The Story of XTRAX STEMS

The Story of XTRAX STEMS

The Story of XTRAX STEMS

Audionamix has been dedicated to the science of unmixing audio for over 10 years. Our mission has been to package smart separation technology in an easy to use interface that is accessible and affordable for all. After many years of development, the new XTRAX STEMS represents the realization of that vision. So how did we get here, and what makes the new XTRAX STEMS unique? Read on to learn (and hear!) more.

2014

Audionamix’s first commercial product line, ADX TRAX, was based on advanced signal processing – identifying the fundamental melody of a voice using a pitch guide and then using filters to separate the vocal from the rest of the music. This method relied heavily on manual user work to refine the pitch guide, thus better informing the separation algorithm. As a result, the software required a complicated feature set including an integrated spectral editor, and it often took a long time to get a good result.

We received many requests from DJs, musicians, engineers, producers, and artists for a simpler, faster solution.

2017

The initial release of XTRAX STEMS 1.0 was created in response to this feedback, featuring an easy-to-use workflow balanced with value and quality. We partnered our legacy TRAX algorithms with a new machine learning AI drum separation. The result was a simple, affordable product that could separate and isolate vocals, drums, and remaining music. One song in, three stems out.

“We Shine” by Fabulous Saturdays

Original Clip

“We Shine” by Fabulous Saturdays

XTRAX STEMS v1.0 - Vocals

“We Shine” by Fabulous Saturdays

XTRAX STEMS v1.0 - Drums

2018

Remixers, DJs, and educators quickly recognized the creative potential of such a tool and at NAMM 2018, MusicRadar honored XTRAX STEMS with the Best Software Award.

June 2018 – XTRAX STEMS 1.5 further delivered on our promise to make our technology accessible; it was our first Windows- compatible product, a long-time request from patient Windows users around the world. Version 1.5 was also available in English, French, and Japanese and added export support for Native Instruments’ STEM format, streamlining workflows for DJs.

October 2018 – Since the initial release of XTRAX STEMS, our research team was working hard behind the scenes to apply their latest breakthroughs in artificial intelligence to stem creation. The fruit of this labor, XTRAX STEMS 2, featured the new machine learning-driven “Advanced” algorithm, which represented a huge leap forward both in processing speed, and the quality of the vocal and drum stems.

In an effort to allow users to tailor and improve initial separations without creating a complicated post-processing editor, we added the real-time separation balance feature. The intuitive triangle matrix allowed for even better quality stems without sacrificing usability.

In recognition of these improvements, XTRAX STEMS 2 was crowned the 2019 Best Tech Tool for Schools and was nominated for a 2020 TEC Award for Best DJ Production Technology.

Our main focus became how to maintain and deliver consistent improvements on the technological side and address community feature requests from our growing user base. Considering our regular cloud-based algorithmic updates, it followed naturally that we couple a subscription-based product that allows for consistent and reliable support with a user-facing membership community. We unveiled our vision for the new XTRAX STEMS at Amsterdam Dance Event 2019 and took home the top prize for ADETech’s Companies2Watch.

2020

The launch of the brand new XTRAX STEMS subscription product, reworked from the ground up. The new version adds a fourth bass stem, separation options tailored to specific use-cases, and a De-bleed feature on each track for post-processing control. With GPU processing and updated algorithms, the new XTRAX STEMS is faster and sounds better than ever.

“We Shine” by Fabulous Saturdays

Original Clip

“We Shine” by Fabulous Saturdays

XTRAX STEMS v 4.0 - Vocals

“We Shine” by Fabulous Saturdays

XTRAX STEMS v 4.0 - Drums

Thanks to the subscription model, our team has dedicated resources to bring you regular updates and make continuous improvements in servers and separation algorithms. Our member community includes access to exclusive resources, events, and a forum by which you can directly vote for or suggest new features that are most important to you. We’re committed to our promise to offer the best, most affordable stem separator to our growing community.

Thank you so much for being a part of our story so far. We couldn’t have made it here without you, and we can’t wait to discover what we’ll create together.

XTRAX STEMS Subscription

XTRAX STEMS Subscription

XTRAX STEMS Subscription

We are thrilled to announce an upcoming upgrade to XTRAX STEMS – a completely redesigned enhancement of our best-selling automatic stem creation software. New features include a brand new GUI, and new algorithms which are optimized for your specific needs, whether it be to separate all four stems (vocals, drums, bass and remaining music), to create acapellas, backing tracks, or drumless tracks for practice. Hone your separations using the new de-bleed slider on each track, helping you perfect each stem before you export! As always, our proprietary algorithms separate the entire spectrum of audible frequencies, all the way up to 22kHz!

