# Hearing It Back: Generative Spatial Audio and the Age of Fully Personal Sound > For a century, every listener got the same mix. Generative Spatial Audio changes that: music rebuilt in 3D around one person, shaped to their ears, their headphones and their hearing. This is the story of giving people their music back. Web page: https://spatial9.ai/blog/hearing-it-back-generative-spatial-audio-personal-sound Author: Spatial9 Team Published: 2026-11-03. Updated: 2026-11-03. ## Key takeaways - For a century, recorded music has been one fixed mix for everyone. Generative Spatial Audio rebuilds each song as a 3D scene, so it can be shaped around one listener instead. - Spatial9 starts by separating a song into its instruments with AI, then places each one in a sphere around the listener and renders it in open formats such as Eclipsa Audio (IAMF) and binaural headphone audio. - Personalization happens in layers: the listener's hearing (an audiogram, per ear, across seven frequencies from 125 Hz to 8 kHz), their headphones (116 models profiled) and their head movement. - Space itself helps hearing: when instruments no longer pile up in one place, the brain can separate a voice from a guitar again. Clarity comes from placement, not only from volume. - Spatial9's hearing features are for enjoying music. They are not a medical device or a hearing aid, and they do not diagnose or treat hearing loss. Quick answer: Generative Spatial Audio is Spatial9's way of turning any song into a living 3D scene: AI separates the music into its instruments, places each one in a sphere around you, and renders it for the way you listen. Because the song is no longer one flat, fixed mix, it can be shaped to your ears, your headphones and your hearing. For people whose hearing has changed, that can mean hearing the voice, the guitar and the room in a song again, clearly and in their own space. [Image: Every song, rebuilt around one person.] Think of the first song that ever stopped you in your tracks. Maybe it was in a car with the windows down. Maybe it was in a bedroom with cheap headphones and the volume too loud. You remember where the voice was, how the drums hit, the moment the strings came in. You did not just hear it. You were inside it. Now imagine that, slowly, without you noticing, that song started to fade. Not the volume. The detail. The breath before the chorus. The shimmer of the cymbals. The second guitar that used to sit just to your left. The World Health Organization estimates that more than 1.5 billion people live with some degree of hearing loss, and most of them never say it out loud. They simply listen to music a little less each year. This post is about giving it back. ## The century of one mix Since the first records were pressed, music has worked the same way. An artist and an engineer make one mix. That mix is copied, a million times, unchanged. Your grandmother, a teenager on a train, a drummer with ringing ears and a child hearing music for the first time all receive exactly the same file. It was a brilliant system for distributing sound. It was never a system for listening. Because no two people hear alike: - Our ears are different shapes. The folds of your outer ear filter sound in a way that is as individual as a fingerprint. It is how your brain knows whether a sound is above you or behind you. - Our hearing changes over a lifetime. High frequencies usually fade first, which is exactly where the clarity of consonants, cymbals and the "air" in a recording live. - Our two ears are rarely equal. Many people hear better on one side, and a fixed stereo mix has no idea. - Our headphones all sound different. The same song on two pairs of headphones can feel like two different songs. One mix for everyone was always a compromise. We just did not have the technology to do anything else. ## What Generative Spatial Audio changes At Spatial9 we call our approach Generative Spatial Audio. The idea is simple to say and hard to build: stop treating a song as a finished, flat file, and start treating it as a scene that can be rebuilt for each listener. [Image: From one fixed waveform, to separate instruments, to a sound field shaped like you.] It happens in three movements. 1. Listen to the song the way a musician would. Spatial9's AI separates a stereo track into its instruments: vocals, drums, bass, guitars, keys, strings and more. It recognises what each one is, so the voice is known to be a voice and the kick drum a kick drum. 2. Give every sound its own place. Instead of squeezing everything between two speakers, each instrument is placed in a full sphere around you: the singer in front, the guitars to the sides, the room above and behind. Then it can move, breathe and react, the way music does in a real space. 3. Render it for the person who is listening. The scene is delivered in open, modern formats such as Eclipsa Audio (IAMF) for immersive playback and binaural audio for any pair of headphones. Because the instruments are still separate when it is rendered, the final sound can be tuned to one person, not to an average. That last step is where personalisation stops being a slogan and becomes physics. | One fixed mix | Generative Spatial Audio | | The same file for every listener | A 3D scene rendered for each listener | | Instruments fused together in two channels | Each instrument separate, with its own position | | Clarity only by turning it up | Clarity from space, placement and per-ear tuning | | Ignores your headphones | Tuned to the headphones you actually wear | | Ignores your hearing | Can follow your hearing profile, ear by ear | ## Space is the most powerful hearing tool we have Here is something that surprises almost everyone: when music becomes hard to follow, the problem is often not that it is too quiet. It is that it is too crowded. In a flat mix, the voice, the guitar, the piano and the reverb are stacked in