How I Generate It
Conversations with people creating things with generative AI
How I Generate It
Chen Effective: Musician turned AI animator
In this episode, I talk to Chen Effective, an AI animator and musician who uses Pika Labs to extend 3 second AI clips into long morphing animations set to music. We chat about his journey into AI, how creating AI can be an obsession, and of course, how Chen generates it.
đź”— Follow Chen Effective
YouTube: https://www.youtube.com/@cheneffective1983
TikTok: https://www.tiktok.com/@cheneffective
🍿 Watch Chen's movies mentioned in this interview:
Beast https://youtu.be/p6PghRPpJ0Y?si=GlnldVKsroTcVznl
Evolution https://youtu.be/7TeOhwuukgA?si=uonGR4gC2HmlnaSk
Blendr vs Pika Collaboration https://youtu.be/7Av_pMIghes?si=A8Q1qxQGXkgDHUTS
Welcome to how I generate it, where I talk to AI, video creators and filmmakers. To understand their process, their inspiration, And of course how they generate it. Today, my guest is Chen effective. I first encountered Chen in P because discord server, where he had shared a really cool video. right after Pika one. Oh was released. know, there's two different kinds of morphine video. One is accidental and one is intentional. And Chen is really good at the intentional morphine videos using tools like Pika.
Mike (2):We'll talk a little bit more about that specific video a little bit later, but for now, Thank you for being here. Thank you for listening. Here's my interview with Chen effective. I am happy to have Chen Effective on the show today. Chen Effective, thank you so much for being here today.
Chen:Thank you for having me. Thank you so much.
Mike (2):It's great having you here. And I know you started doing, you know, your background is more in like music production and producing music. So what, got you interested in, AI
Chen:basically since I little kid, I was always interested in cartoons and animations and I'm drawing myself and I'm sculpting a little bit and I, I love art. In general, but how I came to AI, I want, I wanted to make First, some covers you know photos for my music, then maybe to make some clips for my music. But at first I was only learning Mid Journey and After Effects. And then when wanted to, to, to make something for after effects, and then I, I discovered Pika, PikaLabs, and when I start to start to use Pika for only for um, and how, how I say it, to see how it looks, my pictures in animation before I start to animate in after effects. And after I see what Pika does with my photos, I say. And I don't need after effects. I will animate on, on peak hours, platforms like that.
Mike (2):It's really interesting to hear that you started with after effects and then realizing like, you know, I actually like the way this is working more in these newer tools because I think a lot of people , feel the other way of like, well, this does cool stuff, but I can do a lot more in after effects. I have more control over like these kind of details. What was the thing that drew you to Pika versus After Effects?
Chen:The first time that I see video generation, it's only when Runway started. It was not so, not wow. But it was cool, but not wow. And then, after I started again to, to see Pika, And I see the the huge amount of progress and it's like it's I cannot compare with after effects because in after effects you do specific stuff that you need for production, but when you when we talk about movement of the person, for example It's not the same. I cannot move it naturally like Pika does or Hyper or Runway. Like it's actually a real person there and moving. So in After Effects you can do Post production or or effects to add to the Animation of pico or something like that because it there's stuff You just cannot do on after effects or any other program. Maybe only Unreal engine and stuff like that. Yes.
Mike (2):Yeah. Yeah. Because you were it sounds like the main way you were using After Effects was to try and find ways to animate images that you had created in mid journey. If you have a great, Mid journey image and you want the character to walk down a hallway or something I'm not even sure how you would do that in After Effects.
