It came out looking like a high quality promo video. There is a long way to go getting AI to do motions that are coherent but its getting really good. Right now working on something in LTX Studio, I am impressed with the storyboard flow and the control you have
Kling has issues. I had access and made a ton of videos, then one day it said I couldn’t log in and I’ve not been able to make anything or get through to support since then. They want me to verify my account with a Chinese ID number that I obviously don’t have, and even pointing this out has gotten incredibly hostile responses from the community.
@@ramanandh1261 I requested it shortly after this channels first video about it using WeChat as the login on KwaiCut. I had access for a while and made a few videos then one day I couldn’t login. So even if you get access they will try to collect your information and may restrict it again based on your nationality.
No one has success with Kling except people with real chinese number. I tried different virtual numbers for 2 weeks and that didnt work- either the verification code doesnt arrive by inputting the code you have a error
1000%! Thank you! And yeah, I think it's important to maintain that transparency! Lots of either snake oil, or some hand waving (esp w/ cost, which can add up)-- That said, on the cost front: I mean, when you really look at what you can do vs a traditional pipeline-- I mean, even if you were to double the costs of the AI version, you'd still be thousands upon thousands less.
I went ahead and uploaded a full AI video to RUclips just was curious how hard it would be. I’m a mine site worker so the video is an educational video about working on a mine site. The video is called: Career Progression in WA's FIFO Industries Growth and Advancement
Oh, that's excellent to hear! I'm actually going to be putting together a "starter video" pretty soon-- hopefully, that'll help folks enter a little easier! I know it can be a bit hopscotch on youtube!
I jumped on here to do my usual "Slam Tim for his overly ambitious video titles" , but damn bro.. For once you didnt exaggerate. Thats some promising work. Yarrrrrr Shiver Mi Timberrrss 👏👏👏👏
9:09 You could of zoom in to the pirate leaning forward toward the camera (sped up) then cut straight to a shot the pirate with the pirate clip that has the mouth that was talking gibberish, but trim the shot to show his face only from the top of his moustache upwards and his menacing eyes in the shot! That way you don't see the mouth moving talking gibberish but you can just get some movement just below the nose so you can see that he is talking (you could slow the gibberish talking pirate clip down, so the top part of the moustache moving isn't moving too fast), and you get a great menacing close-up of his face and dark menacing eyes too. Badaboom badaba.
One thing I would change is the way the voices are done. It sounds like a recording studio and not like the voices are actually in the video. May need some light reverb and a little panning.
That's a super valid point. I'm not the best at audio engineering, and to be honest, I'm always a little (personally) worried about L/R panning too hard. But, great note: I do have the Izotope visual mixer, so maybe I'll give that a run through next time around! (and more reverb!)
Your most useful video to date. Well done! So good to see we can move past trailers into actual stories. And thank you for the breakdown of your process. Much appreciated. Will show to my students! Please remind your audience that it's still about story, not tech. Thanks!
10000%! This really does serve as my "call to arms" to start focusing on actual narrative! Trailers are still fun and all, but really, we should start crafting new and interesting ideas on narrative at this point. It can be done, and we have the ability and resources to get in on the ground floor and set the standard for this new medium!
Fantastic work, brother! I am both impressed and thankful to see that the level of work I am doing right now is at least in the same ballpark with what you've done here. That's heartening! You did some great storytelling in this, and that's really half the magic. Good storytelling sells it! Love the genre, love the art. Nice work! Thanks for the entertainment!
Lately Ive been coming back to pirate seas through your short and Our Flag Means Death. I appreciate the masterclass and it goes to show a person still needs to devote time, energy, funds and creativity to make your freak flag fly. I was also reminded that in The Lair Of The White Worm, Ken Russell did the horrific dreamy parts in video resolution which was jarring and f-ed up 😂. Rock on, brother!
Oh, I gotta get caught up on Our Flag...funny enough, I've been marathoning through What We Do In The Shadows...so, maybe that's where the vampire come from? Haha
Thanks so much!! At some point, I need to circle back and do a full on "beginners" tutorial. I do feel at this point there's SO much info out there, that someone popping their head in right now is likely to explode!
Appreciate that so much! That's something I like to do from time to time-- It's always tough when you're reporting on the newest thing, and realizing: "Wait, I don't have time to play with the thing!" Going forward, I want to make sure that I make a "thing" at least once a month. Keeps the skills sharp, and (to some degree), I think it's important to show, not just tell!
Great video Tim! The way you explain it makes it sound so easy, but I'm sure quite frustrating, and a lot of man hours! However, a perfect example of how creative you can be using A.I. Steve
Haha, yeah-- I was joking around earlier that this is the "baking show" version. Y'know, when you're watching a cooking show and they're like: "Just add this, and this, and this, and voila! You have a perfect meal!" That said, it's always fun. Even the messy parts.
Ha, Tim! Cool look behind the scenes. I kinda guessed that this was made before the new lip-sync programs (but they may cause problems with aspect ratio?). You know what would be cool... to make this exact movie again one year from now! To show how fast things change. The possibilities with "face-acting" and lip-sync with Life Portrait are gonna lift this to new heights already.
So, I did have a bit in there that I cut for time-- but basically, the idea would be to take the 1:1 output, and then you'd have to mask it over the 16:9 footage, and scale/keyframe it to match. That, as I was in the middle of shooting, is sort of a tutorial unto itself-- and considering we aren't far off from LivePortrait and the next generation of Lipsync tools, it didn't make a lot of sense to spend 5 minutes talking about it!
