Wednesday 20 April 2016
Week 31
Hand in week. So I attended a drop in session with Chris on Monday to just double check somethings with the metal map and roughness map (if they're black and white). I was able to make the metal map rather quick due to the fact that its very binary in the sense it's either on or off and the only thing that wasn't metal on my piece was the carbon fibre that surrounded the mount for the limb to go in. The roughness map took a bit longer with Chris showing a good technique to produce over spray and to lift the paint up. For example the infinity logo on the top, I used a dark grey underneath that part on the roughness map that really brought that logo up and makes it look as if its sitting on top of the rest of the paint. I was also able to make the finger tips look warn down with the roughness map and on the wrist I was able to make it look as if oil had been dropped on the surface of it. I can really see the potential of a program like substance painter that allows you to paint directly on to the surface of an object because it ends up giving a more cohesive look to the piece. Where as doing it the old fashion way like I did it can be hard to match up the exact amount of ware on each piece. I've also ended up putting my pillar into Marmoset to get some cleaner more professional shots as you can see below. I'm really happy with the way they look and how they've come together.
Week 30
After completing my low poly last week it was for me to begin baking the normal maps for the robotic hand. I attended the drop in session on the Monday to do this due to the fact that I hadn't baked multiple objects before, as with my pillar I merged the three objects and baked them as one. Chris said it basically exactly the same except each item has its own cage and when you produce the normal's you need to make sure they're set to save as PNG's so you can then open them all up in Photoshop and bring them all on to one normal map.
I began with one of the finger tips to my hand i put the required parts into the XNormal program, I then opened up marmoset as we'd been shown that morning how to use it to create turntables for hand in and it works as a quick way to view work without having to open up UE4. As soon as I put the normal maps on to the low poly instantly it looked bizarre the best way to explain is that it looked as if the normal was too big for the low poly. The other issue was that because of how many items i was trying to fit on a 2k texture was causing some pixelation around the object and losing a lot of the detail I was trying to retain from the high poly. I asked Chris and we came to the same conclusion that to reach the required poly count I had removed too much of the geometry and there wasn't a lot I could do about this issue. Chris said to just complete the baking process at home and see what it looks like. I was rather annoyed as I'd spent a lot of time and effort designing the high poly and it just wasn't coming out remotely nice. This was when I came up with a solution, hopefully, to the issue it was to take the low poly that has the geometry that has been changed to far and start adding in edge loops to create a "low high poly". This was an attempt to try and bring back this project and have my final piece look nicer than what was being presented to me at that time.
I then spent the next couple of days making such a piece, of course I tested one piece, a finger tip and compared it to the old work which Chris and I agreed looked much better. Although I'm disappointed I couldn't use the high poly I am happy that I was able to come up with a creative solution to the problem. I encountered some issues that I had to overcome with the typology due to the fact that I'd been trying to reduce the amount of polygons some edges were causing my edge loops to band together causing pulls in the mesh which I had to fix manually. In the end however I'm happy with overall outcome of the piece as you can see below, I also was able to get the ambient occlusion done this week and worked over the weekend to get an albedo done. My inspiration for the albedo was a gritty Borderlands and overall I don't think it looks too bad, unfortunately however this has put me behind a little as I still have the metal map and roughness map to complete. Thankfully however the week I've left for getting things ready for hand in has allowed me the time to do this work.
I began with one of the finger tips to my hand i put the required parts into the XNormal program, I then opened up marmoset as we'd been shown that morning how to use it to create turntables for hand in and it works as a quick way to view work without having to open up UE4. As soon as I put the normal maps on to the low poly instantly it looked bizarre the best way to explain is that it looked as if the normal was too big for the low poly. The other issue was that because of how many items i was trying to fit on a 2k texture was causing some pixelation around the object and losing a lot of the detail I was trying to retain from the high poly. I asked Chris and we came to the same conclusion that to reach the required poly count I had removed too much of the geometry and there wasn't a lot I could do about this issue. Chris said to just complete the baking process at home and see what it looks like. I was rather annoyed as I'd spent a lot of time and effort designing the high poly and it just wasn't coming out remotely nice. This was when I came up with a solution, hopefully, to the issue it was to take the low poly that has the geometry that has been changed to far and start adding in edge loops to create a "low high poly". This was an attempt to try and bring back this project and have my final piece look nicer than what was being presented to me at that time.
