Today we are going to look at the recent rise of generative A.I. programmes and how they can be used in film production. We will see if it is possible to currently use A.I. to make films. Additionally, we will look at how it could potentially change the industry.

“No Strings To Hold Me Down?”

Currently, there are many A.I. tools available that can create a movie’s building blocks. For example, OpenAI tools like ChatGPT, with enough input information, can create a refined script and shot list. Programmes like Speechify can convert text to speech using a variety of voices. Including celebrities like Snoop Dog and Gweneth Paltrow. Applications like Midjourney have already been used to create background images for animated productions. Such as the short The Dog and The Boy. And some tools allow you to edit videoes. Meaning a lot of a movie’s core elements can already be created and accomplished by A.I. But what about a film itself? 

Currently, projects like Pictory and Synesthesia use either stock footage or selected human avatars to make up their videos. Making them more useful for YouTube or business videos. However, trials are being undertaken to produce A.I. that will be able to create original video work through text inputs. Google is researching two A.I. projects aimed at making films. One aims to create realistic-looking A.I. video (Phenaki). Another wants to create high-definition short clips (Imagen Video), both from text prompts. Although these projects currently are only aiming for small videos it could potentially pave the way for longer video creation further down the line and also provides the opportunity to create original videos not based on prior material.

“A Machine, An Imitation Of Life” 

With all of the possibilities and the rate of progression of A.I. what are the implications for the film industry?

Some may say that the errors that A.I. runs into when rendering subjects in images would make it unsuitable for creating films, however, as technology improves, it could conceivably reduce those problems. On the other hand, even if this is not a possibility there is room for discussion of how this technology could be used to create experimental short films, that deal with more surrealist or unconventional imagery. If these programmes become available to the public these A.I. programmes could, like the advent of consumer film equipment and mobile phones before them act as a way of giving space to more independent filmmakers who may not have the finances and time needed to make a film.

It’s hard to see A.I. being a wholesale replacement for human film production. while these programmes can emulate human speech for writing and delivering dialogue, make artwork and other things, the spectacle of A.I.-created art will never wholly do away with the experience of seeing something achieved by humans. Whether it be acting or effects. The art of filmmaking is a collaborative process where teams of people with their own visions and experiences come together to make something with all the idiosyncrasies and sometimes improvised or unintended moments that entails. A.I. could probably never replicate that. There are also definite issues regarding image safeguarding and copyrighting and more to consider. But A.I. does present new ways to challenge and potentially help the industry in the future.

Also Read: A.I Generated Art & Movies: A Collision Course

Also Read: A.I Generated Art & Movies: The Rise A.I In Acting Performances

Your contribution to Big Picture Film Club will allow us to continue to produce content. No matter the amount your contributions are greatly appreciated.

The post A.I. Generated Art & Movies: Can A.I. Make Movies? [Part 3] appeared first on Big Picture Film Club.

Top Gun’s story is centred around a training academy that produces the very best pilots. Tom Cruise plays the ambitious main protagonist (Maverick) who wanted to be number one and Top Gun shows us the rivalry Tom’s character has with the current number 1 (Ice) played by Val Kilmer. Tragedy, coming of age, romance and action, with a cracking 80’s soundtrack. I found that Top Gun is still an entertaining watch as an adult.

Val Kilmer

Val Kilmer plays the confrontational Ice Man. The rival to Tom Cruise’s Maverick. The American actor later starred in Heat (1995), Batman Forever (1995) and worked with Top Gun Director Tony Scott again on True Romance (1993) and Deja Vu (2006).

In his interview with the New York Times Kilmer revealed he had developed a tumour. It was 2014 when a doctor diagnosed him with throat cancer. The treatments that followed involved chemotherapy and a tracheostomy. The Batman beat Cancer, but was left in a debilitating condition. His ailment meant that he now has to eat through a tube and use an electric voice box in order to speak. Kilmer featured in the autobiographical documentary Val (2021) after his recovery. You can hear how the actor sounds post throat cancer towards the end of the trailer.

Remarkably, despite his unfortunate circumstances, Val is still able to work. How could an actor who uses an electrolarynx be able to work? Through the power of technology.

