CAS Quarterly

Spring 2018

Issue link: http://digital.copcomm.com/i/987065

Contents of this Issue

Navigation

Page 58 of 79

C A S Q U A R T E R L Y S P R I N G 2 0 1 8 59 CAS: RX has changed the face of noise reduction, and each version brings with it huge leaps in capabilities. Your team is constantly pushing the technology to do things we've never been able to do. Can you talk about iZotope's plans for RX in the future? Mike Rozett: Source separation: separating signals like speech from noise, or separating musical instruments in a mix. One of the things we're trying to do with the software is make it more usable to people who have music applications. We're looking to take Machine Learning and isolation/ separation technology and apply it to things beyond dialogue, like vocals and other musical applications. Usages can range from production sound to restoration applications to cleaning up things that have been mixed where you don't have the ability to go back to the original stems or tracks. CAS: A lot of folks are looking for the ability to work with multitrack audio. What are your thoughts on that? MR: That is definitely something that we're looking to include in a future release. CAS: Where do you see this all going? (I know that's an impossible question!) MR: There are some general themes that keep coming up for us. People continue to ask for more and more help with dialogue. People are constantly saying that with tighter production schedules, smaller budgets, and more to do, how can we salvage, improve, and use our existing dialogue. So we're always going to be looking at trying to help people save their dialogue and get it to where it needs to be faster. As a company, we've put a stake in the ground with assistive technology. Things that we've put into Ozone and Neutron help people make decisions when one sound is being masked by another. Assistive tech is something that we're always looking at for all of our products. For RX, we're looking to build in features that help people get through audio problems faster. We're also looking into other markets (like podcasting) to introduce technology that can help them get their sound cleaned up faster. When we talk about assistive tech, it's really to help people get better results more quickly and maybe identify some things they hadn't seen. These tools are meant to be supplemental to the person using them, not a replacement for the person using them. We'd like to eliminate some of the easier, more common problems and then provide tools that allow people to dig into the harder stuff that seems unsalvageable. This way, users of our software will have more time to focus on the things that will make the track shine. I don't think I've ever heard anyone say, "Well, we had extra time on this project so…" You never hear that! Developing ways to do things that were impossible in audio improvement and recovery—those are things that the entire company is looking at putting into all of our products CAS: Anything else people don't know about RX? Hidden tips/tricks or Easter eggs? AL: Oh, the Easter egg? It exists in RX, too, but I don't think anyone has discovered it yet. Hint: try searching in the About screen. Seriously, there are a lot of subtle things in RX. You can change the duration of crossfades that RX applies while editing via Preferences > Misc menu. The last, but not the least—our user manual is pretty good, it explains many advanced features with sufficient detail. Clean speech database Isolated noise database Mix x Noisy mixture Neural net Speech mask Error to be optimized – + + + Repaired speech

Articles in this issue

Archives of this issue

view archives of CAS Quarterly - Spring 2018