I then spent the next couple of days making such a piece, of course I tested one piece, a finger tip and compared it to the old work which Chris and I agreed looked much better. Although I'm disappointed I couldn't use the high poly I am happy that I was able to come up with a creative solution to the problem. I encountered some issues that I had to overcome with the typology due to the fact that I'd been trying to reduce the amount of polygons some edges were causing my edge loops to band together causing pulls in the mesh which I had to fix manually. In the end however I'm happy with overall outcome of the piece as you can see below, I also was able to get the ambient occlusion done this week and worked over the weekend to get an albedo done. My inspiration for the albedo was a gritty Borderlands and overall I don't think it looks too bad, unfortunately however this has put me behind a little as I still have the metal map and roughness map to complete. Thankfully however the week I've left for getting things ready for hand in has allowed me the time to do this work.
Week 29
After finally completing my high poly, I attended my workshop with Chris. Although I'd made a start on my low poly I hadn't gotten too far with it due to the fact that I wanted to get some tips for bringing my poly count down. As I knew I was going to be struggling to reach the number Chris was after due to the sheer fact that I had eighteen objects I wanted as many tips as I could get my hands on. One tip was to get rid of the insides of where the finger joints are due to the fact that you can't really see them and from a distance they won't have that much of an impact. Other things I could do was too remove the holes on the fingers due to the fact that they can baked on in the normal maps and they won't look out of place, this would also help for when I go to unwrap the piece due to the fact that there will be less fiddly parts to have to work with.
So I returned home and that was exactly what I did, I removed the holes drastically bringing the count down. This was a very monotonous task that had to be done but I was really rather happy to bring my overall scene down into the 9000 mark. This took a lot of reducing in geometry and meant that each item needed to be around 500 polys. Which took an awful lot to get too, I removed as much geometry as I could trying my best not get rid of any important parts that would affect the baking of the normal's at the end.
After this I moved on to unwrapping each object which was interesting as, each object was relatively similar to the last and I wanted to unwrap each the same as the one before. Which meant I would unwrap the first one as best I could do the second one and realize a better and more efficient way of unwrapping the object then return to the first one and do it again. Which as you can imagine meant unwrapping the object took longer than first expected but at the end of the day meant the baking of the normal's should be good. After finally reaching the end of the unwrapping part of this task I attended a drop in session to get Chris to give my UV's the once over. After seeing the work he showed places to improve my UV's due to when I run the script that hardens and soften edges their may be some issues as a lot of the UV's need to be on separate islands. This is due to the fact that I want my object to be relatively hard to resemble the metal I want to imitate. I then returned home and began breaking the UV's and trying to create a more readable UV map layout. Below is the finished low poly after the script has been ran. So far I appear to be keeping to the adjusted time table I'd given myself.
So I returned home and that was exactly what I did, I removed the holes drastically bringing the count down. This was a very monotonous task that had to be done but I was really rather happy to bring my overall scene down into the 9000 mark. This took a lot of reducing in geometry and meant that each item needed to be around 500 polys. Which took an awful lot to get too, I removed as much geometry as I could trying my best not get rid of any important parts that would affect the baking of the normal's at the end.