Sonantic

John Flynn and Zeena Quereshi founded Sonantic in 2018. Their aim was to improve upon the common and robotic-sounding text-to-speech solutions that convert written text into spoken words. They achieve a hyper-realistic outcome by creating a model of an actor’s voice using the power of Artificial Intelligence (AI). In 2020 Sonantic created Faith, who they claim is the first AI that can cry. Check out this demonstration of the emotion their algorithm is capable of conveying.

Return of the Ice Man

The Washington Post reported that Sonantic were approached by Val Kilmer’s representatives to make a custom AI voice model for him. The usual process involves having actors record a script and feeding the recorded audio into their Voice Engine (VE). The VE then trains the AI model. But because Val couldn’t record, Sonantic had to research and implement new algorithms. This resulted in more than 40 different voice models, so Sonantic selected the best one to use for Val. After the model is chosen, it can be fine-tuned in Sonantic’s desktop application. You can read more about the full process here and check out a clip of Val’s voice model below.

Sonantic developed desktop software with a beautiful UI (User Interface) that resembles DAW (Digital Audio Workstation) tooling. Operators have options to choose what kind of emotion the audio is delivered in and choices range from anger to happiness. One could opt for a shouting style and even configure the intensity. Perhaps a director wants the actor to convey a high amount of fear in a scene? Well, there’s an option you can toggle to make the voice sound like it is extremely fearful. A custom dictionary allows one to add new words to the model. Editors can adjust the pitch, the pacing and alter the emphasis too. One can even batch process multiple voices at once. It sounds powerful, doesn’t it? Check out the video demonstration to see this all in action.

Is an AI Voice Model Ethical?

Kilmer felt confident enough in his audio doppelgänger to pitch an idea to Top Gun Maverick Director Joseph Kosinski. The voice model and accompanying software offer filmmakers unparalleled control over an actor’s voice. But this leap forward in voice technology raises ethical questions. What if the actor is deceased? Should the model still be used? What if the software is misused in combination with deep fake technology? Could the actor be impersonated? Sonantic addresses these valid concerns in an article sharing their guidelines and principles.

The AI audio pioneers offered their assurances to actors worried machines are coming for their jobs. Sonantic’s mission is to empower actors, not to replace them. A synthetic voice can be used on multiple projects simultaneously. This could drastically increase earning potential and give actors a passive income. Film production budgets will be able to retain funds that would otherwise be used to cater for actors on site. Maybe the benefits outweigh the drawbacks?

Conclusion

I previously wrote about A Space Odessey 2001. A movie which featured the computer AI – HAL 9000. Back in 1968, the fictional precursor to Siri and Alexa was voiced by an actor. But in 2022 the actor could be at home while his voice model went to work for him. Synthesised voices do carry concerns about potential misuse and some may be wary of their potential to replace actors. But Val Kilmer’s story is a shining example of the good that voice tech can do to empower actors. Hopefully, fans of the original Top Gun movie will be glad that Sonantic’s Photoshop for voice exists. After all, you wouldn’t be able to bring back the Ice Man without it.

Also Read: Artificial Intelligence: The New Art of Storytelling?

Your contribution to Big Picture Film Club will allow us to continue to produce content. No matter the amount your contributions are greatly appreciated.

The post How Artificial Intelligence Was Used To Recreate Val Kilmer’s Voice In Top Gun appeared first on Big Picture Film Club.

As filmmakers constantly look for ways that technology could be used to help them, such as what Disney has been doing with recent Star Wars and Marvel projects. With deepfakes becoming more common, the use of A.I. in media is becoming more common, so is it possible that one day we could go to the cinema and see a film written by an A.I.?

How does it work?

The intelligence part of Artificial Intelligence comes from the fact that it can learn. For example, Netflix looks at shows you like, dislike and even the time of day you watch, compares it to similar viewers to offer you recommendations. The more you watch, the more it learns and can make better recommendations. Of course, everyone is different, and it may be harder to determine some viewers’ habits than others.

A.I.s like these rely on information, the more they have, the better they can try to do their jobs. In this example, VFX artists attempt to improve on Luke Skywalker’s appearance in The Mandalorian using a deepfake. The A.I works by looking at various photos of both Mark Hamill as Luke Skywalker from Return of the Jedi, and of the stand-in actor. These then give it lots of information to “replace” the actor’s phase with Luke Skywalker.

A similar experiment was done with writing a Sci-Fi Film. A programme was fed several examples of sci-fi films, before creating its own, called Sunspring. The results are mixed, with some sentences making sense, but the overall plot comes off as nonsensical.