In celebration of this momentous early 2020 release we are offering two new low-cost subscription options for XTRAX STEMS that will allow you to get free access to the upgrade as soon as it’s available!

Why Subscribe?

Not only are our subscriptions more affordable when you’re on a budget or working on a short-term project, they also help us to offer consistent updates and upgrades to our advanced cloud-based servers which handle the heavy processing in XTRAX STEMS.

In conjunction with the release of our upgraded XTRAX STEMS in 2020 we will be launching a new user forum where you can share your requests and vote for the features that are most important to you. Expect frequent and regular updates as we work together to build XTRAX STEMS into the product of your dreams. The subscription means you’ll get every improvement at no additional cost, and will help us support a consistent and fulfilling development process dedicated to our user base.

How Much Does it Cost?

Pick up a 6 Month Subscription for only $40.00 or subscribe for a full year for only $60.00. Subscribe now and you’ll get immediate access to our current release XTRAX STEMS 2.2 as well as free access to the upgraded XTRAX STEMS as soon as it’s released in 2020!

6-Month Subscription

$40

1-Year Subscription

$60

 

Interested to get your ears on XTRAX STEMS 4 as soon as possible? Current subscribers and owners of XTRAX STEMS are invited to apply to join the final beta testing phase by reaching out to beta@audionamix.com.

If you have any questions about our subscriptions to our software please contact support@audionamix.com.

We can’t wait to share all of the new and updated features of the next version of XTRAX STEMS in early 2020!

The Sound of an Epic Night

The Sound of an Epic Night

The Sound of an Epic Night

Stephen Oliver and JP Quicquaro recount their experiences creating and mixing the new eko interactive series “Epic Night”

We at Audionamix have been familiar with interactive video company eko for several years. Formerly known as Interlude, eko’s tools helped us to create interactive tutorial videos for our TRAX products. Eko is now focused on creating choice-based series, including the college-age romp, “Epic Night”.

In this interactive rom-com, three friends celebrate their college graduation in a night of house parties, burrito challenges, and eccentric rideshare drivers on their way to the biggest music festival of the season. Stepping into the shoes of the protagonist, Martin, viewers make choices at key moments in the story, leading to twelve completely different outcomes. Although a viewer can navigate through an installment in about 10 minutes, each episode contains nearly 30 minutes of content to cover all of the possible storylines.

We sat down with “Epic Night” producer JP Quicquaro and post-production sound mixer Stephen Oliver to discuss the making, and mixing, of this choice-driven series.

What was the biggest difference working on Epic Night versus non-interactive productions?

JP: Production wise, you’re planning a whole bunch of scenes that the user may not see or experience, so logistically there is a lot more to plan for. You’re not writing a through line, a linear story from point A to point B. You’re getting from point A to point Z and every combination of the alphabet in between. When you have this type of interactive content, it exponentially increases the work on the post side. One of the most difficult things is figuring out how to process sound and handle all the different nodes, and marry all those together and get a seamless integration of audio, video, and performance.

Stephen: In most shows you mix one version of a scene and it’s done. For “Epic Night”, because of the interactive nature, there were often two, three, or four versions of the same scene, depending on what choice you make as the viewer. For example, early on in the show the group splits up and you can choose to go with Lillis, or go with Jess. While that’s a simple choice, it leads to entirely different scenarios in the episode, and even to callbacks to that decision in later episodes. We had to keep close track of each choice and where it led to make sure things were mixed consistently. When you click the button to make your decision, it’s a seamless transition, so you never notice a pause or change in the sound. Since the different possible scenes following a decision weren’t always consistent in length, we had to make sure that the audio happening at the end of a scene, and the audio at the beginning of the next scene were exactly the same across all the choices, so they’ll flow together naturally no matter what you choose. I found my most important tools to be Pro Tools for its quick editing abilities and IDC for fast dialogue denoising.

What were some audio challenges that needed to be solved in post-production?

JP: Often when actors in a movie or show are in a car, they aren’t actually driving, there is something used called a process trailer, which is a low flat bed truck that the car is driven onto. So a truck driver is driving the car on the back of the truck so the actors don’t have to worry about actually driving. You might have 15-20 people on the back of a truck filming the car. You add lights and sound equipment and stuff it in the car and everyone stands on the back of the truck that’s driving while the car just sits there. We used a process trailer for the opening scene. The trailer we had was a bit rattly, and there was road noise and engine noise. That was very challenging and that’s our opening scene so it was really important to make it smooth. Everything sounds really great after the post-production work that Stephen did.

