Tadaam! we have finally completed this.
Added 2024-07-02 14:03:34 +0000 UTCThat's what we got!
-
to be very brief:
Yes, this is a very short video. Its value is not in duration. In fact, this is just a test of a whole bunch of technologies that we have learned to work with. If we already knew everything that we have been learning all this time, creating such a video would require about 2-3 days of work with animation and about 2-3 days for rendering.
That's if we had a location and ready-made characters. If you count the characters and the location, it's a couple of weeks of time and a couple of days per location.
Now that we have come all this way, we will be able to move much faster.
-
details. This text probably won't fit the whole thing here, so I'll continue it in the comments.
Comments
So, for the story. - In this project, initially we hoped to make a certain number of static pages per month and increase the speed as soon as funding allows. It was a dream to move through this story relatively quickly, to reach a speed of 50, maybe 100 pages per month. But pretty quickly we came up against the fact that, having reached a certain point, funding was no longer growing. And we couldn't expand the team. Even for the existing team, this was too little funding - a little less than a thousand dollars, which had to be divided into 3-4 team members and given a share to the translator. - Nevertheless, in the hope that the project would attract more subscribers in the near future, we continued to work fulltime and moved as fast as we could. We thought the quality was too bad on the first pages and started improving the quality to make this project more attractive. It periodically took a lot more effort and time than a “typical working day.” - We spent weekends on this job, working much more than 8 hours a day. But even so, we were sometimes behind schedule. The improvement in quality also led to the fact that it began to take much longer to create one page than we expected at the beginning of the project. As a result, over the several years of the project's life, we have accumulated a significant backlog in our work schedule. We tried to optimize the whole process, constantly learning and improving models, environment, technology. The appearance of 3090 video cards on the market has greatly facilitated the situation, but the cost of buying them... well, you know the prices, right? - Our team is part of a large artel and this gives us some advantage - we received “co-financing” from other, more successful projects so that we can develop our project, make it profitable and return investments to the “common fund". But time flew by, the backlog increased, despite the increase in quality - the project did not attract enough subscribers to reach at least self-sufficiency. Not to mention any profit. Several times we thought about how we could solve this problem. - Options were voiced - just spend as much time on it as the financing can recoup. And completely close projects in this genre (vore). In the first case, this would mean reducing publications to 7-10 images per month, which would be dramatically small and most likely a significant part of subscribers would be too disappointed with the slowdown and would leave this project. As a result, we most likely came to the second option anyway - to close the project. But you, our subscribers, have invested a lot of money in this project and we could not let down your expectations. Our team also invested a lot of their own efforts and money in this project (let me remind you that this project was unprofitable 90% of the time - in any other project of our artel, we would have made a profit, not losses). We didn't want to let you down and we didn't want our investments for 3-4 years to be in vain. - And all this time, we have seen how some animation projects - even if they had much simpler models, much simpler technology and plots - received 10-20 times more funding. - We came to the conclusion that the time of static images is probably a thing of the past and people appreciate animation more. We like animation better ourselves, because it can convey much more subtle shades than even a very large number of static pages. And we decided to take a leap of faith and immerse ourselves in animation technology. At that time, our knowledge in this area was ... relatively “average depth”, but there was little practice. We looked at the Unreal Engine advertisement and decided that this should be the perfect solution. At the beginning of the journey, when all our models and environments could not fit into the video memory of powerful video cards, we had to render on the processor. And it was excruciatingly long. It takes several hours to get a render of 1 frame. There were often errors and problems in it, so we had to do 2-3 renderings to get the final page. UE5 seemed like the perfect solution, because it promised realtime renderers. We posted an announcement here that we want to take a break and explore this engine to make a video for you. - A man came to our team who knew the topic deeply and he promised to teach us everything. We paid him a substantial sum and... he let us down. Once, twice, three times. He obviously wanted more money than he was willing to do any good. We broke up with this man, our team became depressed... but we couldn't let our subscribers down and we resolutely rushed