Queer Eye & Netflix: Recomoji
Netflix teamed up with Queer Eye’s Fab Five to debut the Recomoji bot which recommends Netflix content based on any emoji sent via Facebook messenger. The creative agency Ralph came to us to do sound design, post production, and mix.
For the sound design, I needed to utilize the mobile Facebook messenger sounds. Having had less than 24 hours for turnaround, requesting these assets from Ralph or Facebook was not an option. The FB messenger sounds available online were all outdated and the only way to capture current sounds was to record them from the app itself. I plugged an aux cable from the headphone jack of an iPad into the interface. As I hit record, I noticed that using the headphone jack disables the messenger sounds. After some researching, it was clear that there was no way to change settings or trick the iPad. My solution was an unconventional one, I set up the iPad in an isolation booth and used a Zoom H4N’s microphones to record the iPad’s built in speaker.
This is less than ideal, but these sounds are meant to be heard out of mobile speakers, so capturing them as we hear them shouldn’t result in much degradation. The isolation booth helped to avoid any unwanted noise or reflection. I used the Izotope RX suite to decrease the noise floor and then applied some subtle EQ. I was impressed with how clean and natural these sounds ended up in the mix. This situation helps illustrate how the best solution to an audio problem isn’t always the most conventional one.
For this mix, I was provided with one microphone stem for all of the dialogue. With one mic and five individual voices, the recorded sound varied from person to person. Jonathan specifically seemed to be in-front of, or off axis from the mic, and he sounded more muffled than the rest. My approach to this dialogue repair was as follows:
- Split the Dialogue into individual tracks for each member of the Fab Five (as they needed unique processing).
- Izotope RX Suite: de-verb, spectral repair of any unwanted sounds (de-click, de-plosive, etc), gentle de-ess
- Adjust clip gain so each piece of dialogue is consistently but barely kissing a limiter at -11 dB. Careful to make it sound natural and retain dynamic range
- Fabfilter (EQ, compressor, multiband, and de-esser) to compliment each voice
- For some of the dialogue I used Izotope RX to ambience match from a different speaker to try to obtain a homogenous room tone across all dialogue.
- Utilized a plugin called Audreio to stream my session to my iPhone speaker. This allows me to adjust my mix for mobile in realtime, no bouncing out required.