Ozone 12
iZotope's flagship mastering suite, enhanced and redesigned to scale with the product's growing ambition.
Target Audience
Audio engineers and producers
Environment
Digital audio workstations
Role
Lead Designer
Scope
End-to-end product design
What is Ozone?
The industry standard for audio mastering
Ozone is a professional mastering suite used in the final stages to prepare music for release. It combines over 20 specialized audio processing tools into a single environment, giving engineers everything they need to shape, balance, and finalize their sound. It's a dense, parameter-heavy tool where speed, clarity, and precision are critical.
The challenge
Before any design work began, I conducted qualitative research sessions with users through Lookback and distributed surveys to our beta groups to identify where Ozone was falling short. Two clear themes emerged from that research. A third challenge, fragmentation and inconsistency in the design system, was one I identified myself and chose to take on alongside the feature work.
A design system that couldn't keep up
Over the years, the interface had become fragmented, with inconsistent components and one-off patterns that increased cognitive load and made the system harder to scale. This lack of cohesion also slowed development and limited our ability to evolve the product efficiently.
Lack of trust in AI-assisted mastering
Trust in Master Assistant was eroding. Its processing felt opaque, limiting user control and making it difficult for engineers to confidently refine results.
Working with poor mixes
These challenges were amplified in real-world mastering scenarios, where users often work with poor mixes that are difficult or impossible to fully fix with traditional tools.
Risks and considerations
Preserving trust while evolving the system
As one of iZotope's highest-performing products, Ozone carried significant legacy trust with its user base. The risk wasn't just breaking workflows. It was breaking the mental model engineers had built around the tool over years. Every change had to directly resolve a documented pain point or improve scalability. Where the tradeoff was unclear, I defaulted to preserving the familiar pattern and flagging it for a future iteration.
From Automation to Collaboration
When the assistant felt like a black box
In our research, users frequently reported that fully automated results felt impersonal and often overstepped without sufficient context. After several generations of improvements to the underlying processing technology, it was clear that the answer wasn’t a better black box. We needed to improve how users interact with the assistants.
Turning a one-click tool into a collaborative workflow
I introduced a guided pre-processing step that lets users shape intent before running the assistant, shifting from a one-click outcome to a more collaborative workflow. Internally, we framed this as a “just add eggs” approach, giving users meaningful input to increase control and ownership while still leveraging intelligent processing.
1. Target selection
Users can select their target ahead to augment the applied processing.
2. Module selection
A simple list and a toggle function enables users to quickly select which modules they want the assistant to consider using.
3. Intensity and output controls
Intensity and loudness controls help set the desired level of transparency with the processing and the target output loudness.
Outcome
Master Assistant's new custom page earned a 96% satisfaction score in beta, the highest of any feature in Ozone 12. Users could now shape the assistant's output before it ran, turning a one-click black box into a workflow they felt in control of.
Designing for Imperfect Inputs
When the mix arrives with problems mastering can't solve
The democratization of music creation tools has expanded the pool of artists and producers dramatically, but not always their mixing expertise. Mastering engineers increasingly work with mixes that carry fundamental problems, tonal imbalance, over-compression, or low-end issues, that traditional mastering tools weren't designed to address. When a master falls short of expectations, the engineer's reputation is on the line regardless of what they received.
Tools built for the realities of modern mastering
I introduced a set of targeted processing tools designed to make imperfect mixes more workable within the mastering stage. These tools address common upstream issues directly, giving engineers control over tonal balance, dynamics, and low end response without breaking their workflow.
1. Stem EQ
Enables adjustment of tonal balance and gain across individual mix elements from a single stereo file, allowing for more precise corrections without needing access to the individual tracks that make up the mix.
2. Unlimiter
Restores transients and dynamic range in overly compressed or limited mixes, recovering impact and clarity that would otherwise be lost.
3. Bass Control
Provides targeted control over low end strength and presence, helping balance muddiness or lack of weight in drums and bass.
Outcome
Bass Control earned 94%, Unlimiter 92%, and Stem EQ 89% satisfaction in beta. Together, these modules gave engineers the ability to deliver outcomes their clients couldn't get anywhere else, turning a liability into a competitive advantage.
Taking on the design debt
Years of incremental decisions, compounded into inconsistency
As Ozone evolved through the years, incremental UI decisions led to a proliferation of component variations, limiting consistency and making the system harder to scale and maintain. This wasn't part of the original scope. I identified it as a risk to the product's long-term scalability and made the case internally to address it alongside the feature work.
A systematic overhaul, built to last
This work focused on reducing redundancy and unifying component behavior, consolidating variations into a smaller set of flexible components that work across light and dark contexts. Hover states were standardized, and inconsistent skeuomorphic elements were removed in favor of a more cohesive digital aesthetic, creating a more predictable and scalable system.
1. A scalable module architecture
The growing module library was reorganized into three clear categories, creating a structure that accommodates new additions without requiring design rework each release.
2. Component consolidation
Components were consolidated into a smaller, more flexible set that work consistently across both light and dark contexts, reducing redundancy and improving predictability.
3. Standardized visual hierarchy
The use of light, mid, and dark values was standardized to create more intentional contrast and a clearer, more consistent visual hierarchy across the interface.
Getting it done without slowing everything else down
Because this work sat outside the core feature brief, getting it approved required building a case. I put together a presentation for relevant stakeholders outlining the long-term cost of inaction and the efficiency gains of addressing it now. To reduce the burden on engineering I implemented much of the visual work myself in code, recruiting our PM to help with additional implementation. I also started a weekly council with the PM and tech leads to align on which more ambitious changes were feasible within the timeline, turning what could have been a blocked initiative into a collaborative effort.
Outcome
Component variance was reduced by 40%, and beta users rated the visual design at 98% satisfaction. The redesigned system gave the product a more consistent and scalable foundation, reducing the overhead of each new release and setting a new standard for how iZotope products look and behave.
Impact
“The updated interface feels more modern, clear, and less cluttered. Metering is easier to read, and overall navigation felt more fluid. The visual feedback in modules like Unlimiter and Stem EQ really helps speed up the workflow.”
A 20-year product, carefully evolved for the next generation
Ozone 12 generated approximately $3.7M in revenue since its release in fall 2025. The update successfully evolved a product with over 20 years of legacy, preserving the core workflows users relied on while strengthening trust in intelligent features, expanding mastering capabilities, and establishing a more scalable system foundation.