Chen:Yeah, it's not the same. It's
Mike (2):So the video of yours that caught my attention was when Pica 1. 0 first came out. They moved out of the Discord server only and they had their own ---=like web interface. When they launched it, they had all of these other tools where you could extend a clip and you could paint in one area and change an area of something. And I was really excited about it too. It was fun to go on the Discord server and see what other people were posting. And I saw one that you posted called Beast and I want to just kind of for people who haven't seen it I want to walk through it because I I think what you what you did and that was amazing one of the complaints people have about AI video stuff is like it's only four second long clips You can only do something for really short and and what you did was like it's one minute long And it's like This was five months ago. This is, way before Sora came out with their stuff. So putting it into proper perspective on the timeline of AI innovation. But it was the first one that I'd seen that wasn't done in stable diffusion. Or, you know, a comfy UI workflow or something. It was something that was done with a web based tool that had a lot of morphing to it. Like where it changes over time and then it was really long. What gave you the idea to even try that first of all?
Chen:To be honest, it was my first video long on Pika 1. 0, I did all my first videos I did in the better and the, the, my first videos are, I did always long, but it was not the same extension because in the old speaker, you extend with all final frame, you take the last frame and make a picture again, and then another three seconds. But it's not the same and it's not looking the same, but when pickup came out with this new extension, , this is the first thing that I need to do to see how it work without all this final frame sites or other tools to, to, to extend the video and it, and it just did amazing job. You just know yes, you need to know the parameters to keep it low So it's not morphed too much and not be weird But it's I after I see what it does. It's like whoa, it's The first thing that I needed to test how long I can keep it Because in beta when you start to work with the final frame and the last frame you slowly Lost color the color start to change slowly and you in the end you have like two colors and that's it If you see my first video, you will see that the end of the video. It's only blue and yellow. That's it
Mike (2):Mm
Chen:So here I see it's like you can extend and it's still Remaining the shape and the color sense like whoa That's a different game now Yes, I love to extend stuff and Change to from one thing to another
Mike (2):And I want to talk a little bit about like that last frame first frame animation style but I do want to walk through so this is So if you're for our listeners who maybe haven't seen this video, I watched it again this morning and I was like i'm going to take notes of what's happening So imagine you know, this is a one minute long video and it starts with a wolf on the edge of a cliff. And I think this was one of the original, like, their templates, right?
Chen:I take it from there. I take it from there. yeah, I take it from there. Yeah, I just take another Retry and it generate the same but a little bit different. But yes, it's from there
Mike (2):So it starts with that wolf on the cliff, , the camera's panning around him, but then one of the legs, the front legs of the wolf, lifts up, and as it lifts up, it turns into a human arm, and it's like, And then the other arm that's on the ground turns into a human hand and arm as well. It's like a hand on the ground now. And then the wolf just kind of turns its head. And while it turns its head it turns into a human face. And then the body becomes more human. And now all of a sudden it kind of starts to stand up and it, and the wolf has turned into a man. Like it's basically like a reverse werewolf, like where it's, you know, it's now transformed into this shirtless man with like long hair and pants and the cliff that the wolf was on. It's now just a boulder and the man leaps off the boulder and starts running directly at the camera. And when I first watched this, I was like, what? Oh my God, I've never seen anything like this. Cause like, we're like, you know, 20 seconds in or something at this point. And I'm used to like three or four second clips. So he's running directly at the camera and he's turning into a different kind of beast. Every footstep, there's like this explosion of sand and the sand is also maybe sand and water. And then you notice, well, there's actually waves crashing all around. And then the, the beast turns more into a werewolf and still running at the camera. And then it kind of pauses and you get those like Pika sunglasses, which was a really fun like way to end it. But it was like, it honestly was like, I had never seen anything done. using a web based tool like that before, like something that long. I I've seen a lot of morphing stuff with stable diffusion. The kinds of stuff that you could do, I think, with stable diffusion would be pretty amazing. Have you ever looked into to stable diffusion at all?
Chen:To be honest, I tried to play, I'm playing around with comfy UI. Yes. It's, it's not easy because I'm on Mac and then you have, you have lots of stuff that you need the NVIDIA. And, uh, Mac and NVIDIA, it's not not working together. So I, I, I do the simple stuff. I can generate, I can make from picture to picture or from prompt to picture in Conf UI, but not really big workflow. I'm learning. It's not easy. I hope they will do something in the future from to collaborate with Mac and stuff. But for now, yes,
Mike (2):Yeah.