Fantastic work Tim! Great to see this format coming from you :) looking forward to seeing more of that! And never give up on those reflective comments. That woman's restaurant rating gave me my daily morning loud laugh 😂
You, Tim, are informed and funny and articulate and talented and give great advice which is why your channel is my favorite. (I'm sure lots of people told you it's Dream Machine rather than Dream Factory)
Haha, I could just see this lady stepping into the Pirate Tavern and being all "This is NOT what the pictures on Google looked like. The food had BETTER be good!" 100% she left a one star review!
I’ve dropped my mid journey subscription down to the $10 option and it’s actually still more then enough. So you could save $20 there. Just a suggestion.
Totally true. I started at the $60/mo level, but ended up dropping that down. That for sure was overkill for me. I might be able to get away with the $10 plan, but I do go through spurts where I generate a LOT. Most of that is due to the channel though...I'll say, I'm good w/ $30 since MJ is a daily driver for me.
Don’t disagree. At some point I need to do a beginners AI Editing video. I wouldn’t want to do a full fledged editing tutorial, as there are plenty of those online, but I think one focused on new AI filmmakers might be worth it.
Oh, that's EXCELLENT to hear! I tend to (obviously) hear most from people that didn't like the sequence, but I knew there would be a few people that liked it! Thanks so much for that!
Great brekdown! Your work is so important! I've been directing for 30 years and these tools are the best things to come along since NLEs. Could you possibly do a comparison video wit Dream Machine and Gen3?
Hey Tim, great job, I have been following since December and you have been a great point of reference for my own youtube AI/horror channel. I loved Dead Sea and I would love to be able to create something similar for my own channel, so this video is really helpful to realise how much in subscriptions I would have to spend, and I am pretty far from being sustainable with my small channel but I still appreciated the hard work you put on a daily basis!
Having also played around with all these models for my channel over the last few months, I totally feel you. It's like "playing a game". Each tool, at the moment, seems to be good at 1 or 2 specific thing. I also got some pretty good results throwing Haiper, Pixverse and Luminosity in the mix ^^ So grateful for your content!!
I did have a bit in the video that got cut around the part where I talked about not writing a full script. Basically, I'm not sure writing a script (at least in industry standard) is in anyone's best interest right now, as it'll lock your thinking into a particular shot or imagery-- which, (currently) the generators might (or might not) be able to pull off. It's currently a game that feels a bit like free sculpting. You have something in mind and get to work, and then discover the piece in the process. Not to say you shouldn't write anything out, but at the same time, I don't think you need to fire up Final Draft!
@@TheoreticallyMedia I totally agree about how a script can lock our thinking into a particular image or shot, and it gets frustrating when the tools cannot replicate what the script suggests. Your notion of "free sculpting" is a better approach. I made the mistake of writing a script first and got frustrated. Better to have a good story idea and see what images emerge via the tools, and then sculpt the story from there.
@@TheoreticallyMedia I found the same for my project (my first ever!). I wrote a script but modified it along the way based on the final sequence generations I was able to get. Since it was my first time making anything, I had no expectations going into it, so it worked out. 😆 It was a lot of fun learning the tools and it's encouraging to see I followed a similar workflow to you just by trial and error, lol. Thanks for the video!
@@TheoreticallyMedia That's actually sort on. Sometimes I find myself fighting 30-40 minutes with 4 secs of my video because I'm stuck to my initial script. Next time I find myself and that situation I'll try an adjust the text rather than the video. Appreciate the feedback buddy!
Appreciate that! Always the goal with this channel! Inspire, Educate, and Entertain...well, your milage may vary on the Entertain part. Depends on your threshold for dad-jokes!
Mind boggling that you can create something in a couple of days and pocket change, that would take weeks and 10s or even hundreds of thousands of dollars. Super impressive and an exciting harbinger of where we are heading .
Aren't we going to speak about the Bram Stoker in the room? Without this little Irish lad we would have never see the birth of "Dead Sea" and personally "The Captain's Log" may be one of my favourite part in Dracula. A strong story is spine on which a movie, series, animation,.. Etc is based but writers are very often overlooked and cheap, low quality writers are now often use. We all know the state Hollywood is in right now and many of us remember what Hollywood use to be. I truly love your spin on the "The Captain's Log" and to see the time and money you spent on the project blow my mind (obviously your expertise shorten the time🙇♂️). The future is bright for AI artists 👍
Congrats on the exposure & hitting 100K! And thank you for generously sharing your process and price breakdown. Do you think you could do this without Runway and use a free upscaler like Viva? I can't download Kling in the US - it might be my old Android device, lol. Looking forward to your tips!
Nice video Tim ! I have been using Dream Machine to generate scenes but it's kind of a hit or miss app. Sometimes I can generate 10 or more videos to get some video that will work for a scene. Other times I nail it on the first attempt. And this is for image to video generation. This drives up the cost of production for everyday use. Maybe they could include some sort of preview so you can see what you get before actual generation of the video. Thoughts?
So, I've made a few AI trailers. You seem to have great success in marketing and showcasing your gen AI content. Do you have any tips for someone trying to get their AI content out to festivals and in front of large audiences?