Stephen: With the car scenes there were different engine noise between cuts; sometimes a low drone, but during acceleration it was a rising frequency. Since we are cutting back and forth between the characters in the car, this engine noise would jump back and forth quite noticeably. IDC helped to isolate just the dialogue so that we could build a more consistent car tone underneath.

JP: We had another particularly difficult location where we were filming outside a store on the side of a road, and we were there for a long time to shoot all the different outcomes, we went from early evening to 2 or 3 in the morning. The background sound was completely different because we had a bunch of different levels of traffic. If you have consistent ambient traffic, you can work around it, but we went from half the shots having busy traffic, to it being dead quiet in the middle of the scene and it was a really challenging thing. That’s common across all of film but it was compounded because we were there for so long shooting all these different scenarios.

Stephen: We found that the best solution in that case was to both remove traffic noise from the dialogue and then add consistent traffic background from a sound library, to make everything have the same average level of noise. Again, IDC was instrumental in reducing the traffic noise. I was able to go pretty deep without making the dialogue sound unnaturally thin.

Stephen: In one scene, Martin and Jess are walking together through a park at night. It’s a tender moment where they’re really connecting. To properly light the scene the crew needed a generator running, which unavoidably bled into all the mics.

JP: It was a wide open shot, and we only had a certain amount of cable, so the generator had to be close. The actors were also walking towards the generator so it was getting louder and louder as they approached it.

Stephen: There was a hum at different frequencies but also an underlying pulse and mechanical noise, so hum removal helped, but wasn’t enough. IDC let me significantly reduce the generator noise while keeping the speech clear and free of artifacts.


Epic Night was directed by Andrew Rhymer, executive produced by Benny and Rafi Fine, produced by JP Quicquaro and Lisa Steen, and written by Scarlett Bermingham and Sierra Katow.

About the Products Mentioned in the Blog

Watch the full mini-series now on eko.com or the eko app.

IDC: Instant Dialogue Cleaner

The best tool on the market to address common podcast audio problems including room ambiance, wind, traffic and nature sounds.

Audionamix Professional Services Assist Remastered Baywatch Release

Audionamix Professional Services Assist Remastered Baywatch Release

Audionamix Professional Services Assist Remastered Baywatch Release

Separation Stories: ADX Technology in Action for Remastered Baywatch Release

Audionamix Services Specialist Stephen Oliver recounts his experience removing music from the blockbuster TV series’ worldwide HD rerelease.

With its water rescue action and eye-popping scenery, the smash hit show Baywatch has been broadcast in over 200 different countries, and watched by over a billion people worldwide. Its success spawned the star-studded 2017 feature film by the same name, and the original story and characters still have a dedicated fan base across the globe.

It’s no surprise then, that in 2018 the company behind Baywatch, Fremantle, started the process of creating a fully remastered HD version of Baywatch for worldwide distribution, including popular streaming platforms like Amazon Prime Video. The original series was shot on 35mm film, making it possible to digitize and clean the negatives for conversion to high-definition digital picture. This involved painstaking work by many talented people, and the end result looks fantastic.

On the audio side, the project presented another challenge: due to the popularity of Baywatch, many of the songs used in the original series have since become very expensive to license due to their commercial success. Licensing these songs for the remastered version would make the entire project cost-prohibitive.

To make the re-release possible, it was decided to replace over 350 songs with new, more affordable music, which allowed the project to get off the ground. This wasn’t a problem for the English version of the show, as the separate music, effects, and dialogue stems were archived and available, so the music stem could be easily be muted and replaced.

Stems for dialogue music, and effects. Each element is on its own track.

With stems available, the original music track can be muted and replaced.

For the foreign versions of the show, however, these separate stems were no longer available. The only audio available was the full master track with music, effects and dialogue elements combined into one, making it seemingly impossible to remove and replace just the music.

This full composite mix includes dialogue, music, and effects elements all combined on a single stereo track. There’s no way to independently control or mute the separate elements.

Enter the Audionamix Professional Services team, a group of audio experts working with proprietary artificial intelligence technology that aims to find and extract specific elements in a mix without processing or destroying the rest of the information within.

Using a patented process known as “Music Dissociation”, we were able to separate only the music from each scene, leaving the dialogue and effects intact. Music Dissociation works by teaching the AI the sonic fingerprint of the song, or source, in this case provided to us by Warner Bros. The AI then analyzes the full audio file and separates it into two elements, the learned music and the remaining dialogue and effects.

