Don Kotlos Posted December 18, 2017 Share Posted December 18, 2017 1 hour ago, Matthew Hartman said: But don't be surprised if an AI bot figures out how to tell a better one and in much less time. But it is getting better. A novel made with the help of AI actually got pretty good reviews http://bigthink.com/natalie-shoemaker/a-japanese-ai-wrote-a-novel-almost-wins-literary-award I expect AI to significantly improve color grading as well based on the content of the image. And if you haven't heard AI generated music before I strongly suggest visiting this link and playing with the parameters: https://deeplearnjs.org/demos/performance_rnn IronFilm and maxotics 1 1 Quote Link to comment Share on other sites More sharing options...
IronFilm Posted December 18, 2017 Share Posted December 18, 2017 23 hours ago, mercer said: Sound can be cleaned in post No. That is just not true. (Only true to an extremely limited sense, but you need a certain reasonable threshold of acceptable material to work with first. And saying "clean recording" is kinda redundant, when referring then to cleaning it) Would you say it doesn't matter what lens is used or where the camera is pointed?! As you can always "fix it in post". Or would you call a person out for saying that? As it is unacceptable misinformation to be spreading (not without some serious level qualifiers attached to it). Quote Link to comment Share on other sites More sharing options...
mercer Posted December 19, 2017 Share Posted December 19, 2017 3 hours ago, IronFilm said: No. That is just not true. (Only true to an extremely limited sense, but you need a certain reasonable threshold of acceptable material to work with first. And saying "clean recording" is kinda redundant, when referring then to cleaning it) I get it, you’re an audio guy, you don’t want to believe that in 5-10 years time, technology will make it a lot easier for no to low budget films to not need a sound guy. I’m a screenwriter first, so when I heard a script was written with AI, it kinda ticked me off too. As far as cleaning being used twice, sorry it was mistyped... I meant if you get a clean recording, (with a microphone close to your talent) you can raise it in post. Half of the dialogue in Hollywood movies are ADR’d anyway and the on location audio is used as a scratch track. And anyone who has worked on DIY/Backyard films knows you can record audio in camera, with good preamps, if you get the microphone close to the talent for a take or two... with the right microphone. But there are a few programs that can completely clean bad recordings as well... they’re used all the time in the record industry when transferring vinyl to digital and those same programs can be used for cleaning up wind noise and dialogue pops, amongst other issues. But as I wrote earlier, I was referring to the future when camera makers include a better audio set up inside the camera. Hell, for all we know, the Panasonic GH6 could have XLR inputs directly in the body of the camera. 3 hours ago, IronFilm said: Would you say it doesn't matter what lens is used or where the camera is pointed?! As you can always "fix it in post". Or would you call a person out for saying that? As it is unacceptable misinformation to be spreading (not without some serious level qualifiers attached to it). To add, I would say this is already happening... Fincher has been doing this for a few years now and I would assume that the next generation of filmmakers will have no issue at all working that way. Right or wrong, technology moves forward and artists embrace the possibilities that technology offers. Quote Link to comment Share on other sites More sharing options...
Kisaha Posted December 19, 2017 Share Posted December 19, 2017 It is not that simple as having a couple of XLR inputs on a camera, and writing a script is easier, just imagine the databanks of written material already. Google has an interactive library and experimenting with AI for a few years now. Also, you have that on a GH5 already, and a couple of friends mess that up, recently. To replace a sound division you need a robot that has to adjust the mic(s) in position, gain, move fast to one sound source (or person) to the other- and silently, put wireless mics to actors as backup - or primary sound gathering, have algorithms that can judge how to place the mics and how to move them while in action, a myriad acoustic measurements (most of those are the easy part, really), critical judgment if a take is right, or not, and thousands other things. I am pretty confident that sound for video will be one of the last things that robots will do and it is pretty obvious that sound for video will be the last thing you could/would do on a set. But of course we will see. Quote Link to comment Share on other sites More sharing options...
mercer Posted December 19, 2017 Share Posted December 19, 2017 @Kisaha I’m sorry but if screenwriting were easier than audio, then every Hollywood movie would be a great story. Yet every Hollywood has great sound. But I’m not here to say that one is easier or harder than the other, I also wasn’t saying that you’d have a robotic sound guy following you around on set. Quote Link to comment Share on other sites More sharing options...
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.