To be honest, I love them all. LTX has that great creative workflow, where you can get a birds eye view of your whole film. Plus, it serves as a great brainstorming engine. I rarely take what LTX first offers (that's a recurring theme on the channel), but I love that it always gives me a great springboard to try things out. Currently, I'd say Kling is the best Image to Video on the market. Although, it has a lot of access issues, which is a real hamstring. That said, I haven't seen anything yet that has its level of coherency when you provide it with an input image. That said-- I mean, it's Tuesday. That could change by Thursday! (Luma's dream factory is a close second on that front!)
@@TheoreticallyMedia yeah, I haven't been able to access Kling myself (hence why I was curious if it was worth the effort). LTX is interesting as far as an AI with some built-in consistency so it's cohesive. But it often doesn't pan out as intended. I prefer to simply create the media and edit it like you do, since there's more control on the backend. Thanks for everything you do on your channel - it's a great resource for understanding the latest and greatest, which, as you mentioned, can change on a daily basis!
The costs are great but what about your labor fee 😢 your time is worth something! Thanks always for another great educational upload :) *edit: just for clarity because people tend to not be able to read between the lines or want unnecessary attention, yeah you did mention it but it’s always worth to mention more about self worth on anything people do :)
oh, 100% I did have a longer bit where I did more of a "my cost" breakdown-- but then I thought better of it, since that might come back to bite me in the butt! "In that video you said you cost X amount!!" I may actually do a breakdown of what folks should charge at some point though!
@@TheoreticallyMedia The answer would be to break down the tasks and say that you spent 5 hours working as a video editor and 3 hours working as a script writer, and then use glassdoor or something to get an average pay for those positions.
Currently, if not Kling, I'd say Luma's Dream Factory isn't far behind. You do need to experiment w/ that Enhance Prompt feature though. Sometimes it's great-- sometimes is goes SUPER insane. (see the "walk the plank" segment as an example!)
If people are starting to learn movie making with AI then you still need to sketch up the concept and the lore. Of course AI can help you with that, but it takes a human mind to flesh out the sketch into a real concept to where the AI can learn the concept so it can create better results from that concept. So scripting and storyboarding may still be necessary.. And also, these LLM models may have to learn to generate actual 3d interactive environments..
I actually thought about that, but currently Hedra tends to get pretty wonky when you don't have a "head on" character-- although, I have been told that won't be an issue in the next version. I think I still could have attempted it, but that would have required masking out the mouth movements, and then keyframing the head movement. It could be done, I think-- but it also would have added hours of time. I think if I had another 8 hours or so-- maybe? But at some point-- you just gotta ship! Next one, I think!
What fascinates me about AI creations so much is that it "thinks" "outside the box". Ok it doesn't really do either, but if anything manages to "think" outside of human constraints, it's AI.
I’ve always considered it to be a really weird collaborator. And one that might have gotten drunk before showing up to work. But sometimes that drunk has some golden ideas!
Question about new website for Kling when You get a chance could You help us to understand and take us through each how to get up and running I mostly use Web on the desktop not much mobile too small of screen for my tried eyes
Haha, I know the feeling. Check out the video I did here: ruclips.net/video/CfTnMXodtns-/видео.html- you'll need to set up a Kuaishou account (the mobile app) regardless. Then when you log in to the website, it will give you a QR code that will open the app. From there-- It MIGHT work? I'm not sure if there is a waiting list for the website or not.
I mean, yeah- I didn’t want to get pandemic about it, but that’s true. There are some basics that I’ll take for granted. As far as the learning part, I kind of cover that by talking about what would be my daily rate. I didn’t want to hone in on that, mainly because people at all levels watch these videos. Never want to discourage people from jumping in and starting to forge their own path!
THANK YOU SIR 🙏🏼. I'VE BEEN WAITING FOR THIS VIDEO SINCE I SAW DEAD SEA SHORT. I would love to see if you could pull this off with using either Free AI tools or at least the Free plan . Additional bonus if you can do it on an android phone. If anyone can pull it off I believe you can.
Just extrapolating from your experience, with increased productivity from doing this all the time and likely improvements in the technology, it's easy to imagine how a single artist might produce a half hour episode in a month the same way they might a single comic book
As a self professed “comic book guy” I couldn’t agree more! There’s an obvious ton of pushback on AI in the comic book world, namely amongst the fans- which is kind of a shame, considering the grind that most artists in the industry face. But you’re absolutely right, if I could pull this off in 4 days, I’m sure we could extrapolate that out to a 10 minute film easily. Combine that with a small team, and you have a 22 minute monthly film/comic no problem. You could even take a (ahem) page out of the old Manga phonebook style. Multiple ongoing teams/stories released as a two hour block every month. It’s wild…
Haha, yeah-- I knew it was a wild swing. I did get a few comments from people that DID like it-- but, ha-- for the most part, viewers didn't love that section. Oh well, I still stand by it! Glad overall the production side was helpful for you!
While the looks of it is great, there is no comparing with what is being done in real movies. There's no control ho the scene itself will be acted, where everything will be placed and the production itself is of missing to make it feel consistent. Again, it looks great and 10 years ago there was no such thing at that level, and still, it's a computer doing its best to calculate what will be working in the universe we are in. It is not there yet.
I'm making an AI movie, already 24 minutes in. It will get better with time, because you can just keep getting better takes. Your movie is close, it will get better too.
firstly Congrats for 100k subs, i also make ai movie trailer, like you, i had made the ai trailer but with 0 dollar, without sending money, but it require little bit time and editing, due to free tools, if you dont mind, i can send you the trailer which is made by free tools in 0 dollar
For close up piece to camera talking you could have used Hedra or another TTV app, there's a really good one now on hugging face. Also don't mention 'coffin' instead describe it as 'wooden casket used to bury deceased humans safely' or something.
