Talk-Zone #65 Interview with Soundtheory

Thank you for giving me this opportunity. I started to become interested in audio recording and engineering at an early age and I was fortunate enough to receive solid musical training during that time. When I was studying physics and mathematics at university later, I got the opportunity of working with music software companies. Ever since then, researching and developing new signal processing methods has been one of my greatest passions.

However, my physics research work took me to other places. In 2004 I was doing research on certain problems related to the foundations of quantum theory. I then realised that one of the mathematical formulations I was using could potentially be very useful for rethinking the approach to time-frequency concepts in auditory signal processing.

Around that time I met David Pringle who is also a physicist. I must have been very convincing when I told him about my ideas because he decided to fund my signal processing research privately. After a few years we had rewritten most of the usual time-frequency theory and started to apply the framework to auditory perception, with great success! Even more time was needed to develop fast algorithms and useful application concepts. So after a total of ten years of research, we hired an additional developer, Andreas Beisler, and started to explore and design product concepts. Two years ago we founded Soundtheory and started to focus on the development of commercial products. Finding the right first application with market impact wasn’t easy and we experienced a few set-backs.

Gullfoss is an adaptive equaliser that listens to a signal and decides how to prepare the information in the signal so that your brain can take the most out of it. The realtime analysis of Gullfoss uses our computational auditory perception model to understand which audible elements in the signal are competing for the attention of your brain. An intelligent algorithm then tries to distribute the brain’s attention amongst the different audible elements in a fair way, by altering the level of the individual elements but at the same time keeping the total perceived sound level unchanged.

The main fields of application for Gullfoss are mastering and mixing. It helps to improve clarity and definition quickly without adding anything artificial. I regularly use Gullfoss to fix interference problems when recording a single source with several microphones.

For example, if you record a classical Guitar with two or even three microphones to capture the stereo image, you will run into phase problems on the stereo mixdown. Inconsistencies between different notes and in the stereo image are just two problems that are very hard to fix. Gullfoss working on the stereo sum can remove such issues in a very natural way.

Another possible application is balancing your mix. Gullfoss can reveal and fix problems with the balance of different instruments. If you have access to the individual tracks, the realtime EQ graph display can be used to tell you how to improve your mix by changing the levels of individual tracks for example. If you only have the stereo sum, Gullfoss will try to rebalance the individual audible elements in a way that gives a more consistent total mix.

In a mastering scenario, Gullfoss helps to improve the translation of the mastered signal between different listening situations. For me it helps to greatly reduce the time spent on taking masters to my car, the living room stereo and different headphones and monitors.

Gullfoss also works really well as a non-intrusive de-esser on vocals and dialog. In fact, whenever you have something in your sound that annoys you, there is a good chance that the internal auditory model finds it annoying too and fixes it.

Using a tool well generally requires familiarity and experience with the tool. This is no different for Gullfoss. I do believe, however, that a good tool makes learning easy and natural. Therefore we put a lot of thought into abstracting the internal mechanisms of Gullfoss into very few parameters that map very naturally to how we perceive sound. The effective use of these parameters doesn’t require deep technical knowledge. So we’re convinced that Gullfoss is able to help anyone with their sound, independent of their technical understanding.

NAMM was great and we have received fantastic feedback. A lot of very experienced engineers were deeply impressed with our demonstration and the word about Gullfoss spread quickly. For a new company with no reputation at all we probably had the best show possible. Practically everyone who listened to the demo understood the concept and the potential and reacted very positively.

Gullfoss is currently available only for Mac. A Windows version is a few months away. We use C++ with a custom framework for most of our coding. The other Andreas is responsible for coding our GUI, as a team the three of us decide on what we want, and we used a freelance to do our core graphics.

We’re a group of three, with only two of us doing development. With the success of Gullfoss we may well gradually add a few team members. There are many things that need to come together in the development process. Because our product was based on genuinely unique underlying theory, it took a long time to get the fundamentals sorted. Then it took a significant time to be able to encapsulate everything so that it worked in real time. And finally, once we chose our first product, it took even longer than we expected to get all aspects of the app working in the way we were happy with. In total, if you don’t count the prior research, Gullfoss took 18 months from first prototype to final product.

I received classical piano training and picked up a few instruments on my own later. I’ve also started playing bass more seriously a while ago. I’ve also played in a few bands, but to be honest, I much prefer to stay in the background and give the band the best sound possible. So I see myself as an audio engineer much more than a musician.

There are so many astounding artists who have my greatest respect that I cannot pick one single source of inspiration. I can be inspired by Vivaldi, the Beatles or Daft Punk. And someone with a guitar, a great voice and an emotional performance in a pub can definitely give me goose bumps.

The beauty of having developed a new approach to audio analysis is that we have a great many options as to where we go next. Nothing has been decided thus far.

Soundtheory are giving away licences for Gullfoss to three lucky winners of our Noizefield competition. All you have to do is to download the trial, and mention what you think of Gullfoss on Twitter, Facebook, a blog or a video. Send a link of what you’ve done to info@soundtheory.com with the subject “Noizefield competition”. Entries close on April 1st 2018

We’re really happy that after our long journey we have created something that will help everyone to enjoy their own music even more. Music and audio have always been my passion. It’s such a wonderful way to make the world a better place. We’ll keep working hard to create the best tools for this amazing community! Thank you to everyone who supports us! We will not disappoint you!

More information:
Soundtheory
Soundtheory at Twitter
Soundtheory at Facebook