the same narrow space. A young, healthy ear can untangle them. An ear that has lost some detail cannot, so everything blurs into one sound. Turning it up just makes the blur louder. [Image: Give each instrument its own place, and the brain can follow every one of them again.] Your brain has an extraordinary ability, sometimes called the cocktail party effect: in a noisy room you can focus on one voice, because it comes from a different place than the others. Spatial audio hands that ability back to music. When the voice sits clearly in front of you and the guitars move out to the sides, your brain can separate them again. Clarity returns not because anything got louder, but because everything got room to breathe. This is why we believe spatial sound and hearing belong together. Space is not decoration. Space is understanding. ## Your hearing, your mix Spatial9 includes a Hearing Accessibility feature built on exactly this idea. It starts with something many people already have: an audiogram, the hearing chart you receive after a hearing test. [Image: Your hearing chart becomes a sound shaped for each of your ears.] You can upload a photo of your audiogram or enter the values by hand. Spatial9's AI reads it and builds a personal profile: - Ear by ear. Left and right are treated separately, because they are rarely the same. - Seven frequency bands, from 125 Hz to 8 kHz. From the warmth of the bass to the sparkle of the cymbals, each band is adjusted for where your hearing needs support. - Gentle dynamics. Soft details are lifted without making loud moments painful, so the whisper and the crescendo both stay comfortable. - Clarity and space working together. The profile can widen and clarify the scene, so the instruments separate more easily. Then the music is rendered through that profile, in 3D, in real time. The same song. The same artist. But finally mixed for you. An important note. Spatial9's hearing features are designed to help you enjoy music. They are not a medical device or a hearing aid, and they do not diagnose or treat hearing loss. If your hearing has changed, please talk to a hearing professional, and bring your audiogram along to Spatial9 for the music. ## The moment people hear it back Every technology has a number that matters. Ours is not a benchmark. It is a facial expression. [Image: The song was always there. Now he can reach it again.] Imagine a man who played guitar for forty years and stopped listening to records because "they all sound muddy now". Imagine him putting on headphones and hearing the rhythm guitar step out to his left, the voice right in front of him, the room opening above his head. Imagine him turning to his granddaughter, not saying a word, because he does not need to. That is the moment we build for. Not "more immersive content". Not "a new format". A person, reunited with the music that made them who they are. ## Personal does not stop at hearing Hearing is the most emotional layer of personalisation, but it is not the only one. Generative Spatial Audio shapes sound around the whole listener: - Your headphones. Spatial9 includes profiles for 116 headphone models, so the 3D scene is tuned to the pair on your head, not a lab reference you do not own. - Your movement. With head tracking, the music stays anchored in the room as you turn, the way a real band would. Look to the left and the guitar is now in front of you. - Your taste. Because every instrument is separate, the balance itself can change: more voice for someone following the lyrics, more bass for someone who feels music in their chest. - Your moment. A song can be a quiet companion while you work, or a full stadium when you want to be swept away. And because Spatial9 renders to open standards such as Eclipsa Audio, this is not locked inside one app or one device. The same scene can travel to the screens and headphones people already use. ## A future where nobody is left outside the song We are at the start of something big. For the first time, music can adapt to people, instead of people adapting to music. [Image: One song. Millions of listeners. Every one of them hearing it their own way.] Picture it. A teenager hears a song exactly as the producer imagined. Her father hears it with the voice brought gently forward. Her grandfather, whose high frequencies faded years ago, hears the cymbals shimmer again. Three people, one song, three personal sound worlds, all of them true to the artist. That is what Generative Spatial Audio means to us. Not one more way to make music louder or bigger, but a way to make music personal: to every ear, every pair of headphones, every life stage. Sound was always meant to be felt from the inside. Now everyone gets to be inside it. ## Hear it for yourself The best way to understand personal spatial sound is to put on headphones and listen. - Experience a song in 3D with nothing but your headphones. - Explore the features behind Generative Spatial Audio. - Create a free account and turn your own music into a 3D scene. - Curious about the science of how we hear? Read MIT Psychoacoustics Hacks: How MIT Hacked Human Hearing. Your music is still in there: every breath, every string, every shimmer. Let's go and hear it back. ## Sources - World Health Organization, World Report on Hearing, 2021. - E. Colin Cherry, "Some Experiments on the Recognition of Speech, with One and with Two Ears", Journal of the Acoustical Society of America 25, 1953 (the cocktail party effect). - Jens Blauert, Spatial Hearing: The Psychophysics of Human Sound Localization, MIT Press, revised edition 1997. - Albert S. Bregman, Auditory Scene Analysis: The Perceptual Organization of Sound, MIT Press, 1990. - Alliance for Open Media, Immersive Audio Model and Formats (IAMF) specification, the format behind Eclipsa Audio. - AES69, AES standard for file exchange: spatial acoustic data file format (SOFA), for personal HRTF data.