After this I moved on to unwrapping each object which was interesting as, each object was relatively similar to the last and I wanted to unwrap each the same as the one before. Which meant I would unwrap the first one as best I could do the second one and realize a better and more efficient way of unwrapping the object then return to the first one and do it again. Which as you can imagine meant unwrapping the object took longer than first expected but at the end of the day meant the baking of the normal's should be good. After finally reaching the end of the unwrapping part of this task I attended a drop in session to get Chris to give my UV's the once over. After seeing the work he showed places to improve my UV's due to when I run the script that hardens and soften edges their may be some issues as a lot of the UV's need to be on separate islands. This is due to the fact that I want my object to be relatively hard to resemble the metal I want to imitate. I then returned home and began breaking the UV's and trying to create a more readable UV map layout. Below is the finished low poly after the script has been ran. So far I appear to be keeping to the adjusted time table I'd given myself.
Week 28
I spent this week focusing on getting my high poly done. Most of this was fairly easy to create as I frequently used my own hand as reference and trying to reduce it down to some basic shapes that I could craft into metal pieces for my hand. The most trouble I had was with the palm and thumb due to the fact that the thumb needs to be able to move but a ball joint seemed out of the place and offered too much freedom for thumb as you can see below. I also found my polygon count was getting rapidly higher so I needed to stunt some of my creativity if I really wished to get it within the poly count Chris was asking for.
As you can see below I went through a few iterations of the palm due to the fact that it offered the most interesting design options but came with the sticking point which is where to place the thumb. After doing some research into where on a human the thumb joint sat and looking at other peoples interpretations I landed on the joint mechanism on the left that relies on the thumb rather than the palm, I also think aesthetically it looks much better than the one on right. Now back to the issue that was the actual design of the palm as you can see I did go down some rather intricate design paths but found that when I wanted to put in the mechanism for the thumb I was struggling get clean geometry because of all the edge loops and the poly count was getting to a point in which I worried if I would even be able to get it low enough. So I went for a lot more simpler design that I was going to rely on the roughness map to put in some interesting scratches and designs. I did however add this sort of shield to the top of the palm to give it more of an interesting feature and I think it really helps with the overall atheistic of piece. The wrist was fairly simple to work as I've gone for a design that allows all the movement you would want, up and down delivered by the cylindrical pin at where the wrist and palm join. Then left and right being delivered by the wrist rotating on the first part of the wrist design the rest is static. Now that the high poly is done I can now move on to creating my low poly so far I seem to be keeping to the plan I had for this task my only issue may be that I wont be able to get within the limit for the polys as currently my high poly sits at 50,000 (1.5 million when a smoother is applied) which within ZBrush is fine but inside of Maya it means I need to take away each edge loop by hand until it get's within the needed specs for the task. Overall however I'm happy with they way in which this has come out.
As you can see below I went through a few iterations of the palm due to the fact that it offered the most interesting design options but came with the sticking point which is where to place the thumb. After doing some research into where on a human the thumb joint sat and looking at other peoples interpretations I landed on the joint mechanism on the left that relies on the thumb rather than the palm, I also think aesthetically it looks much better than the one on right. Now back to the issue that was the actual design of the palm as you can see I did go down some rather intricate design paths but found that when I wanted to put in the mechanism for the thumb I was struggling get clean geometry because of all the edge loops and the poly count was getting to a point in which I worried if I would even be able to get it low enough. So I went for a lot more simpler design that I was going to rely on the roughness map to put in some interesting scratches and designs. I did however add this sort of shield to the top of the palm to give it more of an interesting feature and I think it really helps with the overall atheistic of piece. The wrist was fairly simple to work as I've gone for a design that allows all the movement you would want, up and down delivered by the cylindrical pin at where the wrist and palm join. Then left and right being delivered by the wrist rotating on the first part of the wrist design the rest is static. Now that the high poly is done I can now move on to creating my low poly so far I seem to be keeping to the plan I had for this task my only issue may be that I wont be able to get within the limit for the polys as currently my high poly sits at 50,000 (1.5 million when a smoother is applied) which within ZBrush is fine but inside of Maya it means I need to take away each edge loop by hand until it get's within the needed specs for the task. Overall however I'm happy with they way in which this has come out.