into this dark cave without a guide. We spent 4 months intensively studying all the tutorials that can only be found on the web. We studied all the free tutorials and bought a dozen paid ones (yes, another impressive investment). All these months, our subscribers have continued to support the project and this has greatly supported our inspiration and determination. If it weren't for your support, we would have stopped wasting time and effort on this back then. - To study the amount of information that many people spend several years on - in 4 months… This is a very extreme strain on the psyche. We had practically no life outside of work. We woke up, worked, fell asleep. Without days off and rest, distracted only by solving everyday problems that simply can no longer be postponed. We did it and did it - you could see those videos on UE5 here. We created a character, learned how to customize the mech (this is a separate hell within hell for UE). We learned how to work with the materials system and transfer locations to the UE. We wrote our own plugin to solve problems that were not solved on the Internet at that time. We even paid for the time of one of the UE developers to tell us how to fix these problems at the stage of interaction between the Blender and the UE. I think you understand that these guys get more in a day than we get in a month, right? - We went 95% of the way and... found that either the fur looks disgusting, or the video rendering takes many, many hours. There was no question of any real time. And the worst thing is that it turned out that if a problem or error was found on the final render, fixing it required a very, very absurdly long tinkering with updating the character animation, transferring it inside the UE, and this process alone took 10-15 minutes of time every time we needed to correct a small movement. In our project, the characters interact in detail with the environment. The environment that we had at that time was too poorly optimized for working with UE. It would take another six months to optimize it, if it were even possible. UE had its own gorgeous environment - of magical quality. We used it, but faced a new problem - how to make the animation of characters from the Blender interact with the environment inside the UE. We had to transfer the environment from the UE to the Blender, do the animation, transfer it all back to the UE… In short, it was an absurdly uncomfortable, painful, monstrously annoying process. But the worst thing is that in the end we got almost as slow a render as we had before. - By this time, dozens of plugins had been released for Blender. A 4090 graphics card appeared on the market (in our country it was sold at a price of 2.4 months of financing for this project) and more powerful processors. And thanks to tinkering with UE, we gained knowledge that allowed us to optimize everything we did strongly enough to fit into the memory of the video card. We've redesigned Rose's character 10 times, the location, the materials, the textures - everything. It was a separate hell. How to describe it more clearly... Imagine that you built a 2-storey country house, but you did it in a bad place and you need to disassemble this house and build it again - in another place. And then again and again. And although in some cases it is possible to disassemble not completely to the foundation, it is like an eternal repair or general cleaning lasting a decade. The situation was extremely nervous, every now and then someone from the team tried to give up everything and just go into other projects. Someone left. Others came in their place. The core of the team continued to crawl towards the finish line, although everyone had thoughts of giving up this idea forever, coming to terms with the fact that it was all too difficult for us. But you, our subscribers, have continued to support us all these months. The very fact that so many people not only provide moral support, but are also ready to spend their resources (money) on this project, even with such an absurdly prolonged pause, allowed us to maintain this pace and persistently dig into the depths. - around the fall of 2023, we could launch the first video, but it would be the simplest video “like everyone else” we've seen in the furry fandom. But it seemed to us that after all these efforts and after almost a year of waiting for our subscribers, we all deserve not to stop at a simple level and step up. - During all this time, we have learned a huge range of technologies - muscles, fat physics, deeper settings for working with fur, character shape, shape keys and many other technologies, the meaning of which will be clear only to specialists in this field. We learned how to optimize all this so that it fits into the memory of the video card and reached a speed of 10-15 minutes of rendering per 1 image. Against the background of 2-3 hours earlier, it seemed like a wild success, but it was still too long - there are at least 24 frames in a second of video, if you multiply all the numbers.. it turned out that we could make no more than 1 minute of video per month. And we studied further and invented our own solutions to some problems. Now it takes 2-3 minutes to render 1 frame, and this is an incredible record, it seems to me. With the level of detail of everything in our frame, I don't think anyone has achieved such indicators even