Chen:we using the tools that can, we can you know, be above the comfy and maybe simpler like Korea now with the new tool. like like the the blending of the pictures and the frames it's uh, basically it's a workflow from confi UI but they give it to More people to use if they don't know to use confi UI and stuff like me
Mike (2):I want to really quickly go back to the beast one. There are a lot of people who try to do the longer clips, the way You did like where you take the last frame of a video, a generated video, and then use that as the first frame of the next video. And then you sort of edit those all together. And it sounds like you had been doing some of that, but that you found the. Extend four seconds to, to work better. Is that, is that right? Did you feel like that was the peak is extend four seconds works better.
Chen:It works better in different ways. It's the most important that see the motion. If a person like walking for three seconds and you extend it, he will keep walking. Pika see the motion and keep going with the motion. If you're using the rest frame. It's like a still picture and Pika have to figure out what you're trying to do in the picture. He's standing, he's walking, so it's not always walking, it's not always, so it's more hard to keep the movement go on. But when you extend it inside of Pika in the, with the new tool, so it's, just continue it. Anything that you put it inside, it will continue it like smoothly. It's crazy.
Mike (2):The other one that has that same kind of like long morphing stuff is the one called evolution. And that's like, it starts, I won't go as, as detailed in this one, but like it starts, you know, you're on, you're underwater and then you see fish and the fish turns into a reptile. The reptile turns into a monkey. The monkey turns into a caveman. Then we're in ancient Egypt and Vikings. So there's just like, you know, you, you bring us from., the primordial days up until the future where there's, robots and alien life forms and I thought that was really cool. And you do a lot like with the camera movements too, which you know, it had that very much of like a stable diffusion kind of vibe where you're looking at something, it changes and then the camera angle changes too. I could tell you'd been playing around with the tools a little bit more from beast to evolution. What were some of the things that you learned making evolution?
Chen:Making evolution, it starts every project is started as a challenge or in Pico or so, and then it's just give me some a start of an idea and then I just keep going and going with the evolution. I wanted to make like Good evolution video and to see really in this one. I really want to see how push can I get one show one big, big short and how I can handle it. And then after I, after I finish it, I just realized that I can do it much better because I did it with the last frames. And I didn't think that I can take the last three seconds of the video, put it again inside Pica and continue it will continue the same motion and stuff that I just figure out after that. But it was yes, I mean, you make 15 seconds, then you put it in the last frame, and then you put make another 15 seconds and then you put another list and then just keep going but it take so much retries really. Okay. It just cannot add so much retries in every, pick a lot of patients, every AI in general, a lot of patients, you can generate, you will get what you need, but you need to generate and retry and retry until you, because for one, three perfect seconds, I need to generate at least between 20 to , 100 cuts to choose my perfect. Three seconds only like that after if not, it will not look like that.
Mike (2):Yeah.
Chen:for sure
Mike (2):And are you doing these all in Pico one dot Oh, or are you kind of going back and forth between discord and Pico one dot Oh,
Chen:No, no, no, i'm not using now pika discord. I only used only 1. 0 And but it started in the start. I was using only pika And then I start to help with you know, with runway because they, they have the brushes that sometimes you need the brushes. You cannot handle without. And I love hyper too, actually. It's not it's not bad. It's really good for stuff and even consistent stuff too. But Pika is my, is my, like, I started there, I, I know, like, every corner inside for, for now. For now.
Mike (2):You mentioned the daily challenge. I know, like you've, you put out a lot of videos on your YouTube and it sounds like a lot of them are from the daily challenge or the weekly challenges maybe in Pika it seems like you're really part of the discord community there and being involved what's the big thing that you like the most about the pika discord community?
