I haven't uploaded in some time, but I've been playing around with some magazine covers. Started out as re-imagining the Sports Illustrated Swimsuit Issue as something a bit more formal, and it went on from there with a number of publications. Hope someone enjoys them.
Although Google Flow is a real pain to use on my phone due to Google deciding not to format the site for mobile browsers, I've really been enjoying Nano Banana 2. For someone who gets REALLY specific with my prompts, it does an amazing job of including almost all the details I ask for, and understands location, perspective, and fore/background a lot better than any of the other generators I've used. Best of all, it can actually do action poses without having the subject floating in the air with limbs splayed all over the place. Plus, it does a great job of keeping the descriptions of different subjects separate, meaning I can control poses, actions, and outfits for multiple people in the "photo" instead of using generic descriptions or limiting the number of people included.
I'm not a fan of its default "happy" facial expressions being over-the-top, comical laughing faces though. Having to constantly prompt away those faces is kinda annoying. Overall, though, I love Nano Banana 2!
Whisk, which I've been using to generate almost all of my AI videos until now, is sadly being withdrawn by Google after April 30th, so I'm now familiarising myself with Flow to establish to what extent I'll be able to continue creating videos. I'm still testing, but thought I'd share this 15 second extended video for a bit of fun!
Kabe22 said: Although Google Flow is a real pain to use on my phone due to Google deciding not to format the site for mobile browsers...
a few days ago google said it had added the choice of format (length v width) so it is still making changes and improvements. I'm having mostly good luck with Flow although I find I have to specify "pretty" or "beautiful" or I get someone who isn't or just average.
Sleazoid44 said: a few days ago google said it had added the choice of format (length v width) so it is still making changes and improvements. I'm having mostly good luck with Flow although I find I have to specify "pretty" or "beautiful" or I get someone who isn't or just average.
Flow works great on my computer, aside from the occasional time-out failure. On my phone, I have to refresh the page every few generations because they just start timing out (sitting at 99% until the generation fails). It seems like every time I download an image from Flow, the next generation fails and I have to refresh the page. Also the top of the text box gets hidden behind other elements on the page, and if the generation fails and I want to retry the prompt, I have to copy it a few lines at a time. If I try to Select All in the the text box, the popup menu is either hidden behind the stuff at the top of the page or doesn't popup at all. It's just really annoying to use on my phone, so lately I've just been generating the first iteration of the prompt on my phone so I don't forget the idea, then continuing it the next time I'm on my computer.
As for having to specify the subject being pretty/beautiful... I haven't had that issue. Even when I do simple, generic prompts involving casualwear or physically-active situations where the default is natural/no makeup, the subject is almost always on the high end of the scale of attractiveness, just without the obvious makeup. I have to get really detailed with prompts to get unattractive women (I regularly use AI for NPC images for D&D if I can't find one I want to use online). Maybe some part of your prompts are unintentionally telling it to make the women ugly?
Edit: I've tried using Flow in desktop mode on my phone, but that's really annoying to use for other reasons. If I don't zoom way in, the buttons are so tiny that I was hitting the wrong ones on accident. If I do zoom way in, then I'm constantly either scrolling way left/right/up/down or zooming out just to zoom in on another area of the page. And now that Flow made the picture menu a hover-to-trigger effect, that's just another annoyance on desktop mode.
Honestly, I just had a picture of that black and red dress in my head and couldn't get it out until I made a picture of it. And then of course I had to get it wet... kinda. The other woman's soaked, but the one in the pretty dress stayed mostly dry because I got distracted and never finished the series of her getting into the bubblebath.