SoundID demo

More personal stereo with SoundID

Image credit: Soundlabs

Is the best possible sounding audio equipment a mirage? Is the ultimate sound in fact much more personal than we ever thought? Dickon Ross was at CES to hear one such solution, SoundID, for himself.

“How I thought about it before was that ultimately there is this Holy Grail of sound,” says Sonarworks co-founder Martins Popelis. “At CES, some company will come up with the perfect headphones and everybody will say ‘Wow! This is it!’ – and that will be it for the end of time; everybody will say that’s the right sound. And that keeps on not happening. But still people keep chasing that idea that there’s this Holy Grail of sound.

“What we came to conclude is that perhaps ‘perfect’ is rather personal,” he continues. “Maybe that’s OK. We have different tastes in food and music and whatnot and why would we actually ultimately hear the same? It’s like trying to prescribe the same glasses for everybody. So that’s the research insight that SoundID was built upon.”

Of course it’s no surprise that some people prefer more bass, while others might prefer a more natural sound, for example. Headphone manufacturers have long known that and design for those preferences. But Sonarworks’ research with listeners found the differences go much deeper.  

At first, it had aimed to give listeners a sound experience close to the studio sound. “We come actually from the recording studio technology side. We started off seven years ago with building a product for the music creators. We solved an old analogue problem with a kind of innovative software approach,” says Popelis.

“All speakers and headphones colour sound. The frequency response of speakers and through the headphones is not equal, like these headphones sound different from those headphones, speakers in different rooms sound different.”

Manufacturers try to make studio equipment with a ‘flatter’ frequency response, and in the music industry’s boom years studio acoustics became a whole art for sound engineers. But the growth in ‘bedroom studios’ means “more and more people are recording music in a very sub-par environment”, says Popelis. “For the younger people coming up, essentially what our studio software does is to take the different frequency responses of speakers and headphones and we can calibrate them to the same reference point. And what they then find is that their work becomes way more efficient, they get more confidence in the sound.”

From that beginning, says Popelis, “at some point we said let’s bring this studio sound to the user; the exact artist’s intention will be delivered to the user. Nice. So we built the product called True-Fi and we launched it into the market and it sort of failed.”

The problem with the flatter, studio sound is that it’s not to everyone’s taste, and everyone hears differently anyway. Play the same music for two people through the same studio headphones and they may well hear something different. So Sonarworks decided to research how people hear sounds and what they like. The company’s studio background allowed it to take listening research out of the controlled lab environment, with samples in the hundreds of users, and into people’s homes, where they could reach tens of thousands. “We have studio technology that enables us effectively to calibrate every headphones to sound very much the same. So that enabled us to scale it through a web platform, because we can control the test environment over distance. We just ask the user, ‘Hey, what headphones do you have?’  They tell us and then we are fairly sure that the user is actually hearing the right sound when we ask them this or that.”

The result surprised even Sonarworks: “If you take one single particular sound, then it’s the best possible sound for no more than 17 per cent of the population, probably even less. So if you take the best possible headphones that you can imagine, the best possible thing that the manufacturer can craft, then it will be the perfect sound for no more than 17 per cent.” That explains why people like different headphones. “To crack the best possible sound you really need these three layers of optimisation. You need to take the device into account, you need to take the preference into account, and you have to take the hearing into account.

“So SoundID is ultimately a technology and a product that ensures individually perfected sound on any device. On the technology level, it’s built on machine learning and big-data algorithms that sit in the cloud and that optimise your sound. It [does this] based on the device that you’re listening on, then we calibrate it to the studio sound level, which is also important, to take the music creation sound reference point as the starting point for this personalisation journey. So we start there with the device optimisation. At the moment we have over 360 headphone models supported, and that number keeps growing.”

The demo SoundID in the app stores takes the user through a series of A/B choices to establish personal preferences (the hearing difference component will be added to the real product when it launches later this year). It then builds your unique SoundID profile, which is stored in the cloud and eventually could be applied to personalise your listening experience on your phone, streaming service, computer or in your car.

I tried the SoundID at CES and it worked for me: the better-​quality headphones coloured the music very differently from a cheaper pair. Yet with the SoundID switched on, they sounded eerily similar – although I could still discern the better-quality set, as I should be able to. Both pairs sounded better to me with my SoundID than without.

Sonarworks hopes SoundID will help its original user base of music creators too. “They have a translation problem where they say ‘I listen to something in the studio and then I leave the room and it sounds not that good’.” They can test their work on various sound sources and then optimise it in the studio so it sounds acceptable everywhere. “But they’re not necessarily optimising for perfection in that way. Now with this, they can still do the best job they can in the studio, but we actually assure that the end user will hear the best possible version of their song personalised for their particular music.”

Musicians can’t create ‘fixed monuments’ in a studio that sound exactly the same outside, even if they wanted to. “What the artists rather want is to be liked, to engage emotionally with people. With this approach, where they create music on the reference level and then allow it to be personalised for each listener, they assure that they deliver maximum emotional impact.”

Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.

Recent articles