Chen:I tell you I started because I was after I after the Covid started I I have a I was without a job. And then after that, I was some surgeries. I was left months in home without working. Then the world started. So I have like really long time. I didn't work and I was 24 seven inside discord 24 seven. I tell you, I was sometimes forgetting to eat and sleep. I was losing, I lost in two months, like 10 kilos or something like that, because I, I cannot control when I create, I forget. So, inside the, of Discord, of Pika, I see it was early, it more, more, more a beginning of them. It not, not really beginning, but it was not so much people there. And I see what people do, and I love that people talk when they find out something new and share. And I found out and people asking me how you do this and how you do that. And I just love to talk about art in general. So I, I love the community and I had like a lot of time to be there. Now I just a little bit less because I start to work a little bit and I have no much time for too much hanging out there. But This is how I start. I just was there 24 seven and people always talk and I love to talk with them too. It was really fun. The crew is really good. I love the crew too. They're really good people.
Mike (2):Yeah, that, that was the one thing I really noticed about Pika early on was like they, they threw so much into the community with those daily challenges and just with being very responsive to people helping out. Pointing people to the right resources and stuff like that. If you have questions, you can, talk to other people who have those same questions and figured it out. So it's great that you're in there and helping other people and that you're also there and learning from other people, your, your collaboration with someone. I forget the person's name, if you could kind of explain how the collaboration started and then what the, what the video was like, what you were given and what you did with it.
Chen:It's actually a friend of mine that I met in discord and we start to chat and he, he showed me this video clip that he make he started to learn blender a little bit and they show me this car with music and I just love how it, how we make the car and and he say, yes, it was experimented in blender. And. Without asking him, him first, I just download the video, it take a few seconds and put in Pika, just for myself, and then I see how it reacts, and I say, bro, let's make a video, you, I, you, like, it start with Blender, 100 percent Blender, I just Take the whole car that he make and just extend it after a few seconds. I extended it and it starts to Become pika and the car starts to be transformed into transformer In the end And this is how this project born. It's actually need to do more like this. Yes
Mike (2):so the transformer that showed up, was that something that he had created in Blender? Or is that something that you were able to generate in Pica?
Chen:it's pika. It's I he's a You He's, he's, he's part, that car is driving, the camera is going around the car, and then the camera stops in front of the car, and from that, it's speaker. The car is just riding forward, and slowly transforming to transformer. This is all pika.
Mike (2):that's really impressive. I was wondering, like, cause I saw it was a collaboration. I was wondering if you had done something with last frame, first frame with, you know, a car and with a transformer and I was really curious, but it's really impressive that you had done all of that from the beginning of the car, to the transformer, all in Pika. So very cool project. Cool collaboration.
Chen:Thank you. Pika is actually always surprising me. This is why I love it's always surprised
Mike (2):A lot of people who might be listening to this would say like, how do you not run out of credits? Or are you part of their creative, do they have a creative partners program or something that you may have been
Chen:I wish I wish. But no, I have I started the, at first with the pro plan and because only at the pro plan, you can have unlimited generations. And then they make a better plan they met they did unlimited plan. It's the It's a it's the middle one. It's something like 35 In month and you have unlimited in chill in chill generation, but it's unlimited so otherwise you you I will lost every one of my credits in the first day if I don't have this plan
Mike (2):That's what it sounds like in 35, like for some people that's like, that seems like a lot of money, but for other people, it's like, you know, you look at what you pay for streaming services or music services and things like that. And I think it sounds like for you, I bet you don't even. Have time to watch streaming services, like to watch movies anymore cause it's like, it sounds like you're just spending, you know, I think this is something a lot of, that a lot of, people are doing, they just spend so much time in generating AI that they don't do, social media or watch things anymore
Chen:It's, yeah, I tell you, I don't, I don't have much time to see nothing. And yes, it's not that it's not expensive. It's a little bit expensive, but if you, yes, if you see the other stuff that you pay in life. And this is what you love. Some people love just to buy stuff for their car and they, they, only spend money for the car and that's it. But this is what I love. I love art. I, I make a little bit money maybe from this here and there, but most of my life. I just spend money on art like a crazy one. Sometimes I not understand myself But this is me. I'll just love art and As long it will not hurting me in my life and not you know Not I not have to choose between eat or do art. Yes. It's okay Audio. Yeah
Mike (2):and it sounds like This kind of AI came across your life at the exact right time where you needed a creative output And I feel like you can be endlessly creative with Generative AI. I know you were doing music and I want to talk to you really quickly too about some of the music things like Suno what has it been like to be a music producer to now be like looking at AI generated music?