Week 27
This week I decided to return to my other idea, the robotic hand due to the reason that I found the police drone was rather easy and since my last task I found rather straight forward I thought it was best to try and push myself a little harder. This has put this project back by a week due to this, this will be taken out of the week that I left for any issues. This still puts me on track to finish well within the time issued but any issues will make this harder due to the reduced amount of time but I believe that this is the best thing for me to do as it means I can really explore what my skills are as a 3D artist.
So just like the previous idea I began by using references and searching through the internet to see what sort of look I'd like to go it also helps to see how to create them and make a process for the work. One artist I've had an obsession with since Chris introduced him to us Vitaly Bulagrov he makes some incredible pieces of work which mostly focus around this mechanical future I was so desiring with my work. He was the most influential person in this piece of work. I also wanted my work to resemble something realistic which was what Bulagrov in my opinion does best. I have however considered that my allowed poly count may be a little low for what I wish to create but I think it should work in the long run, but I do need to be aware of the little amount of poly's I'm allowed to use for this task. So I have been cautious and knowing that cylinders suck up a lot of geometry if you let them I've been using reduced the amount of faces to eight to try and reduce the poly count early on. So far this week I've been able to make the finger tips and the middle section of the fingers. I'm really happy with how this is coming along they look mechanical but realistic at the same time.
So just like the previous idea I began by using references and searching through the internet to see what sort of look I'd like to go it also helps to see how to create them and make a process for the work. One artist I've had an obsession with since Chris introduced him to us Vitaly Bulagrov he makes some incredible pieces of work which mostly focus around this mechanical future I was so desiring with my work. He was the most influential person in this piece of work. I also wanted my work to resemble something realistic which was what Bulagrov in my opinion does best. I have however considered that my allowed poly count may be a little low for what I wish to create but I think it should work in the long run, but I do need to be aware of the little amount of poly's I'm allowed to use for this task. So I have been cautious and knowing that cylinders suck up a lot of geometry if you let them I've been using reduced the amount of faces to eight to try and reduce the poly count early on. So far this week I've been able to make the finger tips and the middle section of the fingers. I'm really happy with how this is coming along they look mechanical but realistic at the same time.
Week 26
This week was the start of our new task, which is the time machine, I wanted to make my time machine more of a mechanic in a video game instead of a narrative piece in a video game. So I had two ideas for my asset creation task. The first was a robotic hand that allows the user to see where an object has come from in time and space and enter a virtual space in which the object originated from to inspect for clues on its origin (a vault key maybe), set in a gritty borderlands art atheistic with lots of guns. The second idea was a police detective drone that scans a room and uses the information inside of the room to make up a virtual re-enactment of what happened in the crime scene, therefore creating a time machine.
For this task my plan on time was;
-First week, make sure I've nailed down what I wish too create for this task and begin on such.
-Second week, begin making and complete making the high poly.
-Third week, begin making and complete the low poly from the high poly model.
-Fourth week, create all maps essential for this task and present the piece correctly finished.
-Fifth week, in case there's any issues that required more time before hand-in.
-Sixth week, get things together for hand making sure everything has been met for hand in.
I decided in the end to go for the drone idea as it seemed a little more simple for creation. So I began creation of this, I got together a few images for reference to try and make something that could mirror a real life drone. I have decided though as both idea's are hard surface items I'm going to stay within Maya due to the fact that I'm thinking any surface detail I can put back in using a roughness map. Below are the beginnings of my drone was a fairly simple process with very few issues so far. Only thing I have noticed however is that I need to watch where I'm putting edge loops in future so I don't have any complicated issues further down the line. I also re-used the special duplicate from the telephone project to create all four wings perfectly adjacent to each other.
For this task my plan on time was;
-First week, make sure I've nailed down what I wish too create for this task and begin on such.
-Second week, begin making and complete making the high poly.