Mr Puzzles

Netflix recently tried the same experiment, this time using horror films. The result Mr Puzzles Wants You To Be Less Alive, was published online.

The “film” is quite obviously inspired by Saw, complete with traps, puzzles and a strange masked figure. A lot of the elements of horror are there, with some lines even working. However, it seems that the program hasn’t quite figured out how these individual elements fit together. The result plays more like a nonsensical parody of the genre. The whale twist is certainly original though.

The bot has clearly picked out key elements of horror movies, Jennifer’s having bodies and featuring chainsaws. However, it’s not able to separate them, resulting in it trying to include everything. Interestingly some parts do make sense, and there’s even something of a character arc for the detective.

While worrying about deep fakes is a real concern, it doesn’t look like we have to worry about A.I.s writing films just yet. The technology is impressive and has some understanding of what makes a horror film, it’s very limited. It doesn’t seem able to grasp the elements that make something scary, or what makes a character compelling. Perhaps this A.I. would be better suited to writing comedy?

Also Read: Your Favourite Movies DeepFaked

Your contribution to Big Picture Film Club will allow us to continue to produce content. No matter the amount your contributions are greatly appreciated.

The post Netflix Horror Movie Written Entirely Using Artificial Intelligence appeared first on Big Picture Film Club.

The same can be said for media texts; they are stories which the writers and directors feel need to be shared and they have particular ideas about how these tales should be told. These decisions are influenced by numerous factors including environment,  personal experience and emotional ties.

So is the emerging trend of using Artificial Intelligence (A.I) taking away the personal touch that writers, directors and editors bring to a production? Or is this simply the direction that the film industry is moving in and we just need to get on board or get left behind?

AI has been used in other areas aside from the film industry to help predict the success of a piece of work. Wattpad is an online program and app which currently has 65 million users worldwide. People are able to publish books via this website and it has seen some become bestselling authors. Sounds brilliant, but where does AI come into it? The website uses ‘plot-analysing software which is able to determine which subjects and content have previously been successful, profile audiences and basically create the perfect plot for a target audience -all a user has to do is write the book to fit.

One person who did just that is Beth Reekles, author of the highly successful book ‘The Kissing Booth’ which she published at just 15 years of age. In 2013 she sold the film rights to Netflix and it is now one of the streaming sites most popular searches.

Wattpad now has its sights set on the film industry with its additional creation, Wattpad studios, and aims to continue its successful use of artificial intelligence.

Jack Zhang’s company, Greenlight Essentials, uses a specially created algorithm to discover which plot will be successful based on existing lucrative productions.

Greenlight used this strategy to come up with a script for a new horror called “Impossible Things” – they made the trailer for just $30 and it accumulated 2.3 million views on YouTube. Needless to say, potential investors quickly showed an interest.

Surely though, this reliance on computers is sapping the feeling from the scripts being created? That’s one thing artificial intelligence is yet to master: emotional intelligence.

This possibility was tested when director Oscar Sharp and Ross Goodwin, a New York University AI researcher used a ‘recurrent neural network’ (that helpfully named itself Benjamin) to create a sci-fi script called Sunspring. The machine was fed scripts of existing sci-fi success stories including classics like Ghostbusters and more recent offerings such as Interstellar and created its own script in a similar way to predictive text on your smartphone. (The term “predictive” implies that creativity isn’t a crucial factor here!)

The result? A script that consisted of completed, coherent sentences individually yet, when these were put together, it all started to go a bit downhill! The film created looks professional and the actors played their parts well but these aspects had nothing to do with artificial intelligence, highlighting that human influence is surely a prerequisite of a successful production.

Zhang agrees and feels that the creativity is still the most vital aspect of storytelling stating; “If you give 50 screenwriters the same elements, they’ll still come up with 50 different screenplays.” This seems to be the crux of the issue; artificial intelligence can help in the creative process prior to production by looking at what elements have proved successful, although even that could take away from the next surprise hit.

I’m sure Zhang won’t mind me saying that he sums it up perfectly:  “ is like a compass, but someone still needs to sail.” Based on Sunspring, I don’t think scriptwriters have too many stormy seas to navigate just yet!

The post Artificial Intelligence: The New Art of Storytelling? appeared first on Big Picture Film Club.