i love how you’ve embraced the morphing of characters and done something WITH that tendency of the AI, turned it into something of intrigue, rather than left it as an annoying artefact. As we've exchanged in various emails Peter, it's a real journey you've been on here, and i feel I'm learning vicariously through your stumbles and wins... so thank you for sharing... Keep on!
Richard, it's a pleasure to share with you, David and Alix. It's a most unusual journey, so full of frustration when things go wrong, and with these bots, they really do go wrong, as I detailed on a previous post. But as I've discussed with you and others, I like to push the limits, and try to turn things around. So when I saw I could not get Luma to take directions (e.g., "Pan left" gave me a pan to the right) and then I noticed that they offered a "first frame " option as well as a "last frame" option, I experimented with those and found that sometimes they panned in the right direction.
The next step was to feed in a "first frame" from a still of the man and a last frame of a still of the woman. These two stills had been created for me by another AI bot, not Luma. But Luma then produced the morphing I hope would occur. So I only got results by pushing and mucking around after many frustrations.
We can discuss more of these things later Richard. Just remember that these bots are in their infancy... and Alix agrees with me that they will learn from all that we and others are feeding in. The question is, "how long will they take to learn?" pt
I'm responding now to all of the above and to one extra comment sent to me by email:
"I find it very strange in an unsettling way, yet on the other hand very interesting and feeling a desire for more. Not more information, but more time spent. It’s good, but I’m not sure I’m sold on the change of identity/back and forward of it. One back and forward/change of identity, would have been better for me."
My response to that was complex, and I feel should share it, because it states as clearly as I can the crazy relationship I have to these Ai forays.
Well, I understand that... I think of it as a sketch! I sketched it enough for the time I was sketching it. Sketching with it, or within it.
There were so many options arising from it, but I felt a bit constrained by the sheer enormity of effort required to implement any of these options, so now I feel that it is a series of moments with possibilities that come into being (out of nothing) and then recede into a loop, and then I felt it was "enough"... no need to do more!
Also it was just a series of "tests" of these crazy Ai engines which do not take instruction, they have not been programmed to be very accurate. Ask for one thing they give you another, or the VERY OPPOSITE.
"Noir" as it stands, is just me accepting that it offered me these characters which are sort of Hollywood archetypes, not real people, but constructed out of some "universal" consciousness and filtered through Ai engines. They are like lots of characters in 40s Film Noir movies, but also they are like "none" of those characters.
Another thing I had in mind was a sort of "Last Year at Marienbad" where everything is unreal, and all threads "loop" back upon themselves, all ideas are uncertain. But to explore that in depth with these Ai "fragments" would have been too difficult and painful for me. So it remains a simple sketch which took me a few days, including heaps of frustration, plus learning what I could take out of that frustration and turn it into something else. That was probably the best part of it.
I'm not sufficiently in love with it to take it any further, so it is "finished" although incomplete. It's all a game, creating things out of nothing. But it becomes deadly serious when I try using AI with images of Gaza, or some shots from the 1950/51 film. I can't bear what it does with any ideas I text in as prompts in either of those, it's totally destructive and in some ways horrific.
So, I came to this conclusion: it is good enough for the time I was engaged with it!
Darko has recently introduced me to multimedia artist Maria Korporal. Of Dutch origin, Maria lived in Italy for many years but she has lived and worked in Berlin since 2014 . Her artistic production includes video art, interactive projects, installations. Her multimedia work has been designed using a wide variety of techniques. The narrative aspect, both personal and social, plays a major role in her work which together with the directness of the images and the sounds leads to greater participation of the viewer. In her work she plays with virtuality and reality, with undergone and artificially generated experiences. Most of her projects deal with social and environmental issues, often with an ironic approach. A good example is her project Korporal Zoo , a series of video works made between 2010 and 2018, which observe the animal and human world from several perspectives: cultural, social, environmental. Over the last two years Maria experienced the world as extremely turbulent
Recently I posted this: From Richard: We knew it was coming, but it's still hard to believe February 20, 2024 In that blog Richard introduced me to SORA which I found astonishing. Some of you may also have been as astonished as I was. Since then Richard sent me a new lead on "LUMA DREAM MACHINE" which offers in some respects more than what SORA did. Here's a link to their main page which has been going viral: https://lumalabs.ai/dream-machine Well I 've been playing with Luma for a few days now, and to-ing and fro-ing with Richard and Alix... some of the results took my breath away, some were totally disappointing. Let's start with an early test piece... from one of my interests in Cave Art. I asked Luma to create a video only from TEXT, so I fed this text in as the first prompt without using any images: An elderly neolithic man and two younger companions ent
Eyeless in Gaza... Footnotes: I wanted to leave these images up clear of any text. My friend Richard suggested we should acknowledge the sources so I've made a big effort to track them down out of respect for the incredible job they do, and especially for their courage, working in the dangerous zones where so many aid workers and journalists have lost their lives. But some are not credited on the pages of publications I drew them from. Here's a list I've just completed, numbered from the image No.1 at top of page down to the final one, No.9 1. Abdulqader Sabbah/Anadolu 2: Mohammed Dahman/AP 3. P hotograph: Anadolu Agency/Anadolu/Getty Images 4. Photo: WFP/Ali Jadallah 5. April 18, 2024. (AFP) 6: Mahmoud Issa: Reuters: Redux 7. Photographer: Mahmud Hams/AFP/Getty Images 8. P hotographer: Anadolu Agency/Anadolu/Getty Images 9. Mohammed Abed/AFP via Getty Images
Very impressive, given what it COULDN'T do with your cowboy clip.