-Third week, begin making and complete the low poly from the high poly model.
-Fourth week, create all maps essential for this task and present the piece correctly finished.
-Fifth week, in case there's any issues that required more time before hand-in.
-Sixth week, get things together for hand making sure everything has been met for hand in.
I decided in the end to go for the drone idea as it seemed a little more simple for creation. So I began creation of this, I got together a few images for reference to try and make something that could mirror a real life drone. I have decided though as both idea's are hard surface items I'm going to stay within Maya due to the fact that I'm thinking any surface detail I can put back in using a roughness map. Below are the beginnings of my drone was a fairly simple process with very few issues so far. Only thing I have noticed however is that I need to watch where I'm putting edge loops in future so I don't have any complicated issues further down the line. I also re-used the special duplicate from the telephone project to create all four wings perfectly adjacent to each other.
Week 25
After getting a reasonably nice bake first time on my pillar last week, I asked Chris what I could do to improve some issues that were appearing on this bake of my normal maps. For example on one of the corners by the base of the statue their appeared to be a black spot appearing, which he explained was an issue with the cage not reaching that corner quiet correctly. I also had some weird lines appearing on the rear of the statue that were due to some poor topology which he said was something that would take some time but isn't impossible to fix. So I looked at what time I had left before we moved on to the next task, the time machine, I really wanted to have a go at unwrapping and baking the low-poly of the Mayan statue I posted on here last week. Yet I assessed what I believed I could get done within the time period we had left for this task and decided it was best to take something I knew was possible and make it the best possible iteration I could. So I spent the next day re-working the topology until I was happy that this was the best I could make it, also going back and seeing these issue's really helped me to get an understanding of how to miss these issues in the future.
After working all the issues out with the topology and re-aligning the cage I ran this script that essentially hardens all the edges that surround an UV island and soften the inside. After running my first bake for normal maps I soon saw that this was going to cause some issues in places and needed a little bit of adjustment for example where I had run cut down the main the section of the pillar to unwrap it correctly the harden edge was causing it too look like a crease running down this section of the pillar. It was a simple fix, all I had to do was change this particular part of the edge of the UV island to a soften edge to compensate for the issue. After I done this all over the normal maps were coming out really well I couldn't be happier with the outcome of them. I also then ran my ambient occlusion maps through and this just helped emphasis all the details that were somewhat being lost, this really helped my Roman pillar to look realistic. I generally don't think I could have made this particular pillar any better without starting again. Looking back at the amount of time I had left for this part of the project I believe that this was by far the best option. As I think the Mayan pillar would have caused some real issues due to the inspection of the geometry I did before making my decision. After last week's attempt at painting a landscape In colour I've also been giving my characters ago trying to become more diverse with the styles I can do, so below is the beginning of a painting of a Disney inspired character.
After working all the issues out with the topology and re-aligning the cage I ran this script that essentially hardens all the edges that surround an UV island and soften the inside. After running my first bake for normal maps I soon saw that this was going to cause some issues in places and needed a little bit of adjustment for example where I had run cut down the main the section of the pillar to unwrap it correctly the harden edge was causing it too look like a crease running down this section of the pillar. It was a simple fix, all I had to do was change this particular part of the edge of the UV island to a soften edge to compensate for the issue. After I done this all over the normal maps were coming out really well I couldn't be happier with the outcome of them. I also then ran my ambient occlusion maps through and this just helped emphasis all the details that were somewhat being lost, this really helped my Roman pillar to look realistic. I generally don't think I could have made this particular pillar any better without starting again. Looking back at the amount of time I had left for this part of the project I believe that this was by far the best option. As I think the Mayan pillar would have caused some real issues due to the inspection of the geometry I did before making my decision. After last week's attempt at painting a landscape In colour I've also been giving my characters ago trying to become more diverse with the styles I can do, so below is the beginning of a painting of a Disney inspired character.
Subscribe to:
Posts (Atom)