Chen:At first, when I heard about AI music, when I, before I started to use AI, yes, it sounds a little bit weird, especially for singers and people like that. But it's not. Well, after I started to use it and I see that some, sometimes you have like a, a, creative block and this kind of tools can spark idea. You don't have to use them. You can use them. I love to use them and make my own layers on top. If I, if I can make, I will do that. And in the end you, you like building layers on, on top of the ai and in the end, UTG can just erase the AI and you left with your own stuff. That, I don't know, it's, I love it. It's really open your creativity and I don't always have the idea to write music and it really help, especially now with how it's become, whoa. And it's like, whoa, crazy. I love it. I love it
Mike (2):Yeah. It seems like it can really let you experiment quickly and try different styles and try different things. And when you're creative, it seems like that's what you want to do. It's like, , find, find a way to express yourself and you. you connect with something at some point and say, yeah, this is the direction I'm going to go with this, , and now you've got a tool that you can create with a new tool that you can create with.
Chen:Yeah, especially that you don't have budgets for to bring some singers or other players It's really good. It's in general ai it's like for simple people that want to create But they don't have all these studios with money. Why not? It's it's it's this is how it's supposed to be You don't have to limit people because they don't have money or something You
Mike (2):Yeah. That, that, that's what I like about it too, is like, I feel it has the potential at least to really open up who can create really cool high end, music videos or movie trailers, whatever it would be the The tools of production aren't like the barrier anymore because like you said, you know for 35 You can have unlimited video generations I forget what it is for suno, but you know for 15 or so like around 20 You can you know generate music so it seems like that's still a barrier for some people, but I feel like completely When you compare the costs for other things, it's, it really opens the door for a lot more people to, to get in this like world of creativity that they weren't
Chen:Yeah. especially Especially you want to do it professionally and you're trying to do it professionally at first you You have to spend before you earn it's always like that If you really want it, you need to spend a little bit at first one month I realized that I have so much subscriptions that I'd start to delete half of them. Yes. It's too much because it's 5 here, 5 there, 10 here, 10 there, 30 there. And at first, when I was paying 70 for PICA, it was, whoa, too much. And then yes. And now I have my three or four subscription and that's it for now, it's okay
Mike (2):And I feel like that the key to a lot of that you mentioned hyper too, which I currently, I believe they're still free. I feel like when a lot of these tools come out for the first little bit, they're, they're free while they're trying to figure things out and grow a user base and then they switch to paid plans. So I, I feel like that's a, a thing to look out for, for a lot of people too, is like, see what tools are out there that are brand new that you can use for free, you know, for a month or however long it is until they switch over.
Chen:Yeah, but I tell you that they are keeping their beta versions free So if you don't have you cannot pay on pika you can use I can use it pika beta even now and do crazy stuff. You just need To, to know the parameters and how it work, you don't have to pay for a, you know, if, if you cannot, I, I think only maybe runway now it's don't have the free version of them in anywhere, but I think hyper and beaker will continue to be free and better. Yes. You have the logo, but still. You have the same quality of the video and
Mike (2):I want to ask you one other question that this comes from from people commenting on my YouTube. I think some people get into this and they want to create a video and they want it to go like really big and really viral or whatever. When you spend a lot of time generating a video. and you want to reach a big audience and it doesn't hit the views what's the thing that like motivates you? Or is that even part of your, you know, what motivates you?
