This took a while to create. If you remember, Francesca owns a food fillings factory, and she pranked the office staff when she took over, dumping many of the young women in huge vats of caramel. The AI photos were here at the time.
Now, I've managed to create a short video depicting some of the highlights from that day.
For those wanting to know some of the technical aspects, here is what I did.
1. Bing was used to create the original images. 2. Krea was used to enhance them and make them a little spicier. 3. Kling was used to convert each enhanced image into a short AI video. 4. Power Director was used to piece the video segments together into one video with smooth transitions and to add the background music. 5. Topaz Video AI was used to enhance frame rate and resolution
I've already made mistakes and learned a couple of things. For example, the images I used were square, but the video is 16 x 9 resolution, so most images were cropped and stretched, making them blurry.
I'm now creating images with Kling, and they start out at the 16 x 9 resolution, so when I enhance them, then create a video with them, they will be much sharper and show more detail.
In this scenario, a young lady is painting her bathroom, but slips and winds up in the bathtub. A friend, off-camera, decided to pour thick blue paint onto her.
Your video compilation really inspired me to explore AI Video creation. But the constant tease in your material is when the young lady is sitting getting gunged above a nice gunge pool below, but never gets knocked off of her chair to fall into the pool of gunge below. I'd love to see that, and I'm thinking that you could do that with Kling. Just a suggestion/request.
Your video compilation really inspired me to explore AI Video creation. But the constant tease in your material is when the young lady is sitting getting gunged above a nice gunge pool below, but never gets knocked off of her chair to fall into the pool of gunge below. I'd love to see that, and I'm thinking that you could do that with Kling. Just a suggestion/request.
Believe me, I'm trying to make it happen. It's a surprisingly difficult motion to describe in simple terms, and a lot needs to happen just so: the seat needs to tip forward suddenly. The woman needs to fall forward off the seat face first under gravity. She needs to hit the surface of the gunge and submerge. She optionally needs to reappear above the surface but covered in goo.
Trying to get all these things to happen without weird jankiness is not straightforward. I'm getting closer, but it's not there yet.
Believe me, I'm trying to make it happen. It's a surprisingly difficult motion to describe in simple terms, and a lot needs to happen just so: the seat needs to tip forward suddenly. The woman needs to fall forward off the seat face first under gravity. She needs to hit the surface of the gunge and submerge. She optionally needs to reappear above the surface but covered in goo.
Trying to get all these things to happen without weird jankiness is not straightforward. I'm getting closer, but it's not there yet.
What if you made the movement happen in separate clips? Have her fall forward off of the seat in one clip. Do another where she hits the gunge and submerges. A third one where she emerges covered in goo.
I havn't tried this yet, but you could work backwards. Do the last one first, then take a screen capture to be the end screen of the previous video. (since they let you use an end screen) Repeat and hopefully the three videos would have consistency and could be run end-to-end to create one video scene. Just a thought, and as I said, I haven't tried it.
Bobographer said: What if you made the movement happen in separate clips? Have her fall forward off of the seat in one clip. Do another where she hits the gunge and submerges. A third one where she emerges covered in goo.
I havn't tried this yet, but you could work backwards. Do the last one first, then take a screen capture to be the end screen of the previous video. (since they let you use an end screen) Repeat and hopefully the three videos would have consistency and could be run end-to-end to create one video scene. Just a thought, and as I said, I haven't tried it.
Yep, I've tried that approach. Both the motion of the seat and the falling motion of the woman have to happen in a very specific way that is difficult to describe, and it's presumably difficult for the AI to match this with its training. Whereas a falling, flowing liquid seems to be simple for it to match with its training. I'll get there
To be honest, with the images, I got tired of what sometimes felt predictable and routine after a while. But with the videos, every reaction, every second is new and so realistic. I enjoyed every moment.
Brand new AI video. A woman takes off her bikini and sunbathes nude by the ocean, laying on an air mattress. She forgot her sun screen, but for some reason brought gallons of honey:
Another new one. This one features a woman in a bikini at a beach where blue paint seeps through the sand. Eventually she removes her bikini, but finds out that paint canisters are exploding under the sand, sometimes covering her in thick paint.
Aceman38 said: Tried lumlabs and kling with same two images and prompt
Honestly lumalabs wins hands down which surprised me
1 Lumalabs 2 Kling ai
Let me know which is better.
I think I like both equally. The Luma one is 'on target', but the flow of the Kling one seems more realistic to me. I've found Luma to be very variable, often not producing any real motion at all in the video. If you don't mind me asking, what was your text prompt for this?
Aceman38 said: Tried lumlabs and kling with same two images and prompt
Honestly lumalabs wins hands down which surprised me
1 Lumalabs 2 Kling ai
Let me know which is better.
I think I like both equally. The Luma one is 'on target', but the flow of the Kling one seems more realistic to me. I've found Luma to be very variable, often not producing any real motion at all in the video. If you don't mind me asking, what was your text prompt for this?
I created the pics on bing then used them with this prompt
A two panel photo showing the progression of a pretty woman in her 30s with brunette hair and a red suit is sat on a stool locked inside a tall glass door sized perspex box red frame on stage the background is a gameshow set with a live cheering audience clean looking nervous. she cringes hands over her face as a wide downpour of royal blue smooth gunge pours out a pipe inside the box above her head on top and canopys over her head. And shoulders leaving her covered in royal blue smooth gunge.
Another attempt at an AI video. . . .this one features young women taking a bath in caramel while in an outdoor tub. I added the sounds because AI software is silent at the moment.
and now . . . (drum roll) . . . the first trailer for the non-existant full-length feature movie called "MUDSLIDE- THE MOVIE" which tells the tale of a small city being struck by a large mudsilde.
For some unexplained reason, this city is inhabited by about 97% young women between 20 and 40, and all of them attractive. Hey, we can blame Hollywood for that.
So here is the link, and I'd appreciate any comments on it.
thereald said: Splat the Brat playlist with all episodes:
That represents one hell of a lot of work. It really brings your vision to life, and I see you've managed to drop a couple of the contestants into the gunge pit below in a very realistic manner.
Another AI video from me. I'm actually a little surprised that more of the group aren't playing with video and posting here. I thought it would have been something to explore more.
Anyhow, here is my latest. The well-known Francesca is laying on her bed in her famous red satin dress and playing with maple syrup.
I've been experimenting and have created my first of many compilations. The stuff I'm generating now is better than this so expect part 2 at some point. Ive never edited video before but my god this took a while - please enjoy: https://we.tl/t-Q3Is9Qae0O
Bobographer said: Another AI video from me. I'm actually a little surprised that more of the group aren't playing with video and posting here. I thought it would have been something to explore more.
Anyhow, here is my latest. The well-known Francesca is laying on her bed in her famous red satin dress and playing with maple syrup.
I have been trying. So far, I haven't received a lot of good results from Luma. Kling so far has been amazing quality, but... I have been able to get 2 videos in a little over a week. "Generating" for 48 hours or more, then about a 95% fail rate.
moptopumd said: I have been trying. So far, I haven't received a lot of good results from Luma. Kling so far has been amazing quality, but... I have been able to get 2 videos in a little over a week. "Generating" for 48 hours or more, then about a 95% fail rate.
I started with Kling with about a 20 or 30% fail rate, but now mine is also up to about 95%. Some people say their servers and processors are just too overloaded. If I knew if their paid service was any better I may sign up, but at this point they are not providing a very good service, even thought it's free.