@@phizc its called Liveportrait, its opensource. Sorry I didn't mention that in the comment... what am I like? There are a few other vid on YT covering it I think. And its on a few spaces on hugging face to give it a go! It might run on colab? Maybe? Also its existing video, to video output via an uploaded image... eg. Image of your dog, video of Margerat thatcher saying something... output your dog lip synced to that... or a person... or two or three people...
I’d say roughly. I got the idea on a Thursday and started generating some images up. Over the next few days, I was creating the Midjourney images, and funny enough, single Kling has a mobile app, I’d generate the video while I was running errands and waiting in line at various places. I only did one full 8 hour day on it for the edit on the 4th day. It was interesting, while in “production” mode, I’d have a good amount of time between generating video, and that really let my mind wander to decide what the next shot was going to be. All in all, a pretty stress free production!
@@TheoreticallyMedia wow congrats 🎉. I imagine if you created a small production pipeline of people who did it all day every day the man hours would go way down from even that!
oh yeah-- you gotta release a little of that right now. At least in its current state. I tend to think of it as working with a crew of toddlers, a drunk cinematographer, and super uncooperative actors. It's basically creation through chaos. Haha-- Still, it can be a lot of fun!
@@TheoreticallyMedia I don't think text to video will be tge future of cinema you cant fit enough detail into wording or should I say most people can't turn what there thinking into a single prompt the new ai video company Odyssey has it right
@@JaxonBeckett I'm actually think a combo of Image to Video and Video to Video is the workflow of the future. I've got something on that next week that I think will open some eyes! Stay tuned!
Congrats on 100k subs! Well deserved 🎉🎉
appreciate that so much!! I'll do a little celebration video when I get the thingy!!
Agreed this the Best AI news channel 🎉
Agreed, keep up the great work Tim!
@@taoprompts I wish the same.
Clickbaiters dont deserve 100k
Oh, and Congrats on 100K Subs......... you deserve way more! 👏 👏 👏
Brilliant showcase of your workflow. Thank you sir!
It came out looking like a high quality promo video. There is a long way to go getting AI to do motions that are coherent but its getting really good. Right now working on something in LTX Studio, I am impressed with the storyboard flow and the control you have
I could see that. We’re pretty close at this point. Maybe not photorealistic, but for sure good enough to tell a coherent story!
@@TheoreticallyMediadefinitely!
Kling has issues. I had access and made a ton of videos, then one day it said I couldn’t log in and I’ve not been able to make anything or get through to support since then. They want me to verify my account with a Chinese ID number that I obviously don’t have, and even pointing this out has gotten incredibly hostile responses from the community.
@@ramanandh1261 I requested it shortly after this channels first video about it using WeChat as the login on KwaiCut. I had access for a while and made a few videos then one day I couldn’t login. So even if you get access they will try to collect your information and may restrict it again based on your nationality.
How you got access, can you please tell me!!?
Same. I had access and now it says I'm back on the waitlist
No one has success with Kling except people with real chinese number. I tried different virtual numbers for 2 weeks and that didnt work- either the verification code doesnt arrive by inputting the code you have a error
@@artemedutech you don’t need a verification code. It’s just a waitlist thing once you have your account setup.
Congrats on 100K! Very well deserved. I rely on your channel to keep up with AI and love being part of the community!
Dude!! That was excellent! I’m glad to see you pushing towards narrative storytelling! That’s were I wanna take it! Thanks for the inspiration!👏
I am sure that the BLOOPERS of every single Short Film or Movie that used A.I. to be made, will be one of the most funny things to watch ever, lol!
Hey man - you f'ing rule!!!! Thanks for all the transparency; keep up the great work!!
1000%! Thank you! And yeah, I think it's important to maintain that transparency! Lots of either snake oil, or some hand waving (esp w/ cost, which can add up)-- That said, on the cost front: I mean, when you really look at what you can do vs a traditional pipeline-- I mean, even if you were to double the costs of the AI version, you'd still be thousands upon thousands less.
Congrats on 100k🎉 fully deserved and looking forward to congratulating you on 200k!😁👏🔥
Thank you sooo much! One at a time!!
@@TheoreticallyMedia 💯💯👌
Congrats on 100k!
Thank you so much! And thank you for being one of them!!
Finally! Action scenes! I'm really getting excited now! 🏴☠🏴☠🏴☠🏴☠🏴☠🏴☠🏴☠🏴☠
100%! To be fair, the action scenes are still a bit wonky-- but, at this point: Usable-- and only getting better!
I went ahead and uploaded a full AI video to RUclips just was curious how hard it would be. I’m a mine site worker so the video is an educational video about working on a mine site. The video is called: Career Progression in WA's FIFO Industries
Growth and Advancement
Congrats, Tim! This is revolutionary and some day we will look back at this and see how it all started. Thanks for the great content. Keep it up!
Congrats on the big 100K Tim.
Thank you soooo much!! I still have to do a community post on that, and we'll have a fun video when I get the thingy!
I'm brand new to this and your content is accelarating my understanding very quickly. Excellent effort, thank you.
Oh, that's excellent to hear! I'm actually going to be putting together a "starter video" pretty soon-- hopefully, that'll help folks enter a little easier! I know it can be a bit hopscotch on youtube!
