An easy and great improvement that I found, is that if you rotobrush out the main focus, in your case the guy+scateboard. And run that through ebsynth separately, you get a much better result. you get separation from foreground and background so ebsynth doesn't blend the two together and struggle with color bleed. You also get some more creative freedom, you can create outlines, or effects on your main character, or add some blur to the background. :)
I know right 😂 I was intimidated by it to be honest. Now it’s easy. Thank you for watching, I feel like I need to update this video because I’ve learned a lot more since then.
@@enigmatic_e i am getting into AI stuff for video. Would you recommend starting off with stable diffusion or with ebsynth? and why? Any info is appreciated
Bro, you do things the hard way. Just blend it with the original, and overlap. Shoot in 24, output 12fps in media encoder, make each frame 3, then cross desolve on frames 1 and 3. Its way easier. Img2img means image to image.
I would swear I watched this video years and years ago and not just the last year. Am I wrong? Did you reupload it or something a year ago? I have been seeing ebsynth tutorials from 5 years ago, not last year and I thought I saw yours about at the same time.
its a bit more tricky when you try to use something like Midjourney or any AI art generator for that matter. Here I just used a style transfer which transfers a style that's very consistent. Might make a video about style transfer.
I exported a png sequence of a man and used 1 frame of Thanos as the style example thinking it would make the man look like thanos. but what EbSynth is producing is just different frames of the 1 Thanos example I gave it. Its completely ignoring my PNG files of the man.
Thanks for your tutorials. I"m trying to animate a still image using a video as the "animation/displacement" source. So the video drives the movement in the still image, which is a face. Any suggestions for the technique? Thanks.
hey. when you put your keyframes(style)does it have to be from the video (png photos) or any style you want can be put it out there to generate the animation? Thanks
Great video! I was able to create a few seconds of stylized video from this tutorial. I get an error when checking Google drive files for my to Neural Style Pt and was wondering if you knew what the problem was.
Hey, my brother, you don't need to add Chinese subtitles, that will increase your work, I can understand it, I hope you can produce more tutorials for stable difiusion animation, the animation I did was not stable, my face would change, come on
Thank you so much in sharing this info Brother and yeah it was great watching this video and it was neat and clearly explained, I will surely come up with a creative work and let you know bro ❤️
How did you make that frame with the AI, I wanted to do something similar and I couldn't, it would be great if you could tell me how with the concept you got that first frame with that cool style. Thanks for the video.
Anyone know a good way for achieving a lot of keyframes? I painted 90 frames for more detail, but i'm having trouble figuring out how to even load that many into EbSynth. The most I can do at a time is 15 keyframes, any help is much appreciated!
do you have to use the first frame 0000 to do the stylized reference? My first frame 0000 doesn't show the face of my talent, can I use any frame in the folder?
I appreciate your knowledge transfer and the content is awesome, but please plan out the video a little or cut out the parts which are unnecessary because it misdirects the viewer and they might lose focus.
Re img2img, you might have already figured that out, but if not, that's a Stable Diffusion option. It might only be in the Automatic1111 Web UI. I think it generates an animation between two image keyframes -- have not tried it yet as I'm still figuring out how to install SD with an AMD graphics card
@@mikealbert728 when i export from premiere or AE i just name it 0 and it will number the frame 0000 to last frame e.g. 0120. Of course the style frame should match the original frame from which it came from e.g. original frame 0000 and style frame 0000.
@@enigmatic_e Thanks. I figured out my issue. I had the naming right but I had too many key frames. It seems if the amount of keys extends the window below the size of the desktop they just won't be added. Trying to find a work around
Man! You are a monster teaching, I love the way you teach. I have had bad results with the EbSynth and Stable Diffusion. My problem was that I tried to edit a video clip of a raper and I had constant failures in the consistency of the video. I didn't do the blending that you teach (I'll try it) but I didn't like the results at all. Maybe it was because only the face appeared in the video and there was too much movement? The truth is that I do not know and I am trying to do it in another way (several steps are from other videos that you have published). As soon as I can I will join your patreon without a doubt. Keep teaching please!!!
An easy and great improvement that I found, is that if you rotobrush out the main focus, in your case the guy+scateboard. And run that through ebsynth separately, you get a much better result. you get separation from foreground and background so ebsynth doesn't blend the two together and struggle with color bleed. You also get some more creative freedom, you can create outlines, or effects on your main character, or add some blur to the background. :)
wow ! cool to see ur progress since this video. u were saying in the beginning u haven’t gotten into stable diffusion and now ur so good at it
I know right 😂 I was intimidated by it to be honest. Now it’s easy. Thank you for watching, I feel like I need to update this video because I’ve learned a lot more since then.
