Gregory Albert Posted August 6 Report Share Posted August 6 Dear JWsound members. I'm curious about the development of Ai in film making and any thoughts around our profession being at risk. What do we think will happen to production sound recording in the next 10 years ? Many thanks Greg Albert Quote Link to comment Share on other sites More sharing options...
PMC Posted August 6 Report Share Posted August 6 I just completed revisions on 21 five minute interactive muliimedia training modules for one of my clients. Three years ago I hired four voice artists for the original project. Re-voicing all the changes plus recording time would have cost around $3k. I did it with a high quality real-time AI voicing website for $50. Plus my editing time. I can't tell you how much of it was actually AI vs HQ voice synthesis. The term AI is often misused. I could change gender, ethnicity, timbre, speed, emphasis, pause length, etc all on the website interface. Then tweak the .wav export in Adobe Audition for timing to video and gfx. There was a learning curve to get the voice synthesis to pronounce some technical terms correctly but not as bad as when I used to use the Votrox SC-01 voice synthesizer when designing games for the Tandy TRS80 and CoCo Computers in the 80's. Quote Link to comment Share on other sites More sharing options...
codyman Posted August 6 Report Share Posted August 6 I'd say post is going to use it more than production (at least at first). I do fear a future where even large shoots will be 1 person sound teams that have you lav up each speaking actor, record everything even if there is clothing rustle etc (just make it "good enough") and then they take all the recordings and do an AI version of ADR, automating the process and discarding the actual audio used on set. Quote Link to comment Share on other sites More sharing options...
Philip Perkins Posted August 6 Report Share Posted August 6 We have the same issues that actors, extras and all production crew have. A main attraction of AI to producers is the reduction in the number of shoots days required for a project Quote Link to comment Share on other sites More sharing options...
PMC Posted August 7 Report Share Posted August 7 I remember a Frank Zappa interview from many years ago in which he was asked about using the new Mellotron and synthesizers to emulate an orchestral sound in his work. This was during his fight with local government and Studios who were pushing restrictions on home studios with zoning laws. He said the new tech was creatively freeing and didn't require union breaks. Production and Post workflows will definitely change and actors will strike to restrict the cloning/reproduction of their performances. Oh wait, they are. Quote Link to comment Share on other sites More sharing options...
Constantin Posted August 7 Report Share Posted August 7 22 hours ago, codyman said: I do fear a future where even large shoots will be 1 person sound teams that have you lav up each speaking actor, record everything even if there is clothing rustle etc (just make it "good enough") and then they take all the recordings and do an AI version of ADR, automating the process and discarding the actual audio used on set. I have the exact same fear. In the long run, a mic on cam will suffice and no sound crew will be needed anymore at all. That’s just how things are progressing. It’s also what happened to thousands of other workers in other industries. Sad but true. The one thing that might save us is the actors striking to prevent AI from altering any of their performances. Quote Link to comment Share on other sites More sharing options...
Matthias Richter Posted August 7 Report Share Posted August 7 as long as AI isn’t working in real time right on set there is hope that the Director and other on-set crew need a proper mix to hear the dialog which requires a dedicated sound crew as of now. Quote Link to comment Share on other sites More sharing options...
Constantin Posted August 8 Report Share Posted August 8 6 hours ago, Matthias Richter said: as long as AI isn’t working in real time right on set there is hope that the Director and other on-set crew need a proper mix to hear the dialog which requires a dedicated sound crew as of now. I think that will happen. The next SuperCMIT (or the one after that) will have AI built in to ignore everything but dialog. Maybe it’ll have a choice of setting like „Dialog“, „Music“, „Ambiance“ or whatever. With several cameras, there will be one mic on each camera and the auto-mixer in the ifb receiver will create a nice mix from that. But, as sound crews are cheap compared to everything else, I think there will be a stronger drive to get rid of the set altogether. Just let AI create the entire movie. It’ll start with movies that are mostly created by computers already, like large chunks of Marvel or animated movies. It‘s possible that audiences will be bored by this fairly soon and we’ll return to more regular film-making and some directors will insist on realism, but the temptation will be huge. Quote Link to comment Share on other sites More sharing options...
Derek H Posted August 8 Report Share Posted August 8 Quit giving them ideas! Quote Link to comment Share on other sites More sharing options...
PMC Posted August 10 Report Share Posted August 10 Universal Music Group and Google are in discussions to license artists’ voices and melodies for AI-generated content. The two companies are considering developing a tool that people can use to create AI-generated music, otherwise known as deepfake, with popular artists’ voices, the Financial Times reported. Quote Link to comment Share on other sites More sharing options...
Constantin Posted August 11 Report Share Posted August 11 On 8/8/2023 at 5:39 AM, Derek H said: Quit giving them ideas! Sorry… but they have these ideas already Quote Link to comment Share on other sites More sharing options...