Amazing job !
Big fan of your RUclips channel.
Cheers
How do you get style consistency with text to video in runway?
Wow!!! Congrats on 100k subs!
this was GREAT! Worth 5 thumbs up.
know the reality and details is invaluable!
Was pretty close to my WAG...
Thank you so much!! And can I borrow a few of those thumbs?
A student observing the effort put into this master piece. 😮
Thank you for that!!
I jumped on here to do my usual "Slam Tim for his overly ambitious video titles" , but damn bro.. For once you didnt exaggerate. Thats some promising work. Yarrrrrr Shiver Mi Timberrrss 👏👏👏👏
Ha! See, every once in awhile, Everything Just Changed!
Congrats on 100k subs
Thank you so much!! still kinda pinching myself!
9:09 You could of zoom in to the pirate leaning forward toward the camera (sped up) then cut straight to a shot the pirate with the pirate clip that has the mouth that was talking gibberish, but trim the shot to show his face only from the top of his moustache upwards and his menacing eyes in the shot!
That way you don't see the mouth moving talking gibberish but you can just get some movement just below the nose so you can see that he is talking (you could slow the gibberish talking pirate clip down, so the top part of the moustache moving isn't moving too fast), and you get a great menacing close-up of his face and dark menacing eyes too. Badaboom badaba.
One thing I would change is the way the voices are done. It sounds like a recording studio and not like the voices are actually in the video. May need some light reverb and a little panning.
That's a super valid point. I'm not the best at audio engineering, and to be honest, I'm always a little (personally) worried about L/R panning too hard. But, great note: I do have the Izotope visual mixer, so maybe I'll give that a run through next time around! (and more reverb!)
Your most useful video to date. Well done! So good to see we can move past trailers into actual stories. And thank you for the breakdown of your process. Much appreciated. Will show to my students! Please remind your audience that it's still about story, not tech. Thanks!
10000%! This really does serve as my "call to arms" to start focusing on actual narrative! Trailers are still fun and all, but really, we should start crafting new and interesting ideas on narrative at this point. It can be done, and we have the ability and resources to get in on the ground floor and set the standard for this new medium!
@@TheoreticallyMedia Quick follow-up from England: What editing software do you use? I could not tell from your video.
This was so great. You're in trouble. Tuesday, October 1st...
Fantastic work, brother! I am both impressed and thankful to see that the level of work I am doing right now is at least in the same ballpark with what you've done here. That's heartening! You did some great storytelling in this, and that's really half the magic. Good storytelling sells it! Love the genre, love the art. Nice work! Thanks for the entertainment!
Lately Ive been coming back to pirate seas through your short and Our Flag Means Death. I appreciate the masterclass and it goes to show a person still needs to devote time, energy, funds and creativity to make your freak flag fly. I was also reminded that in The Lair Of The White Worm, Ken Russell did the horrific dreamy parts in video resolution which was jarring and f-ed up 😂. Rock on, brother!
Oh, I gotta get caught up on Our Flag...funny enough, I've been marathoning through What We Do In The Shadows...so, maybe that's where the vampire come from? Haha
@@TheoreticallyMedia Ah! You will be happy to know there is so much crossover! I love all those NZers. So warm, yet so dark.
First - I thought that film was a treat.
Second - thank you for the breakdown. It is very appreciated
Great breadown of the whole film mate, well done this was great for anyone wanting to get into film making with ai.
Thanks so much!! At some point, I need to circle back and do a full on "beginners" tutorial. I do feel at this point there's SO much info out there, that someone popping their head in right now is likely to explode!
One of maybe one actually making things while bringing the latest AI news. Congrats on 100k! I'll keep watching. Thanks!
Appreciate that so much! That's something I like to do from time to time-- It's always tough when you're reporting on the newest thing, and realizing: "Wait, I don't have time to play with the thing!"
Going forward, I want to make sure that I make a "thing" at least once a month. Keeps the skills sharp, and (to some degree), I think it's important to show, not just tell!
That was totally awesome, Tim. Congratulations. Keep going, man!
Great video Tim! The way you explain it makes it sound so easy, but I'm sure quite frustrating, and a lot of man hours! However, a perfect example of how creative you can be using A.I. Steve
Haha, yeah-- I was joking around earlier that this is the "baking show" version. Y'know, when you're watching a cooking show and they're like: "Just add this, and this, and this, and voila! You have a perfect meal!"
That said, it's always fun. Even the messy parts.
Ha, Tim! Cool look behind the scenes. I kinda guessed that this was made before the new lip-sync programs (but they may cause problems with aspect ratio?). You know what would be cool... to make this exact movie again one year from now! To show how fast things change. The possibilities with "face-acting" and lip-sync with Life Portrait are gonna lift this to new heights already.
So, I did have a bit in there that I cut for time-- but basically, the idea would be to take the 1:1 output, and then you'd have to mask it over the 16:9 footage, and scale/keyframe it to match. That, as I was in the middle of shooting, is sort of a tutorial unto itself-- and considering we aren't far off from LivePortrait and the next generation of Lipsync tools, it didn't make a lot of sense to spend 5 minutes talking about it!
@@TheoreticallyMedia Interesting nonetheless! Thanks for explaining.
Fantastic work Tim! Great to see this format coming from you :) looking forward to seeing more of that!