ReplyDeletei love how you’ve embraced the morphing of characters and done something WITH that tendency of the AI, turned it into something of intrigue, rather than left it as an annoying artefact.
ReplyDeleteAs we've exchanged in various emails Peter, it's a real journey you've been on here, and i feel I'm learning vicariously through your stumbles and wins... so thank you for sharing...
Keep on!
Richard, it's a pleasure to share with you, David and Alix. It's a most unusual journey, so full of frustration when things go wrong, and with these bots, they really do go wrong, as I detailed on a previous post. But as I've discussed with you and others, I like to push the limits, and try to turn things around. So when I saw I could not get Luma to take directions (e.g., "Pan left" gave me a pan to the right) and then I noticed that they offered a "first frame " option as well as a "last frame" option, I experimented with those and found that sometimes they panned in the right direction.
ReplyDeleteThe next step was to feed in a "first frame" from a still of the man and a last frame of a still of the woman. These two stills had been created for me by another AI bot, not Luma. But Luma then produced the morphing I hope would occur. So I only got results by pushing and mucking around after many frustrations.
We can discuss more of these things later Richard. Just remember that these bots are in their infancy... and Alix agrees with me that they will learn from all that we and others are feeding in. The question is, "how long will they take to learn?"
pt
Cool video,❤️😎
ReplyDeleteWe're all dabbling Alexia, and your work was inspiring to me in leading to this piece.
ReplyDeleteartists inspire each other !
DeleteI like the she becomes he bits.
ReplyDeleteI'm responding now to all of the above and to one extra comment sent to me by email:
ReplyDelete"I find it very strange in an unsettling way, yet on the other hand very interesting and feeling a desire for more. Not more information, but more time spent. It’s good, but I’m not sure I’m sold on the change of identity/back and forward of it. One back and forward/change of identity, would have been better for me."
My response to that was complex, and I feel should share it, because it states as clearly as I can the crazy relationship I have to these Ai forays.
Well, I understand that... I think of it as a sketch! I sketched it enough for the time I was sketching it. Sketching with it, or within it.
There were so many options arising from it, but I felt a bit constrained by the sheer enormity of effort required to implement any of these options, so now I feel that it is a series of moments with possibilities that come into being (out of nothing) and then recede into a loop, and then I felt it was "enough"... no need to do more!
Also it was just a series of "tests" of these crazy Ai engines which do not take instruction, they have not been programmed to be very accurate. Ask for one thing they give you another, or the VERY OPPOSITE.
"Noir" as it stands, is just me accepting that it offered me these characters which are sort of Hollywood archetypes, not real people, but constructed out of some "universal" consciousness and filtered through Ai engines. They are like lots of characters in 40s Film Noir movies, but also they are like "none" of those characters.
Another thing I had in mind was a sort of "Last Year at Marienbad" where everything is unreal, and all threads "loop" back upon themselves, all ideas are uncertain. But to explore that in depth with these Ai "fragments" would have been too difficult and painful for me. So it remains a simple sketch which took me a few days, including heaps of frustration, plus learning what I could take out of that frustration and turn it into something else. That was probably the best part of it.
I'm not sufficiently in love with it to take it any further, so it is "finished" although incomplete.
It's all a game, creating things out of nothing. But it becomes deadly serious when I try using AI with images of Gaza, or some shots from the 1950/51 film. I can't bear what it does with any ideas I text in as prompts in either of those, it's totally destructive and in some ways horrific.
So, I came to this conclusion: it is good enough for the time I was engaged with it!