“This is live audio. Real world, live. Live straight from production to our AV system. No Adobe. All discrete, no encoding or decoding,” explains Nuno Duarte, sound design and audio manager at Olympic Broadcast Services (OBS), of the audio feeds being delivered to rights holders throughout Paris 204.
Speaking to SVG Europe at the Paris broadcast centre for OBS and rights holders, the IBC, he goes into more detail: “We don’t have any loss of quality until the rights holders distribute to the households. So, for example, if I’m a rights holder, I receive discretely, I can handle it as I want. I don’t have to decode and then encode, so I can put the commentator on top, I can put the commentator to one side, I can mix. It’s simple; I don’t have to encode and adapt to the video delays, latencies and all these problems.
“The mic layout creates the immersive mix, it continues with the immersive discrete 5.1 sound. “So we don’t use any processing. That’s not to say that with some sports we don’t need a little help sometimes, but the concept is this: that all the immersive and the treble and the 3D is based on the sound design.
“Then I do everything discretely until (the audio is with the broadcasters who own the rights). Then if they want to encode it in Dolby Atmos or MPEG-H (or something else), they can do it so it gets to the set-top boxes, but (what they receive) is discrete, no processing, nothing.
“What they get, what we hear here, is what the guy in the venue is mixing.”
For the first time with fans
While OBS created a discrete audio mix for the world feed in Tokyo, Paris 2024 was the first time it was offered with audiences in venues. Duarte says: “Since Tokyo, we’ve used it for all sports, 100% top-notch. The big difference here (in Paris) is that we have crowds; in Tokyo it was easier to control. It was the first experience (with crowds, using this sound design concept) for the sound engineers here. It wasn’t the first time they did it immersively, but it’s the first time with crowds.”
For Duarte, that was an extra level of creativity. “My concern was that we have an audience now – because getting these sound effects when we don’t have an audience is easy, very easy – but doing it when we have all these crowds and all these PAs is difficult. Because we have a lot of crowds, the PA has to be high; there’s no other way. But as you can see, we did it.”
Duarte comments on a demo of the audio while watching a rhythmic gymnast perform live on TV: “You use the crowd, you use the instrument (used by the gymnast), the sound system gives out 80 or 90 dB, no problem. We have the reaction of the audience everywhere; you can hear people talking. It’s not a round (ball) sound; there’s an atmosphere and everything is in the space (where it’s supposed to be). We can almost understand how people talk or talk (in the crowd).”
In love with discreet tone
Duarte says the audio for Paris 2024 is going great. Some broadcasters are so excited about the discreet audio that they are offering certain channels without commentary to give viewers a pure, atmospheric experience. Duarte says: “Honestly, the feedback from broadcasters has been fantastic. When I talk to these broadcasters, I find that many people, many channels are offering the sports broadcasts (on some channels) without commentators. I think HBO in Portugal has some channels without (commentary) and people are seeking out the sports broadcasts without commentators to listen to.
“I’m not against commentators,” he adds. “I think they’re fundamental to explaining the sport. But it’s interesting that when someone – HBO and all these telecommunications companies – offers content without commentary, it’s because someone is looking for it.”
Improved Automix
OBS is using artificial intelligence (AI) in some sports in Paris, achieving an advance that has taken months: an automix that follows the camera angles. Duarte explains: “We’ve started doing some things at some levels that we don’t normally do in terms of artificial intelligence. The big difference is that our artificial intelligence is not blind.”
Automixes generally don’t change the audio to match the shot. While some formats will follow a ball, for example, they don’t optimize the mix for a wide shot or close-up. However, OBS worked with a technology provider six months before the games to get the audio in line with the video.
“This means when we hear audio, we see the image. If it’s a close-up or a wide-angle shot, it’s left to right or if most of us are blind, they can’t follow the image. In soccer, for example, when the goalkeeper kicks the ball, the sound should be different if it’s in a wide-angle shot or in a close-up.”
The new improved Automix is being used here in Paris in rowing and experimentally in some other sports. Duarte explains: “The speed of the athletes we use in rowing, for example, is not the same from the start of the heats to the final; it’s almost a difference of 10 or 50 seconds because the athletes are faster. So the sounds have to follow that difference depending on the image. The big difference for us is that the Automix more or less follows what’s out there; it’s not blind, it reads the image, reads the cuts.”
He adds: “AI is a tool. It’s important to say that I’m not using artificial intelligence to replace the A1 or to do the whole mix.”