And never give up on those reflective comments. That woman's restaurant rating gave me my daily morning loud laugh 😂
You, Tim, are informed and funny and articulate and talented and give great advice which is why your channel is my favorite. (I'm sure lots of people told you it's Dream Machine rather than Dream Factory)
Brilliant film and brilliant breakdown, as always! 1 star yelp review made me lol
Haha, I could just see this lady stepping into the Pirate Tavern and being all "This is NOT what the pictures on Google looked like. The food had BETTER be good!"
100% she left a one star review!
Congrats!!! as well TIm God Bless Sir and Thank You for all Your editing and work on this channel and humor 🎉
oh, thank you so much!! We'll do a big party video when I get the Silver Play Button!
you are my inspiration in this ai stuff. thank youuuu
Thank you so much for that!!
I’ve dropped my mid journey subscription down to the $10 option and it’s actually still more then enough. So you could save $20 there. Just a suggestion.
Totally true. I started at the $60/mo level, but ended up dropping that down. That for sure was overkill for me. I might be able to get away with the $10 plan, but I do go through spurts where I generate a LOT.
Most of that is due to the channel though...I'll say, I'm good w/ $30 since MJ is a daily driver for me.
Nice work and breakdown I would say mainly its your work in the editor that made this film work IMO ,Some thing many people overlook
Don’t disagree. At some point I need to do a beginners AI Editing video. I wouldn’t want to do a full fledged editing tutorial, as there are plenty of those online, but I think one focused on new AI filmmakers might be worth it.
@@TheoreticallyMedia Sounds like a good idea , keep up the good work 👍
The VHS part was actually my favourite sequence. Very original for a pirate movie and aesthetically pleasing on top.
Oh, that's EXCELLENT to hear! I tend to (obviously) hear most from people that didn't like the sequence, but I knew there would be a few people that liked it! Thanks so much for that!
Might be my favorite ai channel
ahhhh, man-- YOU are the BEST!
Great brekdown! Your work is so important! I've been directing for 30 years and these tools are the best things to come along since NLEs. Could you possibly do a comparison video wit Dream Machine and Gen3?
The sound part with effects caught me. 🎉
The laughing pirates? Not gonna lie, kinda want to make it my ringtone!
Hey Tim, great job, I have been following since December and you have been a great point of reference for my own youtube AI/horror channel. I loved Dead Sea and I would love to be able to create something similar for my own channel, so this video is really helpful to realise how much in subscriptions I would have to spend, and I am pretty far from being sustainable with my small channel but I still appreciated the hard work you put on a daily basis!
well done!
Having also played around with all these models for my channel over the last few months, I totally feel you. It's like "playing a game". Each tool, at the moment, seems to be good at 1 or 2 specific thing. I also got some pretty good results throwing Haiper, Pixverse and Luminosity in the mix ^^ So grateful for your content!!
I did have a bit in the video that got cut around the part where I talked about not writing a full script. Basically, I'm not sure writing a script (at least in industry standard) is in anyone's best interest right now, as it'll lock your thinking into a particular shot or imagery-- which, (currently) the generators might (or might not) be able to pull off.
It's currently a game that feels a bit like free sculpting. You have something in mind and get to work, and then discover the piece in the process. Not to say you shouldn't write anything out, but at the same time, I don't think you need to fire up Final Draft!
@@TheoreticallyMedia I totally agree about how a script can lock our thinking into a particular image or shot, and it gets frustrating when the tools cannot replicate what the script suggests. Your notion of "free sculpting" is a better approach. I made the mistake of writing a script first and got frustrated. Better to have a good story idea and see what images emerge via the tools, and then sculpt the story from there.
@@TheoreticallyMedia I found the same for my project (my first ever!). I wrote a script but modified it along the way based on the final sequence generations I was able to get. Since it was my first time making anything, I had no expectations going into it, so it worked out. 😆 It was a lot of fun learning the tools and it's encouraging to see I followed a similar workflow to you just by trial and error, lol. Thanks for the video!
@@TheoreticallyMedia That's actually sort on. Sometimes I find myself fighting 30-40 minutes with 4 secs of my video because I'm stuck to my initial script. Next time I find myself and that situation I'll try an adjust the text rather than the video. Appreciate the feedback buddy!
Excellent work as always and a great inspiration to those on their AI creative generation journey
Appreciate that! Always the goal with this channel! Inspire, Educate, and Entertain...well, your milage may vary on the Entertain part. Depends on your threshold for dad-jokes!
@@TheoreticallyMedia I love a good dad joke 👍
Nice mate, yet again
Thank you so much!! Really means a lot! Haha. Lots of coffee today!
Actually, that trailer you made was brilliant. Just goes to show that full movies are just right around the corner.
Mind boggling that you can create something in a couple of days and pocket change, that would take weeks and 10s or even hundreds of thousands of dollars. Super impressive and an exciting harbinger of where we are heading .
Well done! I enjoyed it and I'm inspired to try my own trailer for my book ads.
Aren't we going to speak about the Bram Stoker in the room?
Without this little Irish lad we would have never see the birth of "Dead Sea" and personally "The Captain's Log" may be one of my favourite part in Dracula.
A strong story is spine on which a movie, series, animation,.. Etc is based but writers are very often overlooked and cheap, low quality writers are now often use.
We all know the state Hollywood is in right now and many of us remember what Hollywood use to be.
I truly love your spin on the "The Captain's Log" and to see the time and money you spent on the project blow my mind (obviously your expertise shorten the time🙇♂️).
