Hey Matt. I’m new to your Substack. A graphic designer of 30 years. A fairly good photo retoucher and was using photoshop back before it even had layers, which was V2.5 . Maybe you’ve shown examples of real video vs GAN generated video side by side with your applied curve adjustments, but for me that would be helpful. To know what the baseline is between real and CG. If you cut that into every video you rolled out it would help to sell this.
It would make it easier to show people without a graphics background the difference so they can instantly say, no trucking way!!
It’s just like a weird acid trip without any point of reference even for me, and I kind of know what I’m looking at.
Or releasing an action for a common video editing app so folks that are savvy can do this themselves or having a website where people can upload video, have your effects applied and download again so they can share their findings.
Older folk that still believe what they see on screen need this. They have so much hate because of this scam.
Can you break down, in lay terms, the technical part of what are we supposed to be looking for in these images? Are there other "neutral", unrelated examples of the same technology being used somewhere else, for comparison?
I read and viewed the images on your other posts. I did see your analysis of the frames where the Toyota truck's tailgate and soldiers don't move, unlike everything else. However, I still have difficulty understanding what should I be looking for in those images, when it comes to the telltale signs that the engine could not keep up with some parts of the characters it generates.
I think you'll have to really spoon-feed your audience on this one, since this is really advanced and new technology. Not to mention the whole cognitively dissonant aspect due to how reality and fiction blend almost seamlessly. I think there's no too many zooms, arrows, circles and slow motions and replays you can use with this kind of material.
You're doing a great job, thanks.
PS. consider your posts liked. For "some" reason, Substack does never processes my "likes".
Hey Matt. I’m new to your Substack. A graphic designer of 30 years. A fairly good photo retoucher and was using photoshop back before it even had layers, which was V2.5 . Maybe you’ve shown examples of real video vs GAN generated video side by side with your applied curve adjustments, but for me that would be helpful. To know what the baseline is between real and CG. If you cut that into every video you rolled out it would help to sell this.
It would make it easier to show people without a graphics background the difference so they can instantly say, no trucking way!!
It’s just like a weird acid trip without any point of reference even for me, and I kind of know what I’m looking at.
Or releasing an action for a common video editing app so folks that are savvy can do this themselves or having a website where people can upload video, have your effects applied and download again so they can share their findings.
Older folk that still believe what they see on screen need this. They have so much hate because of this scam.
Can you break down, in lay terms, the technical part of what are we supposed to be looking for in these images? Are there other "neutral", unrelated examples of the same technology being used somewhere else, for comparison?
This is just a sample.
"neutral" = a control group.
I have many ;-)
Thanks for your reply Matt.
I read and viewed the images on your other posts. I did see your analysis of the frames where the Toyota truck's tailgate and soldiers don't move, unlike everything else. However, I still have difficulty understanding what should I be looking for in those images, when it comes to the telltale signs that the engine could not keep up with some parts of the characters it generates.
I think you'll have to really spoon-feed your audience on this one, since this is really advanced and new technology. Not to mention the whole cognitively dissonant aspect due to how reality and fiction blend almost seamlessly. I think there's no too many zooms, arrows, circles and slow motions and replays you can use with this kind of material.
You're doing a great job, thanks.
PS. consider your posts liked. For "some" reason, Substack does never processes my "likes".