Chen:I try to to see how it works, all this view stuff, because I know if you don't have a channel that Reach at least 1000 views. The views are slow at first, after 1000 subscriptions, it start to go better. The, the stories is really good for this. I, I noticed because if I upload the video in, in the channel in the regular way, I can have maybe. You know, 100 and something, maybe less, but when you put in story every few days it's like at least 10 times more views. I don't know why, but in stories it's much more. And then you can like a play a little bit in story and then little bit in the, the regular channel and then like that, but motivates me when I see other people do stuff. I just motivated by watching others. That motivate others. Yeah. It was not my plan to open a channel and to start. I see. Let's see what happened. I will, because I'm making videos anyway. Let's see if it work. It work. I'm not crying. If it's not working with the views.
Mike (2):Yeah, it's more about the creativity I imagine than the views. The views are nice, but it seems like the views are like a way of sharing it, maybe.
Chen:Yes. If you keep doing it like every few days, it'll work after. One first 1000, it will work. I hope sooner we'll reach and let's see. I will tell you for sure if it's working.
Mike (2):great. Is there any advice that you might give someone who might just be starting out, anything like that,.
Chen:Advice. Yes. I always have advice. It's first of all first of all, patience and anything. If patients don't rush to end because sometimes I rush to. But don't rush if you see something that's almost good. It's almost good Try a little bit more. It will be better and About the Maybe about the generations It's always if you have the option of the choosing the speed of the of the parameter of the generation You know the motion speed and stuff and the consistency Always on the low it will not fall apart when you try to extend or you're trying to make some movement inside of the video Low parameters. It almost work Give it a try a few times and you see it will not Fall apart and with not too much blurry and stuff And just keep experimenting. I just always experiment. I see some new tool. Let's see how it work with other tool. Oh, let's see. Let's see how it's blend with, you know, with other stuff. Always experiment and be patience.
Mike (2):Yeah, that's great. So it's, it's, I like that advice too, about like keeping the motion parameters low, because I think the, the temptation is like, I want there to be a lot of movement in this. And so you'll put those motion parameters really high. And I think that's, like you said, exactly. That's what I've noticed too, is
Chen:tell you
Mike (2):just get really out of control.
Chen:I tell you why people doing this why they think because they don't think that after they Make the cut they can make it faster after that They don't think about that, you know, because when they move slow or walk slow or everything slow, it's not matter in case that you need, you just keep it slow motion. But if you need faster, you just make it, yes, it maybe will be a shorter, but it will be faster. So it's, uh, and if you
Mike (2):Yeah. So you're, you're generating it with a lower motion parameter and then you're getting slow motion. But then when you download all, when you extend all of those to 15 seconds and then to a minute or whatever, then you'll speed that up and it happens more quickly. Is that what you're saying?
Chen:yes, if you just use the part that it looks very good and that's it, you don't need all the 50 seconds that everything that happened, you just need what you need and that's it. And I always, If I generate stuff and that I not need them, I keep them maybe later I will need, because if you see something, not perfect. In, in some other case, it will be perfect.
Mike (2):Well, it has been a pleasure talking to you. I love that you're very experimental and very creative and it would be great to check in with you somewhat somewhere down the future too and just see like what new tools are you using and hopefully they'll figure out this stable diffusion thing on Macs and we'll be able to be play around with that someday too.
Chen:Thank you. Thank you so much. It was really really fun to talk about all this stuff and Thank you for having me. It was really really good and I really Enjoyed to hear your next podcast
Mike (2):Yeah. And, and where can people find you? I'm going to put links to all of them as well in the description. But what's the, where, where are you at besides YouTube
Chen:the best place is YouTube because all the videos is there and in TikTok
Mike (2):Well, thank you so much for being here.
Chen:Thank you. Thank you so.
mike_1_05-24-2024_075845:Thank you again for listening. If you enjoy conversations with AI creators like this, don't forget to subscribe. And until next time. Keep generating.
