Spotlight Shines on Emerging Technologies for Pro AV at InfoComm 2023
Peavey’s New Single IP-based Network Infrastructure Technology (sNET) is Based on RAVENNA
Roll Uses Generative AI to Simulate Jibs, Dollies, Much More
Roll, a new app for web and iOS, puts an AI spin on timeworn video capture and editing techniques, delivering virtual bokeh, motion graphics and multicam shots, as well as sliders, cranes, dollies and jibs. As the latest consumer artificial intelligence entry, Roll professes to empower anyone to create professional-quality remote video quickly using just an iPhone and Roll’s proprietary software. The Roll app is free to use, with the whole platform offered as a free trial with a paid subscription option. Roll lets users record broadcast-quality remote video calls, add generative AI special effects, and publish quickly.
Roll AI “allows users to add simulated video effects to iPhone footage that would typically require professional camera equipment to achieve, such as stabilized pan or crane shots,” reports The Verge, calling it among the latest “in a boom of new apps and services that utilize AI to simplify technical creative processes.”
Roll AI relies on its own generative AI models “to recreate the filming environment in iPhone footage as a 3D space, allowing users to add text overlay effects and simulate side-panning, dolly, and crane camera movements in post-production and apply various studio effects like bokeh (background blur),” The Verge explains.
The platform also automates the editing process for fast publishing. “Roll captures metadata from audio and video recordings that the Roll editor later uses to reframe hosts and create scene changes based on any on-screen conversations,” according to The Verge.
The company combines two products to form the Roll editing platform: one is the iOS app that relies on the iPhone camera to capture video and upload footage to the cloud for processing and storage; the other is a web-based app that allows previewing, editing, and downloading of the recorded footage.
“Roll says it’s the only remote video calling service that records footage in High Efficiency Video Coding (HEVC), which provides substantially higher video quality compared to other video compression standards at the same bit rate,” The Verge reports, noting that “Roll claims that videos recorded on its platform are ‘sharper than anything on the market for iPhone capture.’”
Each session for recording supports a single host and up to eight call participants.
“Roll is focusing on the ‘prosumer’ market like consumers, independent content creators, podcasters, and influencers who are creating their own content,” writes Newsshooter, which says it’s suitable for “video podcasts, executive interviews, customer testimonials, webinars, fireside chats, virtual conferences, or anywhere where you’d hire a professional video production company or invest in high-end camera equipment to produce professional videos.”
Roll Wants to Recreate Dolly Shots and More Using Generative AI, TechCrunch, 5/31/23
The post Roll Uses Generative AI to Simulate Jibs, Dollies, Much More appeared first on ETCentric.
Character.AI Lets Users Chat with Wide Variety of Characters
Character.AI is a new chatbot that generates facsimiles of conversations with famous personages or original creations. Napoleon Bonaparte, Billie Eilish and Ariana Grande are among the historical or contemporary characters the site recreates using a neural network. Anyone can use the free app to create a character, whether fictional or real, dead or alive, but a paid offering called c.ai+ provides perks including faster response times, priority access and early previews of new features. In addition to a website, the app launched on iOS and Android this month, triggering 700,000 Android installs within 48 hours.
According to TechCrunch, Character.AI could be the viral follow-up to ChatGPT, with “over 1.7 million new installs in less than a week on the market,” compared to “half a million downloads in its first six days” for the OpenAI chatbot.
Backed by Andreessen Horowitz, Character.AI raised $150 million in Series A funding in March for a valuation of $1 billion. The 16-month-old startup created by former Google employees Daniel De Freitas and Noam Shazeer, “offers customizable AI companions with distinct personalities, as well as the ability for users to create their own characters,” TechCrunch writes.
Shazeer and De Freitas previously directed the Google researchers that built LaMDA (Language Model for Dialogue Applications) for a conversational AI. The duo reportedly became frustrated with the pace of Alphabet’s path to commercialization, exiting the company in late 2021.
In November 2021 they founded Character Technologies, and in December they raised $43 million in an angel round. The website says the app is still in beta.
Mashable provides a “quick guide” Character.AI tutorial that says to use it, first “search for the name of a character or for the media (book, TV series, film, etc.) they’re associated with. The search results will usually show the best matches based on your keywords, with the most-chatted-with characters toward the top.” After selecting a character, “a window will open. The character will introduce themselves first, then you can get to gabbing.”
Apparently Character Technologies does not license likeness rights for the real people — including Mark Zuckerberg, Stephen Hawking, Oprah Winfrey and Lady Gaga — populating its site. That’s because “everything characters say is made up!” the website cheerfully advises.
Presumably, when the characters are real, the models train on their online record to approximate dialogue, which could raise IP concerns. As for the user-generated content, Character.AI has posted its terms of service that specify rights to any characters created by users.
Chatbot Startup Character.AI Rolls Out First Mobile Apps, Bloomberg, 5/24/23
One-Third of People Can’t Tell a Human from an AI. Here’s Why That Matters, The Verge, 5/31/23
The post Character.AI Lets Users Chat with Wide Variety of Characters appeared first on ETCentric.
SoundTube Create High-Quality, Reliable Background Music and Paging System
Snapchat+ Introduces ‘My AI Snaps’ for Chatbot Snap Backs
Snapchat is rolling out a new feature for its premium Snapchat+ platform that enables users who send Snaps to My AI let the artificial intelligence know what they’re up to “receive a unique generative Snap back that keeps the conversation going” via My AI Snaps. The feature was previewed at the Snap Partner Summit in April as part of a larger push on AI updates, including the ability to invite the My AI chatbot to participate in group chats with friends and the ability to get AI Lens suggestions and place recommendations. In addition, the My AI chatbot — made free to all users this year — was updated to reply to users’ Snaps with a text-based response.
As for for My AI Snaps, “if you Snap My AI your latest grocery haul, it might recommend a recipe,” Snap said in an announcement that warns “mistakes may occur, so please do not rely on it for advice.” It also says AI chatbot conversations will be saved until the user manually deletes them, which is also true with the generative photos.
Questions are arising as to the sufficiency of guardrails Snap has implemented around the My AI generative photo feature. “This could be a concern for parents of teens, as other generative AI apps, like Lensa AI, have been easily tricked into making NSFW images,” TechCrunch writes, citing “reports that My AI had been responding in an unsafe manner.”
Snap has implemented new insights in a parental controls hub, the Family Center, “to help parents and guardians stay informed about their kids’ interactions with the My AI chatbot,” TechCrunch says, noting “parents can choose to see if their teen chatted with the AI in the last 7 days.”
My AI has garnered some criticism from Snapchat users, who “slammed the app with one-star reviews after its rollout, complaining about the bot being pinned to the top of their Chats, where they can’t remove it or block it,” TechCrunch reports.
Meanwhile, as parents fret, “teens are bullying Snapchat AI,” TechCrunch says, writing that “Snapchat users have been gaslighting, degrading and emotionally tormenting the app’s new AI companion.”
Snapchat’s My AI is powered by OpenAI’s GPT, which “was trained to engage in playful conversation while still adhering to Snapchat’s trust and safety guidelines,” per TechCrunch.
Snapchat Can Now Send You AI Snaps, The Wrap, 5/31/23
You Can Now Send Snapchat’s AI Chatbot Photos and Get a Response, ZDNet, 5/31/23
The post Snapchat+ Introduces ‘My AI Snaps’ for Chatbot Snap Backs appeared first on ETCentric.