John Dombroski Posted August 19 Report Share Posted August 19 Right now, execs are trying to convince post-production to replace striking actors' voices and ADR-time with easily sourced (they have all the files) AI-created audio. I recently found an AI-powered noise reduction Pro Tools plugin that does near everything 833 NoiseAssist (or some time with RX) can do, in the turn of a digital dial... But the neural computing isn't yet at a point where it knows how to mic an actor on an ever-changing film set, nor does it know how to hold a boom, nor does it know how to have an affable conversation with a Director (...or does it? ha)! I don't know if Production Sound has to worry about AI just yet — until Moore's Law brings the world invisible and idiot-proof lavs with mics that can be placed by a chimpanzee. But producers are always going to try to cut corners (after all, it's their job)! [edited for clarity] Quote Link to comment Share on other sites More sharing options...
Philip Perkins Posted August 19 Report Share Posted August 19 I think there will be a change in what "acceptable" production sound is to allow this technology (and the attendant savings) to happen. Many older sound people mourn the advent of the "all -lavs-all-the-time" approach to production sound, vs. what could be done with skilled boom ops, location treatment, set discipline and rehearsals: the "all-lav" sound is now accepted and probably preferred by many. Whatever "AI production sound" ends up sounding like will probably be another step away from naturalness and real-location audio perspective, many of us won't like it, but it will gradually become the accepted norm just as "all-lav" did. Sorry. Quote Link to comment Share on other sites More sharing options...
jason porter Posted August 19 Report Share Posted August 19 I'm on a documentary right now and I have been having issues with doing interviews in loud environments (I'm on a boat) I like ambience, but sometimes it's been overwhelmingly loud, depending where the director/editor wants to shoot. He says not to worry, he uses noise reduction software that makes everything sound like a podcast (SM7b) and I died inside. Quote Link to comment Share on other sites More sharing options...
Philip Perkins Posted August 19 Report Share Posted August 19 You should ask him if that's his motivational speech? What he said is kind of what I meant above. It's also bullshit. Quote Link to comment Share on other sites More sharing options...
PMC Posted August 20 Report Share Posted August 20 Philip, Yes, too many "lav everyone" jobs in resent years. I get that mindset for reality shows but not narratives. I have seen a couple dramatic TV shows in the past year in which the actor's exterior location performance sound (whispering) is so pristine and doesn't vary with camera distance that it is totally distracting. Completely unnatural. I dont think it was ADR'ed. I am sure the post sound guy/gal was very pleased with his/her accomplishment but I sure think it was 100% captured with lavs and very sterile sounding. A good boom person and proper mix, IMO, makes for a far more pleasing and natural dialog track. I say this as the UPM says, "Hey, I can save $600/day if I cut the boom op." Quote Link to comment Share on other sites More sharing options...
Izen Ears Posted August 20 Report Share Posted August 20 This is why *we* need to step up with our unions and demand a three person team beyond a Tier 1! The area standards contract is about to be up for negotiation, and that's the one that we get the most of out here. It's also the one that most of you all use outside of Los Angeles, so you guys could help with this! I have joined my locals negotiation team and I gotta say it's very disheartening. The producers could just decide to say "we won't make a decision on that" and bang! the issue is gone. There's only a few areas that they must respond to, so they usually throw everything else away because they can. The three person sound team is one of those issues. I'm going to propose it anyhow, but I fully expect it to be shot down. My recommendation is to have as many locals as possible put this in their negotiations, that might make them answer. So contact your negotiations team and request a proposal for a three person team for anything over Tier 1! This should actually be its own thread, because I would love to bring other proposals that the national sound community wants in our contracts. Dan Izen Quote Link to comment Share on other sites More sharing options...
codyman Posted August 20 Report Share Posted August 20 13 minutes ago, PMC said: I say this as the UPM says, "Hey, I can save $600/day if I cut the boom op." Studios would love to just have bag jockies with 16 channels of wireless while paying for labor + kit at less than it would cost to rent the kit alone at a rental house. That's their end goal. Quote Link to comment Share on other sites More sharing options...
The Documentary Sound Guy Posted August 20 Report Share Posted August 20 4 hours ago, codyman said: Studios would love to just have bag jockies with 16 channels of wireless while paying for labor + kit at less than it would cost to rent the kit alone at a rental house. That's their end goal. It might be their end goal, but they will be disappointed if they think they will realize any labour savings from it. Managing 16 channels efficiently still requires multiple people to do well (my rule of thumb is 4 tx per person, can stretch to 6, but not a good idea to attempt that long-term). Aside from that, someone from sound still needs to run the set ... this is an important part of the boom op's role that PMs will miss if they try and shrink the sound dept any further (which isn't to say they appreciate how important that is right now). They will lose far more money in botched takes and "waiting on sound" than than they will gain in saving a person on set. From a mixer's / Head of Dept. point of view, it is our job to make clear to the producers that a boom op does much more on set than just swing the boom. They will learn that the hard way if we don't help teach them. Quote Link to comment Share on other sites More sharing options...
Dejan Ceko Posted August 21 Report Share Posted August 21 "Hand Made" is how we do it today . Eventually , most of the content will be created by user and AI , sound is just a part of it . https://www.theguardian.com/technology/2023/jul/20/ai-tool-creates-south-park-episodes-with-user-in-starring-role Quote Link to comment Share on other sites More sharing options...
.png.2bf2d6f7b89ab88ba7729612d7a505f1.png)
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.