Back to News
ProductAI Understanding briefing

iZotope launches Photon, an AI‑powered stem separation and audio‑enhancement plug‑in

iZotope’s new Photon plug‑in uses twelve custom‑trained machine‑learning models for stem separation, vocal restoration and noise reduction, and is priced at $259 through Dec. 7 before rising to $349.

4 min readRead the linked source
Source-page capture accompanying iZotope launches Photon, an AI‑powered stem separation and audio‑enhancement plug‑in
Source referenceSource recorded
Publisher
djmag.com
Source type
Linked source — primary-source status has not been established.
ContextUnderstand this in 60 seconds

Key terms

Feature
An input variable used by a model to make predictions.
Latency
The time between sending a request and receiving the model's output.

What happened

iZotope announced Photon, an AI‑driven audio plug‑in that runs locally on a GPU and can be used as a standalone application or an ARA plug‑in within compatible DAWs. The tool bundles twelve custom‑trained machine‑learning models split into two functional groups—Demix for stem separation and Denoise for audio restoration. Specific models include a Voice Separator that can isolate up to six speakers on a single track, a Voice Restore model for cleaning vocals, and a drums model that separates a drum performance into five elements (kick, snare, hi‑hat, toms, cymbals). In partnership with Boris FX, Photon also adds restoration tools for outdoor wind, traffic and indoor room noise, plus a “Cinema — 4 Stems” that splits video or audio files into dialogue, music, sound effects and an “everything else” stem. The plug‑in is compatible with macOS 14+ and Windows 10+, and is offered at a discounted price of $259 until 7 December, after which the price rises to $349.

iZotope released Photon as both a standalone application and an Audio Random Access (ARA) plug‑in that can be loaded into compatible digital audio workstations (DAWs). The plug‑in operates locally on the user’s GPU, eliminating the need for an internet connection during processing.

The product bundles twelve machine‑learning models, divided into two categories: Demix for separating audio stems and Denoise for cleaning and restoring audio. Specific models include a Voice Separator capable of isolating up to six speakers on a single track, a Voice Restore model for vocal cleanup, and a drums model that separates a drum performance into five distinct elements.

In collaboration with visual‑effects company Boris FX, Photon adds audio‑restoration tools aimed at reducing outdoor wind, traffic noise, and indoor room noise. It also offers a “Cinema — 4 Stems” that can split a video or audio file into four stems: dialogue, music, sound effects, and an “everything else” category.

The plug‑in supports macOS 14 or later and Windows 10 or later. iZotope is offering a launch discount of $259 until 7 December, after which the price will increase to $349. The product can be purchased directly from iZotope’s website.

Source details: djmag.com

Why it matters

Photon represents a notable step forward in bringing sophisticated AI audio processing to everyday music producers, post‑production engineers and content creators. By running locally on a user’s GPU, the plug‑in avoids the and privacy concerns of cloud‑based services while delivering real‑time stem separation and noise reduction. The inclusion of multiple specialized models—especially the ability to separate up to six vocal tracks—could streamline workflows that previously required labor‑intensive manual editing or expensive multi‑track recordings. For independent creators and smaller studios, the discounted launch price lowers the barrier to accessing technology that was once limited to high‑end post‑production houses. However, the tool’s effectiveness will depend on the quality of the underlying models, GPU compatibility, and how well it integrates with existing DAWs, which remain to be independently verified.

AI‑driven stem separation has traditionally been offered as a cloud service, which can introduce , data‑privacy concerns, and recurring subscription costs. Photon’s local‑GPU approach provides real‑time processing without sending audio to external servers, appealing to users with strict confidentiality requirements.

The ability to separate up to six vocal tracks on a single stem is unusual for consumer‑grade tools and could reduce the need for multi‑mic recording setups, saving time and resources for podcasters, musicians and post‑production teams.

The discounted pricing lowers the entry barrier for independent creators, potentially democratizing access to advanced audio‑AI capabilities that were previously limited to high‑budget studios.

The partnership with Boris FX suggests a cross‑industry collaboration that may lead to further innovations at the intersection of visual effects and audio processing, expanding the utility of AI tools across media production pipelines.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
Interactive Concept Check+10 Points
What is AI? Quiz

A route planner searches possible journeys using explicit rules. What does this illustrate about AI?

What to watch next

Future updates to Photon’s model library, user adoption rates among professional and hobbyist audio engineers, and any third‑party benchmarks comparing its separation quality to competing cloud‑based services. Additionally, iZotope’s pricing strategy after the introductory discount and potential expansion of platform support (e.g., Linux or mobile) will be key signals of the product’s market trajectory.

Independent benchmarks and user reviews that assess the quality of Photon’s stem separation and noise‑reduction compared to established cloud services such as iZotope’s own RX suite or third‑party AI tools.

Adoption metrics from professional studios, post‑production houses, and independent creators to gauge market reception and real‑world workflow impact.

Potential updates from iZotope expanding the model library, adding support for additional DAWs, or introducing new AI‑driven features such as automatic mastering or adaptive mixing.

Pricing adjustments after the introductory discount period and any subscription or licensing models that may be introduced.

Related guides & quizzes

Found this useful?