close. - The second epic problem was an extremely unpleasant discovery. Do you know what the problem with open source technologies is? They usually don't have bosses who decisively dismiss fools with their hands out of their asses. There are many good things in the Blender, but there are also quite a few absolutely terrible things. It would be possible to talk about them for several days, I will spare you these details. To put it briefly, using an analogy, imagine that you are running on a summer lawn. Everything is very beautiful around, nature, forest... but almost every burdock has a huge pile of shit piled up. While looking at it from a safe distance, it's a blooming beautiful meadow, but it's worth starting to run through it.... - It turned out that all these many very cool plugins that we bought (even more than $2k in expenses, even taking into account discounts on sales) - all these plugins can almost never work correctly with other plugins. The physics of muscles, the animation of fur, the physics of clothing, the physics of the environment - all this is absolutely not able to interact with each other correctly. And our knowledge of Marvelous, which we considered very high at the time we were making static renderings of clothes, is complete nonsense at the kindergarten level. Animation required completely different knowledge. And we continued to study. Every time it seemed to us that we had reached the right point. And all that remains is to make the final render. And when we are sure that everything is working, we will move on to the routine creation of a content stream and will finally be able to return to a normal work schedule. Afford to play games released over the past year, watch all the cartoons like Kungfu panda 4 and the like. For the last six months, we have lived with the feeling that we are standing a meter from the finish line. But every time some problem came up that ruined everything. We solved this problem, moved a centimeter further - we were completely sure that now there could be no problems, but - no, there is a problem again! - An emotional swing at supersonic speed, that's about how it felt. There was a moment when our team actually disintegrated into molecules crawling in different directions. But a day or two passed and in the gloomy depths of depression we continued to pick at this problem again. One by one. Day after day, month after month. We have learned a layer of knowledge that many probably would not have gathered in 4-5 years. And finally we did it - a video in which you can see: - the physics of hair (it's not perfect, we just didn't delay everything for another month and endlessly edit it. We'll do it later, of course) - muscle physics (Rose's muscles can tense, sway and contract). We will show this in a separate video a little later. - fat physics (Rose's breasts and her belly) - character breathing - usually few people worry about such little things, but we decided to make the characters breathe so that they look as lively and naturalistic as possible clothes. Most even very cool studios, if you look closely, carefully avoid simulating clothes. Their clothes are nailed to the character, they are part of the body. These are usually very simplified clothes with minimal imitation of physics. In most cases. 99.9%. The edge of the garment always stays in the same place. We made the clothes a full-fledged simulation, so that in certain situations the edge of the clothes or the entire T-shirt could move over the character's body or could be completely removed altogether. - And this has confronted us with a problem that the largest studios, with their huge computing power, are afraid of and avoid. Clothes and fur. This is an additional hell within hell. In our model, Rose's hair has limited physics and it requires 1-2 hours of simulation before proceeding to render. And every change in the animation of her head will require going through those 2 hours of simulation again. And there are only a few hundred thousand hairs. There are about 2 million of them on the character's body. - simulating all the hairs required a supercomputer or a month of work on our home PCs. And without the simulation of each hair, they would, like cactus needles, pass through clothes and it would look terrible. Many furry 3d videos often abandon fur altogether and the characters look as if they are made of plastic. We didn't want to sacrifice the quality so much. Therefore, we had to invent a separate package of tricks and tricks to solve this problem. But we solved it - if you look closely, you will notice that the edges of Rose's clothes move within quite noticeable limits. Despite the fur. - Rose's eyes. Her pupils can dilate and contract. fur on the elbows. We still made “pigtails” on the elbows, because without them, the furry character seems a little strange. And it took a week of tinkering with learning additional technology. The fact is that if we just made a slightly longer fur on her elbows, when bending her arm, this fur would be absolutely motionless relative to the surface of the skin. And bending the elbow would completely lose the shape that you see in our video. It would look like a terrible and disgusting “explosion” of hair in different directions. The invention that we have applied allows us to control the direction of fur growth depending on the position of the bones, that is, when