Photon: Crossgrade from RX Advanced

iZotope Photon is a standalone app and ARA plugin for AI stem separation and audio enhancement. Powered by 12 cutting-edge machine learning models, custom-trained on our own real-world audio data, Photon runs locally on your GPU for a fast, easy workflow. No uploads, no cutoffs, no interruptions. Split stems, remove noise, balance levels, and isolate vocals, even in conversations with up to 6 speakers.

When you need to get in, get work done, and get out in a few moves, Photon’s got you. Need to demix? Quickly split a finished file into voice, drums, bass, and other elements. Segment a scene into dialogue, music, effects, and ambience. Plus, separate conversations into multiple speakers and split drum kits into individual elements. Need to denoise? Easily clean up echo, wind, traffic, and background noise – all while balancing levels and vocal clarity. With bespoke models built to match your needs, just set the strength, click anywhere in the clip, and hear results right away.

Plus, stay ahead of the curve with new and improved models as they roll out in the Model Manager. And for pro fixes without ever leaving your DAW, the Photon ARA plugin brings the full engine to your timeline. No choice overload. Just drag, drop, dial, and deliver.

The first joint research effort from iZotope and Boris FX. The continued development of SoundApp. Photon brings the expertise of two teams together, sharing cutting-edge tech to go further, faster.

What does Photon do?

Split stems Turn any finished song, scene, or drum kit into editable elements for quick tweaks and rebalances.

Separate speakers Split up to 6 voices into stems, plus isolate room tone & background noise for clear conversations.

Remove noise Clear out echo, wind, traffic, and background noise with custom-trained models built to fix, fast.

Restore voices Bring distant, muffled speech forward, and balance levels so every word is heard loud and clear.

Features

Demix
Split the mix | Have the final say

A finished file is supposed to be locked in. With Photon, you get the key. Drop in any audio or video file and split it into editable elements – music, voice, sound effects, individual instruments, individual speakers – all without needing to open a session. Experiment with cutting-edge machine learning models. Pick the right one for you. Adjust what you want. Drop what you don’t. And export what you need. All in a flash.

Music 4 Stems
Songs to stems | Unbake the cake

Quickly turn a finished song into four individual tracks for voice, drums, bass, and other elements. Prepping a remix or rebalance? No need to track down the original session files. Just split any mix into stems, preview instantly from anywhere in the track, tweak what you want, and get what you need.

Music 4 Stems Ultra
Songs to HQ stems | Give stems their solo

Just want the vocals? Or maybe the bass line? Music 4 Stems Ultra is your go-to when the stem needs to stand alone, not just sit in the mix. Split any mix into voice, drums, bass, and other elements with finer precision. Because sometimes one stem needs to be the whole story.

Voice Separator
Stems for each speaker | Herding chats

One recording, multiple people talking? With Voice Separator, you can split up to 6 voices into individual stems plus isolate room tone and background. Turn chaotic chats into clean and clear conversations.

Cinema 4 Stems
Split the scene | Undo the cut

Want the soundbite without the soundtrack? Need to replace some expired background music? Split any video or audio file into voice, music, sound effects, and other elements to isolate, tweak, and export what you need.

Drums 5 Stems
Stems for each drum | Unmix the kit

Got the perfect snare to sample? Not sure the kicks are hitting hard enough? With Drums 5 Stems, you can split your kit into kick, snare, hi-hat, toms, and cymbals. Whether working from a drum bus or full song, there’s no more all-or-nothing. Get the drums you want, just how you want them.

Voice 2 Stems
Split voice and music | Skip or solo the vocal

Whether you need a quick instrumental or acapella, Voice 2 Stems is here to help. Split any track into voice and music instantly, then find your balance – all vocals, all music, or somewhere in between. From prepping a karaoke mix to cutting background music in your content, this is your most straightforward separator for voice and music.

Denoise
Clean any scene | Sound good, anywhere

Real-world audio is never perfect. Echoey rooms, rush hour streets, windy hills, tinny calls – distractions are everywhere. Thankfully, no matter where you hit record, there’s a cleanup model for you. And don’t worry about diagnosing the problem. Photon’s cutting-edge machine learning models find and fix it, while balancing levels and enhancing vocal clarity. Just pick the model that matches the environment, set the processing strength, and take recordings from cloudy to clear.

Voice Restore
Voice enhancement | Loud & clear

Podcast guest sitting too far away? Muffled mic talk from your livestream? Traffic noise in your interview a little too loud? Voice Restore gives you the control to enhance speech and handle noise, levels, mic tone, and echo. Keep the context you need and make sure every word is heard, loud and clear.

Indoor (Fast)
Cleanup for room noise | Tidy the room

Shooting content at home? Recording interviews inside, untreated? Anywhere four walls did you no favors, clear out the noise and echo distracting from dialogue.

Indoor Noisy (Fast)
Cleanup for busy rooms | Tame the room

Frantic factory floors? Echoey event halls? For indoor noise that goes past the ordinary, this model has you covered.

Outdoor (Fast)
Cleanup for outside noise | Take the outside out

The great outdoors – without the noise. Wherever the shoot, clean up general background noise from anything recorded out in the open.

Outdoor Traffic (Fast)
Cleanup for traffic noise | Street talk to sweet talk

For content anywhere the motors are running, reach for this model, trained specifically on traffic and urban noise.

Outdoor Wind (Fast)
Cleanup for wind noise | Gusts, gone

Rooftops, coastlines, anywhere out in the elements, the mic can take a real beating. Calm the wind with this model built for the toughest conditions, not just a light breeze.

Across the app

• Photon ARA plugin. The full Photon engine inside your DAW timeline. ARA hands it the entire clip rather than a real-time buffer, so it works with the full context of the file. Run several models together, and let the processed audio cache carry results across every instance of a clip so you can switch and compare without waiting. Available on Pro Tools, Sequoia, Samplitude, REAPER, Cubase, and other ARA2 hosts.
• Model Manager. We’re always improving to help you work smarter and faster. Discover and download new and upgraded models as they’re released in the Model Manager, and try them right away on the project in front of you. No need to reinstall.
• Local GPU processing. Everything runs on your machine for full privacy and optimized speed. Your audio never leaves your device – crucial when the material is under embargo or simply not yours to upload. No queue, cutoffs, or interruptions.
• Preview from anywhere. Move to the moment you care about and start processing there. Hear results in seconds, then adjust and re-listen as often as you need.
• The first joint research effort from iZotope and Boris FX. The continued development of SoundApp. Photon brings the expertise of two teams together, sharing cutting-edge tech to go further, faster.