Adobe Max 2019 comes full swing with Sneaks, a segment of the conference where the software giant shows off its next-gen innovations.
Starring Late Night Host John Mullaney, Adobe took the audience through astonishing image, video and audio tools that have been in development and have a high likelihood of making it to end-users in due course.
Presenting these projects were key Adobe staffers who had worked on them, showcasing functionality in live demos on stage.
Project All In
Project All In allows users to composite two images taken from the same perspective with different subjects, allowing everyone to get All In to the final image. Leveraging Adobe’s Sensei machine learning, All In recognises the subjects in the image, merging one of each subject into the final product and avoiding duplicates.
This functionality is perfect for groups of friends or families who want to take a group photo without a tripod or a third party to help capture the image. One (or more) subjects can take an image of the first group before the first group returns the favour for the subject(s).
Project Sound Seek
For repetitive sounds within an audio track, Project Sound Seek again utilises Sensei to identify such instances for elimination or isolation. Adobe personnel demonstrated Sound Seek’s ability to eliminate rampant “erms” distributed over a recording, and to isolate instances of a word for later selection.
Project Sound Seek works regardless of language.
Project Sweet Talk
Also showcased at Sneaks was Project Sweet Talk, which can animate any image with recognisable facial features for speech. Where Adobe Character Animator was the tool of choice for the lengthy, painstaking process of animating characters for speech, Sweet Talk merges an audio voice recording with an image.
This means that character animation for voice no longer requires multiple keyframes just for mouth movements (with much more for other animations), and everything can be accomplished with a click of a button.
Adobe also showcased this technology on a rough pen-on-napkin drawing of a feline character, a Van Gogh self-portrait, and a caricature of John Mulaney – dubbing voice lines into code that automatically animated the respective facial mock-ups.
Next up was Project Pronto, which transforms a tablet into an AR experience creation tool.
Pronto enables a tablet to be the point of view for both the creator and the subject of the experience, allowing developers to directly control the experience of the end user by more closely replicating user behaviours necessary for triggering events and responses in AR.
Pronto was demonstrated with the capability to simulate proximity, in-air pulling gestures, and surface-tapping inputs to trigger animations and feedback in the on-screen AR experience. This would be useful for commercial applications, where users have the ability to bring up additional information or other immersive content when images or objects are scanned.
Project Image Tango
It takes two to tango, and Project Image Tango takes two separate images and tangles them in a completely unique way.
Demonstrated was a terrible rough outline sketch of a bird on a branch (specified as “shape image”), which was blended with a real-life image of a bird – facing the opposite direction (specified as “texture image”). Image Tango blended both images, imbuing the sketch with the rough colour scheme of the real-life image, filling in the background and the branch with life-like colours in a realistic, unique way but clearly inspired by the real-life image.
Also showcased Image Tango’s ability to blend a “shape image” of a dress, with three different styles of “texture image” to produce three different groups of dresses blended from the style image and each texture image. This was repeated on vastly different handbags, and on real-life images of two completely different species of bird – generating images of a completely new species.
Project Fantastic Fonts
Fonts have been an integral part of the functionality across Adobe’s suite of creative products, and Fantastic Fonts takes the possibilities to infinity.
Project Fantastic Fonts provides an interface for users to easily adjust the design of any font – from horizontal to vertical weight, width, height, ascent and descent, for both capital and regular typefaces.
What’s more, users can also easily animate fonts with effect sliders for types of waves, wind, fire, hearts, melt, dance and “hanging cloth”.
Project Go Figure
Animation has been a painstaking process that requires plenty of work in After Effects, with animators having to adjust positions of the characters frame by frame.
Project Go Figure allows users to capture a video of themselves to animate a character by simply linking automatically-generated points of articulation from the video to the character.
Project Light Seek
Bringing practical AR technology to photography, Project Light Seek enables users to adjust the position of lighting in an image in post by capturing several images of the scene from different angles.
The different angles will be processed as a 3D model which will be used to triangulate the position of the Sun or main light source. Using the 3D information of the building or scene, users can adjust the position of the Sun and the shadows cast in the image.
Demonstrated on an image of Cinque Terre, Italy, Project Light Seek could also work with additional images sources online or from stock image platforms that were taken from different angles.
Project Awesome Audio
Not everyone has access to professional recording equipment, and Project Awesome Audio was demonstrated on a recording from the on-board microphones on a laptop.
Awesome Audio was able to clean up the recording and emphasise the voice just like a professional microphone would.
The app, which is in development, was also shown to work on a recording taken when moving between two spaces with different audio signatures.
Including light-and-shadow effects on Adobe Illustrator in an immense undertaking and minor changes would require complete revisions to fix the shadows.
Project Lightstick works as a plug-in for Illustrator that allows users to input light sources in order to cast shadows in their illustrations without any additional work.
Project About Face
In line with Adobe’s initiative to fight fake news and doctored images, Project About Face might be the next step in detecting and policing the wild west of internet imagery.
Project About Face in essence detects doctored images by looking at the pixels, giving a percentage score as to the probability of the image being edited and doctored.
Most impressively, About Face can also reverse-engineer the edits, reconstructing the original image.
Adobe Max 2019 also saw the launch of several landmark apps, including Photoshop on iPad and Photoshop Camera. Read our coverage here.