The future is bright for AI artists 👍
Congrats on the exposure & hitting 100K! And thank you for generously sharing your process and price breakdown.
Do you think you could do this without Runway and use a free upscaler like Viva?
I can't download Kling in the US - it might be my old Android device, lol. Looking forward to your tips!
It's getting there. That was incredible.
Nice video Tim ! I have been using Dream Machine to generate scenes but it's kind of a hit or miss app. Sometimes I can generate 10 or more videos to get some video that will work for a scene. Other times I nail it on the first attempt. And this is for image to video generation. This drives up the cost of production for everyday use. Maybe they could include some sort of preview so you can see what you get before actual generation of the video. Thoughts?
Good Job Tim!I think it looks great! Good job with the idea and Capt. Renfield joke! 😉
Thanks Tim!! This is great!
Appreciate that so much!!
So, I've made a few AI trailers. You seem to have great success in marketing and showcasing your gen AI content. Do you have any tips for someone trying to get their AI content out to festivals and in front of large audiences?
Good work per usual. Why do you prefer Kling? As far as consistency, do you like LTX Studio? I've gotten mixed results on LTX
To be honest, I love them all. LTX has that great creative workflow, where you can get a birds eye view of your whole film. Plus, it serves as a great brainstorming engine. I rarely take what LTX first offers (that's a recurring theme on the channel), but I love that it always gives me a great springboard to try things out.
Currently, I'd say Kling is the best Image to Video on the market. Although, it has a lot of access issues, which is a real hamstring. That said, I haven't seen anything yet that has its level of coherency when you provide it with an input image. That said-- I mean, it's Tuesday. That could change by Thursday!
(Luma's dream factory is a close second on that front!)
@@TheoreticallyMedia yeah, I haven't been able to access Kling myself (hence why I was curious if it was worth the effort). LTX is interesting as far as an AI with some built-in consistency so it's cohesive. But it often doesn't pan out as intended. I prefer to simply create the media and edit it like you do, since there's more control on the backend. Thanks for everything you do on your channel - it's a great resource for understanding the latest and greatest, which, as you mentioned, can change on a daily basis!
The costs are great but what about your labor fee 😢 your time is worth something! Thanks always for another great educational upload :)
*edit: just for clarity because people tend to not be able to read between the lines or want unnecessary attention, yeah you did mention it but it’s always worth to mention more about self worth on anything people do :)
oh, 100% I did have a longer bit where I did more of a "my cost" breakdown-- but then I thought better of it, since that might come back to bite me in the butt! "In that video you said you cost X amount!!"
I may actually do a breakdown of what folks should charge at some point though!
That would be a great video ! Please do @@TheoreticallyMedia
@@TheoreticallyMedia The answer would be to break down the tasks and say that you spent 5 hours working as a video editor and 3 hours working as a script writer, and then use glassdoor or something to get an average pay for those positions.
Very Encouraging. Congratulations on your film.
And really appreciate the workflow and costings breakdowns.
Amazing visuals and work which ai is best for image to video..."I'm not able to access kiling AI. If there's any other AI, please let me know."
Currently, if not Kling, I'd say Luma's Dream Factory isn't far behind. You do need to experiment w/ that Enhance Prompt feature though. Sometimes it's great-- sometimes is goes SUPER insane. (see the "walk the plank" segment as an example!)
@@TheoreticallyMedia got it thank you so much..
If people are starting to learn movie making with AI then you still need to sketch up the concept and the lore.
Of course AI can help you with that, but it takes a human mind to flesh out the sketch into a real concept to where the AI can learn the concept so it can create better results from that concept.
So scripting and storyboarding may still be necessary..
And also, these LLM models may have to learn to generate actual 3d interactive environments..
Could you have used hedra to lip sync with greenscreen and then inserted in your scene?
I actually thought about that, but currently Hedra tends to get pretty wonky when you don't have a "head on" character-- although, I have been told that won't be an issue in the next version.
I think I still could have attempted it, but that would have required masking out the mouth movements, and then keyframing the head movement. It could be done, I think-- but it also would have added hours of time. I think if I had another 8 hours or so-- maybe?
But at some point-- you just gotta ship!
Next one, I think!
Thank you!!
Great Work, your work inspired me.
Thanks 🙏
This is Good stuff 👏👏👍👍. Congratulations. It's earned
5:30
😆 🤣 😂
Great blooper!
Yep, lip sync is also my absolute biggest problem. Next come jumps and falls. Good to see you struggled with those things too 😉
What fascinates me about AI creations so much is that it "thinks" "outside the box".
Ok it doesn't really do either, but if anything manages to "think" outside of human constraints, it's AI.
I’ve always considered it to be a really weird collaborator. And one that might have gotten drunk before showing up to work.
But sometimes that drunk has some golden ideas!
Question about new website for Kling when You get a chance could You help us to understand and take us through each how to get up and running I mostly use Web on the desktop not much mobile too small of screen for my tried eyes
Haha, I know the feeling.
Check out the video I did here: ruclips.net/video/CfTnMXodtns-/видео.html- you'll need to set up a Kuaishou account (the mobile app) regardless. Then when you log in to the website, it will give you a QR code that will open the app. From there-- It MIGHT work? I'm not sure if there is a waiting list for the website or not.
@@TheoreticallyMedia I did watch that but I cannot get this to work on my moble so I guess I have wait until they get up and running on the website
You forget your computer, your internet access subscription and time spend learning the stuff you did.