the arm is bent, this fur changes shape. - And we had dozens more similar tasks and problems, I just don't want to bore you with listing all these details, and I'm too tired right now to make this text even longer and more tedious. - In general, let's summarize the results: We've completed our huge journey into animation and learned everything we need to continue making tons of content for you. At first it will still be slow (although not nearly as slow as those terrible 17 months of pause, ahahah...). - We need to improve some of these technologies - it won't take long and it won't contain any surprises anymore. Just a bit of routine fuss and long tests. - We need to remake the rest of the characters for the 15th time, taking into account all new technologies. It won't take long either, well, relatively. It will probably take a month or two to update all the characters, but of course we won't keep you waiting for a few more months. We will start producing content from a place that is slightly earlier in the story than the last picture. From this point on, we only have a couple of characters - Jay and Rose, who are walking through the forest and Rose is carrying her heavy belly with another hare swallowed. Here we will need to invent a couple of additional technologies that allow her palms to affect the physics of her belly, and the hare inside to affect the physics of her body and palm. But we already know the solution, it's half done - you just need a little practice with it. We will start this content for now, but we probably run the risk of missing it this month. Although we will try to publish reports much more often and more regularly, now we will have some time to rest and we will regain our strength. - During this time, another part of the team will update the location for this portion of the content. We will move “on the run” and this means that in the future there may still be some pauses and delays in publications, but they will not be as huge as before, their number will decrease and in the next six months we will enter a regular update schedule. During this time, we will understand which stage and how much time it takes, improve our practical experience and be able to move faster. - The only problem is that we don't know what to link the schedule to. How to measure our work and how much content to promise to all subscribers. The fact is that the animation of a different number of characters in the frame, even the same character, is very different. If the character just stands and looks, turns his head, blinks - this can be done very quickly and simply. If the character moves intensively, it is two to three times more difficult, especially if the character is in the frame entirely, and not just some part of it. If a character interacts intensively with another character, it is ten times more difficult. It is simply impossible to measure all this in any units, so we can only measure it in “the hours of work spent on creating it.” There are no other options, but I think this will not be a problem - we see your high trust and will do our best not to disappoint you. As you may have noticed, we would rather sacrifice our personal time than the quality of our product. Although, to be honest - now, after this painful marathon - we plan to allow ourselves more rest and after the next 2-3 months return to work 8 hours a day and with two days off a week. As far as it goes. - - Oh, yes, a few more important nuances. - - We plan to use this video as part of a promotional package. We will publish it publicly. In a few months, probably when we will have a more or less routine content creation process. But for now, I ask you not to distribute it outside the club, so as not to create unnecessary excitement. New people, if they are attracted by this video, will come to us, but they will still only receive pages from 1 to 600+ before they see the video. Therefore, it will be better to attract new people and resume advertising only when there will be a little more than one such video. - And also - of course, we want to eventually make this project of uniform quality and the first static images will gradually be replaced by video. Probably at the end of the project. This will largely depend on the growth of funding and the possibility of expanding our team. - And of course, let me remind you once again that every month that you have supported us will be filled with content. In proportion to the financing. All those who showed this monstrous patience and went through this difficult path with us will receive a lifetime low price bonus, i.e. video prices will not rise for you. Those who were a little impatient will receive a video at the old price in each month in which they showed patience, but after the end of this period, prices will be increased, as well as for all other club members who have not yet reached this pause, as well as for new ones who have not yet come, but will come in the future. - Uh... well, now we are looking forward to your comments and ratings with horror and impatience. I hope our efforts were worth it and were not a big disappointment for anyone. But don't be shy about any criticism - it will allow us to improve our product. We welcome constructive criticism
vorecomicsclub
2024-07-02 14:04:59 +0000 UTChttps://dropmefiles.com/zeYOE loseless version, 800 Mb
vorecomicsclub
2024-07-02 14:04:31 +0000 UTC