@@enigmatic_e i am getting into AI stuff for video. Would you recommend starting off with stable diffusion or with ebsynth? and why? Any info is appreciated
Always making quality content man I learn so many new techniques watching your videos
Bro, you do things the hard way. Just blend it with the original, and overlap. Shoot in 24, output 12fps in media encoder, make each frame 3, then cross desolve on frames 1 and 3. Its way easier. Img2img means image to image.
Yea I know 😅 I’ve learned a lot since I made this video. Might make an updated version of these days.
Hey @johnshow
How can I contact you, looking for someone to do something similar to what you described as regular work
Mathematical 😊
been grinding your channel bro. good shit
I would swear I watched this video years and years ago and not just the last year. Am I wrong? Did you reupload it or something a year ago? I have been seeing ebsynth tutorials from 5 years ago, not last year and I thought I saw yours about at the same time.
Nice video! BUT o_O How do you get so consistent styled images to use as a reference. Midjourney has remix but it wouldnt be so consistent.
its a bit more tricky when you try to use something like Midjourney or any AI art generator for that matter. Here I just used a style transfer which transfers a style that's very consistent. Might make a video about style transfer.
Thanks for this. I did not know EbSynth existed.
Very cool. Gives me a new direction to go on a vid I have in mind.
👌🏽
SUPER helpful thank you so much for making this
Great tutorial! Thank you!
Stable diffusion "img2img" is just using an image to create an image, in stable diffusion. (image to image).
Thank you, i figured it out eventually and even made some videos on it! It’s my favorite thing at the moment.
@@enigmatic_e Cool, I'll check them out :)
Awesome video! Super helpful!
👍🏽
I exported a png sequence of a man and used 1 frame of Thanos as the style example thinking it would make the man look like thanos. but what EbSynth is producing is just different frames of the 1 Thanos example I gave it. Its completely ignoring my PNG files of the man.
I'm still new to editing, I was wondering whether youre using Premiere or after effects in your ebsynth videos? Or maybe a combo of both?
I mainly use after effects because the work is easier in it in my opinion.
Hi can adding more that one stylized key frame or multiple help with the movement? How do you do that?
What type of computer do you need to produce this time and time again ? Im just learning and have an iPad Pro 2018
This is hands-down the most helpful tutorial on EbSynth I've seen. Thank you|
I can't get this to work. It always says the resolution doesn't match. Wha does this mean?
its likely your style frames are not the same dimension as the video frames
Great job..What styleTransfer Ai did you use?
Hey, thank you! I used Neural Style PT
@@enigmatic_e where can I ding this style?
Nice tutorial.
Hi! Does anybody know why mine does not render from the start of my video to the end?? It only renders my reference photos
how many reference photos do you have? i believe ebsynth can only do 20 at a time
Thanks, dude (y) Just that need this /m/_ _/m/
Thanks for your tutorials. I"m trying to animate a still image using a video as the "animation/displacement" source. So the video drives the movement in the still image, which is a face. Any suggestions for the technique? Thanks.
Anybody knows the song in the background?
And.... Thanks for your amazing videos. I love your voice and your videos editing skills.
Interesting thanks ^^
VR creator
Prototype
D for 2 decades!
So when you say you ran the photo through AI @ 3:36 what did you use? Was that Topaz or Stable Diffusion?
It’s an AI called Neural style transfer. I may do a video on that. I get quite a few people asking me about that.
Thanks for that! Amazing result! What did you use for the style transfer on the first pink version, which you use in the thumbnail?
Thank you! I will make a tut on how to use images to get a style. Maybe will work on that today
was this video ever made? the titles for all the videos dont really point out what is being done
Hello, can you do the tutorial of the illustration of the video with Topaz?
Yea definitely will make another video
hey. when you put your keyframes(style)does it have to be from the video (png photos) or any style you want can be put it out there to generate the animation? Thanks
what software is that to export the video to png files so many
You can use any editing software. There should be an option to export as png images
@@enigmatic_e Thanks for replying
can you make an updated version?
Yea probably will
@@enigmatic_e tnxx
Great video! I was able to create a few seconds of stylized video from this tutorial. I get an error when checking Google drive files for my to Neural Style Pt and was wondering if you knew what the problem was.
Nice! Im not sure what that can be but I do intend to make a tutorial on Neural Style. You can get some pretty interesting results from it.
@@enigmatic_e ill keep a lookout for it and subscribe. Thanks for the reply
What did you use to export to images?
Premiere Pro or any editing software. When you export, you have the option to export as images by changing quicktime or h.246 to jpg or png files.
@@enigmatic_e I use VSDC so will have to check if that is there.
