a wide stereo representation with three mono tracks. Particularly as I wanted the vocal to sit central, and the saxophone only appeared intermittently throughout the performance, complimenting the vocal. Therefore it was up to the guitar to really fill the stereo field. Thankfully, the style of playing and atmospheric sound lent itself to be manipulated with phase.
Using one of ProTools' native plugins, ModDelay, I decided to introduce a 50ms delay to the left side. This caused the sonic quality in each ear to be different enough for the listener's brain to perceive it as coming from a much wider source. Though it was not spaced apart enough for one's ears to recognise it as being an echo; it sounded like it was being played simultaneously on both sides.
In addition to this delay, I also included two levels of compression and an EQ.
Compressor 1: slow attack and small ratio, to deal with the dynamic range as the performance progresses
Compressor 2: fast attack and higher ratio, to create a consistent ambient soundscape for the vocal and sax to shine through
EQ: HPF at 50Hz, with a 4dB dip at 270Hz to assist the vocal cutting through, and notch cuts at 2.75kHz and 3.39kHz to cater for problem frequencies
Finally I sent the guitar to an auxiliary reverb channel which featured my favourite reverb plugin, ValhallaVintageVerb. Here I found a preset for a spacious bright room, setting the decay time to around 3 seconds, with a 20ms predelay. I added an EQ after the verb, topping and tailing the frequency range to remove muddiness, enabling a suitable range to be virtually reflected.
SAX
Lastly, I EQ'd the sax to remove all low-frequency content below around 150Hz. I applied a gate so that it would only be present when being played, and then compressed the remaining signal heavily due to its significant dynamic range coming into the desk. After sending a lot of the sax to the auxiliary verb channel in the same way as guitar, I quickly adjusted the level so that it would be complimentary to the vocal, and the sax was quickly complete.
ZOOM
The Zoom's room recording captured the audience talking throughout the performance, so unfortunately I was unable to use it for a 'roomy' feel as I normally would. Hence the heavy use of reverb throughout the process thus far. As a solution I chopped up parts of the crowd cheering, and creatively faded it in after each song had finished.
MIXING | MASTERING
Ultimately, each of the channels' levels were reassessed and appropriately mixed with some minor automation, before being sent to a master fader. This channel was adjusted to provide approximately 5dB of headroom for the basic mastering phase which was to come. After exporting the final mix, I then imported the 48kHz/24bit WAV file into Ableton to provide a mono'd bass below 120Hz, a touch of glue compression, some mid-side EQ to polish, and finally ran it through a limiter to bring the level up to a reasonable level. This allows a file of appropriate level to be sent to the band for approval.
As I am not yet well versed in mastering, I follow this simple process to my taste, and leave the station's engineer to properly master the file before radio play.
Soon I hope to hear back from the band with their approval, and the final product will be available to the public for streaming...
Stay Tuned
- TA
REFERENCES
Brain, M. (2008). How Tape Recorders Work. Retrieved from https://electronics.howstuffworks.com/gadgets/audio-music/cassette3.htm
Prescott, S. (2012). Listen: Wonderfuls. Retrieved from https://crawlspacemagazine.com/tag/negative-guest-list-records/
Sweetwater. (2007). Bias Controls On Tape Decks – And What Do They Do?. Retrieved from https://www.sweetwater.com/sweetcare/articles/bias-controls-tape-decks-what/
TripleJ Unearthed. (2019). Wonderfuls. Retrieved from https://www.triplejunearthed.com/artist/wonderfuls
Waves. (2019). J37 Tape. Retrieved from https://www.waves.com/plugins/j37-tape#butch-vig-billy-bush-j37
Waves. (2019). Vocal Rider. Retrieved from https://www.waves.com/plugins/vocal-rider#achieving-perfect-vocal-levels-with-vocal-rider
Waves & Abbey Road. (2019). J37 User Guide [PDF]. Retrieved from https://www.waves.com/1lib/pdf/plugins/j37-tape.pdf
Waves Audio. (2017) How to Quickly Achieve Perfect Vocal Levels – Vocal Rider Plugin Tutorial. Retrieved from https://www.youtube.com/watch?v=cq0N5gDo13U&feature=youtu.be
White, P. (2010). Waves Vocal Rider. Retrieved from https://www.soundonsound.com/reviews/waves-vocal-rider
Comentários