Monday, February 23, 2009

Speech Animation in Paintings



For this week’s original posting, I would like to address the technological achievements in relation to art and media. Although this video is only a couple years old, the technology used to create this is unbelievable and to this day not used widely. In this video, the director uses numerous famous paintings such as the Mona Lisa which are digitally altered to mimic actions such as lip syncing to begin talking.

The technology used to do this is similar to that used by CG animators in recording movements for character animations. The major achievement here is the ability to alter a still 2D image and enable it to manipulate the area around the mouth without distorting the original image. The director later uses a still photograph of himself in order to demonstrate the programs ability to manipulate various forms of portraits. This program enables the manipulation of images from various angles eliminating the need to recreate a 3D version of an original painting to achieve the same or similar effect. If anybody watches SpongeBob Square Pants, the opening includes a cheesy version of this which uses an overlapping layer of a person’s mouth to simulate speech, but with this type of technology, we can animate all type of images in a similar fashion. Imagine being able to manipulate an entire scene based on pinpoint interactions and movement from a human source. You could quite possibly reenact an image such as events in the Revolutionary War with proper time and development. As technology progresses, it is not certain what we can accomplish with this type of digital ability.

4 comments:

  1. I thought that this was a really cool idea. Although it wasn't 100% perfect, it truly did seem as if the paintings were talking. The way that they manipulated the cheeks and part of the nose really helped and added to the realism. I'm not sure how they did this, but the effect looked similar to using the liquify tool with masks in Photoshop, although that would only be for still images. I think that with this, it would be easy to animate paintings or drawings into simple cartoons. I like what you said about reenacting events from the Revolutionary War. I think that educational-type films to be shown in school would be an excellent use for this technology.

    ReplyDelete
  2. The video reminded me a lot of a character out of the 'Red vs. Blue: the Blood Gulch Chronicles' machinima name Vick. He was a supporting, and the method that the people at Rooster Teeth used to make it seem like he was a speaking was editing his uncomplex, mouthing animation into a sequence to match the audio. This was early in the machinima's life, but late in the final season before Vick's scheme was uncovered, the Rooster Teeth animator(s) did something a lot like this where they used pieces of the original model in a still image to animate in a similar fashion.

    Found a video clip where they were testing it.
    http://www.youtube.com/watch?v=_Io6Nl683hE&feature=related

    ReplyDelete
  3. From the time I saw this video till now, I have been trying to figure out how they did that. This is crazy! Soon we are going to be walk in the museum and we will have paintings talking to us. Now that would be creepy but this new achievement has a lot of potential. They are saying they made this with their own software but I can’t find the software. They probably made a 3d model of the painting and animated the lips. I have this image altering software that can change the facial expression of still images and I can see myself doing animation from that.

    ReplyDelete
  4. This reminds me of all those cheesy animal movies they make where the animals are talking. Some of them make the entire animal in 3D like Garfield for example while others just move their mouths like the move cats and dogs I think it's called. Its really nice how now art has so many options to do what they want to accomplish.

    ReplyDelete