I mean, yeah- I didn’t want to get pandemic about it, but that’s true. There are some basics that I’ll take for granted.
As far as the learning part, I kind of cover that by talking about what would be my daily rate. I didn’t want to hone in on that, mainly because people at all levels watch these videos. Never want to discourage people from jumping in and starting to forge their own path!
THANK YOU SIR 🙏🏼. I'VE BEEN WAITING FOR THIS VIDEO SINCE I SAW DEAD SEA SHORT. I would love to see if you could pull this off with using either Free AI tools or at least the Free plan . Additional bonus if you can do it on an android phone. If anyone can pull it off I believe you can.
still not able to access Kling!! 😢
Your channel RUclips is soooo beautiful man we love your easy way and ai news me and my friends hi 👋 from iraq ❤
hey thank you (and your friends!) so much!!
Very cool. Very inspiring.
Just extrapolating from your experience, with increased productivity from doing this all the time and likely improvements in the technology, it's easy to imagine how a single artist might produce a half hour episode in a month the same way they might a single comic book
As a self professed “comic book guy” I couldn’t agree more! There’s an obvious ton of pushback on AI in the comic book world, namely amongst the fans- which is kind of a shame, considering the grind that most artists in the industry face.
But you’re absolutely right, if I could pull this off in 4 days, I’m sure we could extrapolate that out to a 10 minute film easily. Combine that with a small team, and you have a 22 minute monthly film/comic no problem.
You could even take a (ahem) page out of the old Manga phonebook style. Multiple ongoing teams/stories released as a two hour block every month.
It’s wild…
❤❤❤
can you host an ai movie compeition ? would be dope
This is awesome 🎉
Do you have a discord server by chance?
Great film......... big fan of pirate ship adventures.
Voices, Glitch effects, Actions sounds doesn't fit to the video. But overall excellent review on the production side. Keep on your great work!
Haha, yeah-- I knew it was a wild swing. I did get a few comments from people that DID like it-- but, ha-- for the most part, viewers didn't love that section. Oh well, I still stand by it! Glad overall the production side was helpful for you!
While the looks of it is great, there is no comparing with what is being done in real movies. There's no control ho the scene itself will be acted, where everything will be placed and the production itself is of missing to make it feel consistent. Again, it looks great and 10 years ago there was no such thing at that level, and still, it's a computer doing its best to calculate what will be working in the universe we are in. It is not there yet.
Really interesting
Some of the services suggested here are under several lawsuits for stealing content. Maybe not the best thing to suggest others to use them?
Should have asked for 'ships telescope'
That does look great, but no we are not there yet. Not even close.
I hope Luma adds lip-syncing.
I'm making an AI movie, already 24 minutes in. It will get better with time, because you can just keep getting better takes. Your movie is close, it will get better too.
Which website do you use
firstly Congrats for 100k subs, i also make ai movie trailer, like you, i had made the ai trailer but with 0 dollar, without sending money, but it require little bit time and editing, due to free tools, if you dont mind, i can send you the trailer which is made by free tools in 0 dollar
For sure! Swing it over-- would love to check it out!
I used up 30 usd of credits in Luma Dream Machine in less than 2 days. Making memes :)
It is a super great meme maker, I’ll be the first to admit!
For close up piece to camera talking you could have used Hedra or another TTV app, there's a really good one now on hugging face. Also don't mention 'coffin' instead describe it as 'wooden casket used to bury deceased humans safely' or something.
Do you have the name of the TTV app on HuggingFace you mentioned?
@@phizc its called Liveportrait, its opensource. Sorry I didn't mention that in the comment... what am I like? There are a few other vid on YT covering it I think. And its on a few spaces on hugging face to give it a go! It might run on colab? Maybe? Also its existing video, to video output via an uploaded image... eg. Image of your dog, video of Margerat thatcher saying something... output your dog lip synced to that... or a person... or two or three people...
@@mickelodiansurname9578 ah, yeah, I've seen videos about it. Looks great.
I up voted you bro
So that only took you 14 hrs?!
I’d say roughly. I got the idea on a Thursday and started generating some images up. Over the next few days, I was creating the Midjourney images, and funny enough, single Kling has a mobile app, I’d generate the video while I was running errands and waiting in line at various places.
I only did one full 8 hour day on it for the edit on the 4th day.
It was interesting, while in “production” mode, I’d have a good amount of time between generating video, and that really let my mind wander to decide what the next shot was going to be. All in all, a pretty stress free production!
@@TheoreticallyMedia wow congrats 🎉. I imagine if you created a small production pipeline of people who did it all day every day the man hours would go way down from even that!
I'm far too much of a perfectionist to use ai video in its current gen
oh yeah-- you gotta release a little of that right now. At least in its current state. I tend to think of it as working with a crew of toddlers, a drunk cinematographer, and super uncooperative actors.
It's basically creation through chaos. Haha-- Still, it can be a lot of fun!
@@TheoreticallyMedia I don't think text to video will be tge future of cinema you cant fit enough detail into wording or should I say most people can't turn what there thinking into a single prompt the new ai video company Odyssey has it right
@@JaxonBeckett I'm actually think a combo of Image to Video and Video to Video is the workflow of the future.
I've got something on that next week that I think will open some eyes! Stay tuned!
@@TheoreticallyMedia yes I agree I'd like to see a proper video to video model