Audio professionals see artificial intelligence as a super-smart, super-fast assistant
Artificial intelligence (AI) is already very present in professional audio. It automates and streamlines tasks such as editing, mixing, mastering and, increasingly, music creation and generation. It allows producers and engineers to work faster and more efficiently and offers new creative possibilities. Nonetheless, concerns remain about the potential for loss of human creative contribution and the ethical implications of AI-generated content.
SVG asked some industry thought leaders for their views on this watershed moment in the evolution of professional audio.
Karl Malone of NBC Sports: “The beauty of AI is its ability to analyze a scenario and evaluate many more parameters of an audio signal than we ever could as humans. »
Karl MaloneSenior Director, Audio Engineering, NBC Sports:
I’m excited to have the opportunity to see what can be achieved with new AI tools in broadcast audio. We are already looking to Cedars, Isotopes and upmix engines to handle complex tasks, and the use of Cut The mix of ball spin in European football should be enough for us to take automated and intelligent technology seriously.
The beauty of AI is its ability to analyze a scenario and evaluate many more parameters of an audio signal than we ever could as humans. Although A1s excel at brain-hand-eye coordination, we simply cannot match the computational foresight of AI. He can process information in real time, anticipate results, make decisions or offer suggestions based on both his analysis and the specific training he has received for a given situation.
However, the artistic nuances and creative expertise that define an A1’s work are irreplaceable, making AI unsuitable for mixing a major show on its own. That said, AI could be very effective in handling secondary outputs, such as creating a dedicated mix for second screen feeds – focusing, for example, on mixing close ball effects and commentary radio as a separate audio stream.
AI can also be useful in quality control of large numbers of program streams to be able to check audio and video for all kinds of visual and audio mixing issues: out of sync audio and video, resolution quality artifacts video. fluctuations, missing audio, clipping, metadata sync, phase, etc. It can alert the MCR/BOC operator to take a closer look or listen.
Ultimately, we decide whether we want to use it or not at these early stages, so that no one is forced to implement it.
Gary Dixon of Audio-Technica: “AI is a tool for humans to better respond to unpredictable events caused by interesting human nature. »
Gary Dixon, director, broadcasting and production business developmentAudio-Technica:
The audio is dynamic and the moments worth hearing are usually unpredictable: like a crash during a race, the crowd erupting on the 18th hole, or sustaining a musical note just a little longer during a concert. AI in professional audio and in microphones in particular will be used as a tool to quickly adapt hardware in unpredictable audio situations. Hardware may have limitations in terms of gain structure, dynamics and general EQ, while AI can help humans react to these situations.
However, for the audio to appeal to humans, the final monitoring step and final adjustments will need to be done by humans. AI is a tool for humans to better respond to unpredictable events caused by interesting human nature.
Christian Scheck, marketing content managerLaw :
From a content creation perspective, generative AI is proving extremely powerful. When it comes to video, it is already possible to feed a generative AI engine with some information to obtain usable images.
Likewise, on the audio side, music written and performed by AI engines is starting to scare songwriters and performers, while artificially generated voiceovers for videos and live commentary for shows manage to fool a growing number of listeners by making them believe they are listening to a song. human.
AI has been put to good use in generating closed captions, a task that used to take a long time and can now be prepared in minutes. However, the results still need to be reviewed by a human for consistency, tone of voice, and most importantly, accuracy.
Christian Scheck from Lawo: “AI has been used to good effect for the generation of closed captions, which was previously a tedious task. However, the results still need to be verified by a human being.
In the broadcast industry, more advanced algorithms (can) help audio engineers, for example, cope with a rapidly growing workload, particularly in immersive audio mixing scenarios that require supervision and delivery of multiple presentations and submix formats – all from one console and through a single A1.
Ultimately, AI’s success in real-world production scenarios will depend on its ability to react to unexpected situations. It may very well become a powerful assistant complementing multimedia production based on the use of Lawo solutions, but it remains to be seen whether it will be able to replace high-quality DSP audio or video processing.
AI can add value in other ways. On a modern software platform like HOME Apps, for example, AI can streamline process monitoring, significantly improve debugging times and reduce downtime, as well as facilitate data analysis and help predict failure conditions.
Other applications could include advanced auto-mixing algorithms or intelligent deployment of applications and services to maximize hardware and software utilization in scenarios with limited computing resources.
However, AI must be applied with a level of enterprise governance, as it also presents a multitude of challenges.
Paul Johnson of Q5X: “AI is the enabling technology that will drive the expanded use of live audio in sporting events. »
Paul Johnson, CEO, Q5X:
There are many implications for AI that will expand the scale of audio capture from athletes and officials at sporting events. The dramatic increase in speed and quality of text-to-speech technology will facilitate real-time audio processing, allowing in-game audio to be filtered and corrected to remove profanity and other unwanted language, as well as real-time translation into multiple languages. This will result in increased audio for live games, which is always popular with fans. Once (the audio/video) is captured as text transcriptions, indexing the archived audio/video becomes much easier and the archives much more useful for post-production purposes. AI will also be key to combining appropriate audio with automated digital zoom from a large format video feed. Eventually, AI will be able to mix multiple audio inputs so that the sound focuses and follows the digital zoom target.
From the Q5X’s perspective, AI is the enabling technology that will drive the increased use of live audio at sporting events, and it will continue to expand as processing capacity and speeds increase. We strive to securely capture very high quality audio from players and officials during the match. This high-quality audio is the input required before the AI can work its magic.
Paul Johnson of Q5X: “AI is the enabling technology that will drive the expanded use of live audio in sporting events. »
Ben Escobedo, Market Development Manager, Shure:
Although the term AI Often criticized as just a “fun toy” for generating images or text, the future promises to turn it into a valuable partner for the audio industry. AI will help automate repetitive tasks and solve complex challenges, improve audio workflows and save a lot of time for operators. AI assistants, such as Microsoft Copilot, require top-notch audio quality to effectively capture and process voices. Shure is committed to making this happen.
There is a common concern that AI could replace careers in the audio industry. However, AI should be seen as an assistant and not a replacement. Even as AI continues to advance, the human capacity to understand and communicate effectively remains vastly superior. Live broadcasts and sound productions require quick, immediate thinking to resolve critical issues that can determine the success or failure of a show – something AI cannot replicate at the moment.