@@enigmatic_e I checked VSDC and it doesn't seem to have that feature but I believe VLC so will use that.
I never updated my ebsynth installation 🤣
Haha you have the same voice as John C. Reilly
😂 No one has ever said that to me but I guess there’s some similarities I hear.
Hey, my brother, you don't need to add Chinese subtitles, that will increase your work, I can understand it, I hope you can produce more tutorials for stable difiusion animation, the animation I did was not stable, my face would change, come on
You didnt tel us what you are using to export the video to PNG files... LOL.
Martin Barbara Rodriguez Kevin Gonzalez Angela
Miller Linda Wilson Amy Johnson Eric
you are confusing us with this tutorial, you can be more simple and direct. do you want to make money or learn something from you?
This is one of my very early tutorials so I agree it might be a bit rough. I’ve gotten better.
worst tutorial ra bunda👍
😂 im so confused
Thank you so much in sharing this info Brother and yeah it was great watching this video and it was neat and clearly explained, I will surely come up with a creative work and let you know bro ❤️
🙏🏽🙏🏽🙏🏽
@@enigmatic_e ❤️from India bro😉
Nice video man thanks. How do you get so consistent styled images to use as a reference. Midjourney has remix but it wouldnt be so consistent.
How did you make that frame with the AI, I wanted to do something similar and I couldn't, it would be great if you could tell me how with the concept you got that first frame with that cool style.
Thanks for the video.
I will make a video about it. I will put that on to do list
Do I have to have Motion Bro or can I use something else to export as png?
I typically use premiere or after effects
there's got to be a better way lol
There probably is. I gotta update this video someday.
Anyone know a good way for achieving a lot of keyframes? I painted 90 frames for more detail, but i'm having trouble figuring out how to even load that many into EbSynth. The most I can do at a time is 15 keyframes, any help is much appreciated!
An example using ebsynth th-cam.com/video/Smw4fKunI7k/w-d-xo.html
do you have to use the first frame 0000 to do the stylized reference? My first frame 0000 doesn't show the face of my talent, can I use any frame in the folder?
Two guys doesn't really equate to a bunch of people are playing with stable diffusion + ebsynth.
That's like Van Gogh style
I am coloring and styling an old black and white movie this way. It is so cool to have fun with EbSynth
👌🏽👌🏽
I appreciate your knowledge transfer and the content is awesome, but please plan out the video a little or cut out the parts which are unnecessary because it misdirects the viewer and they might lose focus.
Thank you for the feedback. These were some of my earlier vids. I’ve changed my format in my newer vids. Might update this one some day
Hello, nice and informative video, thank, May i ask you that what computer you are using or suggest for this type of work
thank you! My computer uses rtx 3080 10gb and Ryzen 9 3950X cpu
Man a tutorial to make that final look would be amazing!
i may give it shot one of these days.
Re img2img, you might have already figured that out, but if not, that's a Stable Diffusion option. It might only be in the Automatic1111 Web UI. I think it generates an animation between two image keyframes -- have not tried it yet as I'm still figuring out how to install SD with an AMD graphics card
Yeah I did figure it out. I did a couple of videos about it after this. Thanks for the info anyways 🙏🏽
looks $ick bro
AI shart.
If your video has audio, will it be automatically applied to your final video?
No I don’t think so. I believe you have to add the audio again.
thanks broo your going great
thanks bro!
how did your keyframes get setup automatically i have a many keyframes its hard to set up one by one
Put all your keyframes in one folder and then drag folder to keyframes. Hope that helps
What naming convention are you using for the key and video frames? Thanks
@@mikealbert728 when i export from premiere or AE i just name it 0 and it will number the frame 0000 to last frame e.g. 0120. Of course the style frame should match the original frame from which it came from e.g. original frame 0000 and style frame 0000.
@@enigmatic_e Thanks. I figured out my issue. I had the naming right but I had too many key frames. It seems if the amount of keys extends the window below the size of the desktop they just won't be added. Trying to find a work around
where does one change the export to PNG?
When you go to export there should be an option that says format. There you look for the png option and then export it.
Man! You are a monster teaching, I love the way you teach.
I have had bad results with the EbSynth and Stable Diffusion. My problem was that I tried to edit a video clip of a raper and I had constant failures in the consistency of the video. I didn't do the blending that you teach (I'll try it) but I didn't like the results at all. Maybe it was because only the face appeared in the video and there was too much movement? The truth is that I do not know and I am trying to do it in another way (several steps are from other videos that you have published).
As soon as I can I will join your patreon without a doubt.
Keep teaching please!!!
um.. did you mean a rapper? You left out a p.