Armed with this technology, our team worked on over 400 clips from the show’s nine seasons, allowing for the replacement of music in the French, Spanish Neutral, Spanish Castilian, Italian, and German versions of the show.

In a perfect world one would perpetually have access to all of the elements that create a full mix, whether it be a movie or a song. But this world is complicated, unexpected challenges arise, and archives are often lost over the years. I’ve been honored to assist clients through these worst case scenarios to help them realize their visions for beloved shows, movies, and songs.

Turns out sometimes even the Hoff needs a helping hand.

2018 Music Expo San Francisco

2018 Music Expo San Francisco

2018 Music Expo San Francisco

On November 10th, Audionamix attended the Music Expo conference in San Francisco, hosted by the SAE Expression College. Students, DJs, remixer, producers, and music aficionados from around the country came to see the latest products and learn the newest techniques from industry insiders.

The night before the event, we attended a welcome dinner where we mingled with the organizers, sponsors and exhibitors. Thanks to Sound On Sound for the incredible food!

Visitors to the Audionamix table saw XTRAX STEMS 2 in action and chatted about their projects and how XTRAX could open up new creative possibilities for remixing, mashups, and more.

Our artist partner, Venomisto, delivered an inspiring presentation on how he used XTRAX STEMS to create remixes of “Wolves” by Selena Gomez, and “Havana” by Camila Cabello. After giving the crowd tips and tricks on working with XTRAX STEMS, Venomisto put on an incredible performance integrating light, sound, and live percussion that blew away everyone there.

We look forward to seeing everyone at another Music Expo soon! Thanks so much to Loic and his entire team who worked so hard to make this event happen.

Separation Stories: Steve Wilk – How to use XTRAX STEMS in Recording Covers

Separation Stories: Steve Wilk – How to use XTRAX STEMS in Recording Covers

Separation Stories: Steve Wilk – How to use XTRAX STEMS in Recording Covers

Separation Stories Part 2: Covers

In the Field with Steve Wilk – XTRAX STEMS

Producers and audio engineers around the world use Audionamix software every day to separate, remix, and create. Producer and Playback Engineer Steve Wilk writes of his experience using XTRAX STEMS in his creative process:

While working on our initial EP for A-List Records, we were asked to record some covers. Traditionally we would brainstorm some songs, pick one and start working on it. First we would need to learn the song, and then we would put it in the right key for our vocalist, and only then begin to experiment with the idea we had.  Obviously, this all becomes quite time consuming, and definitely slows down the creative process!

When I started working as a beta tester for Audionamix’s vocal extractor software, XTRAX STEMS, I thought it would be a great opportunity to combine this software with the process of recording covers.

What I discovered was how much of a time saver XTRAX STEMS is. I could hear a song, say in a cafe, or on a playlist, in a movie, etc. Once I had an idea, I could come back to my computer, get the original song, drop it into XTRAX STEMS and let the software perform the vocal extraction.  Within minutes I would be importing files into my session and working on the idea.

With the first song I tried out I had a new version with drums, a new loop of harmony material, and the separated vocals were muted. Ironically, the singer I was working with listened to the cover and liked the music so much, she wrote a new song on top of it. She didn’t realize it was a cover at all!

This ability to keep in the creative mind is so important today. So often we get bogged down with other time-consuming tasks, only to realize that the idea we had may not work, which can be heartbreaking. XTRAX STEMS has made the process of trying an idea with a cover song both seamless and efficient, allowing me to find out quickly what will work and what won’t.

Steve Wilk with Hannah & The Engineers

Steve Wilk with Hannah & The Engineers

Using XTRAX STEMS has also led me to some interesting musical discoveries, as it can give you a breakdown of recorded music that you normally would never get just listening to a full mix. You can listen to the drum track alone, separate from the bass, which lets you hear more nuance and production techniques.

Listening to the music stem, minus the drums and vocal, really opens up the song and is a great learning tool either for instruments or for production and mixing techniques.

This is really a great way to analyze and hear what is going on inside your favorite tracks, whether you are a musician, producer, or engineer. It can be quite surprising to discover some of the elements that become obvious when the track is separated. For example, degrees of swing, mixing of sounds, and production techniques that aren’t obviously audible in the full mix.

Every time I split apart a track with XTRAX STEMS, I truly am surprised at what I find, as I discover something new each time.

-Steve Wilk

About Steve

Steve Wilk is a Producer/Engineer. He currently works with artist/actress Kat Graham, with Musical Director Michael “Nomad” Ripoll (Babyface). Other clients include Keyshia Cole, Hannah and the Engineers, and various show programming for tours.