Lemon Tree

For a recent single titled “Lemon Tree,” I produced, engineered, mixed, and mastered an alternative/indie track with REEM, whom I met at IMI Studios. The project pushed me to treat music production as an engineering problem: designing a workflow that could translate an artist’s vision into a cohesive record under limitations. It was my first time building a full record around an artist’s existing demo at this level of detail. The session quickly became a test of my system design, signal flow, and decision-making under unfavorable conditions.
The recording process had two main challenges. First, REEM recorded vocals in her living room rather than a studio. Second, the arrangement required a large vocal stack of almost 70 layers. My goal was to preserve the intimacy and rustic tone while maximizing impact and ensuring the mix translated across headphones, laptop speakers, and studio monitors.
Engineering-wise, the living room recording forced careful capture and cleanup decisions. I optimized mic placement to minimize reflections using the environment (furniture, distance from artist). In post, I cleaned the vocals using RX-style tools (de-click, de-noise, targeted spectral repair), removing artifacts without introducing over-processing. I also kept a controlled amount of room character when it supported the aesthetic, using natural ambience as part of the sonic identity rather than treating it purely as error.
The song structure was an additional design challenge. The verse was calm and minimal, while the chorus carried the emotional peak. To avoid an abrupt transition, I approached the arrangement like energy management. I built momentum through stereo width and sound choice rather than simply raising volume. We tested variations and ultimately added a late pre-chorus that functioned as a bridge for the listener, improving the perceived rise into the chorus. On the mix side, I used automation as a structural tool, widening the stereo field and increasing effect sends approaching the chorus, plus a subtle loudness illusion: dipping the master about 1 dB right before the chorus drop so the downbeat hits feel larger without needing a dramatic gain increase.
REEM’s vocal layers required organization and routing. I grouped vocals into buses by function (lead support, high stacks, low stacks) and focused processing at the bus level to maintain consistency. I relied on a two-stage dynamics strategy, slow leveling for overall control (LA-2A) followed by faster peak management (Rvox), plus parallel compression on vocal buses to add density while keeping the lead vocal clear. Instrumentally, I emphasized the low end by layering a barely audible sub under the bass from the pre-chorus onward, increasing perceived weight without compromising the acoustic feel.
The finished track became REEM’s first released single, and my biggest takeaway was that creative outcome depends on technical system design. I became especially interested in why similar tools (like different compressor models) create different tonal results due to design differences, which now motivates my interest in DSP, audio software, and building better creative workflows.