Iron Trying

The title makes sense I promise. At the end of this week I found out that I'm borderline Anemic which explains why I've been so incredibly run down this week. But despite that I'm miraculously still on task. Although I'm not as far ahead as I think I could have been, had I not been struck by five different ailments at once as a result of this.

The Project

But enough complaining, this week has definitely had it's highs and lows. So whilst I just said enough complaining, I am going to start out by talking about the lows first. It's better to get the bad stuff out of the way quickly to spend more time focusing on the positives in my opinion.

Engine.

First, I just want to start by saying that whilst I still love engine, I do not love using it on the lab computers because of technical problems. I nearly lost all my work for the day because of those problems and it was not a happy time for me. But the good news is I got that work back, and now I've learnt the valuable lesson of 'WORK AT THE SAME COMPUTER FOR ENGINE STUFF.'

I showed Mike my Whitebox on Tuesday and his feedback was to adjust the proportions a bit, because whilst the actual assets and room might be real world proportions and sizes, things always feel a little off in first person mode. So thats exactly what I did.

This is the original:

And this is the re-scaled one:

It feels much better proportion wise, I was told that often things need to be scaled up a bit larger than they'd really be for first person perspective, it definitely feels more in scale now, and it's starting to look like the room.

I was also told to make the still match up a priority, which I did. This involved a lot of screenshots. A LOT of screenshots, and a lot of camera duplicates (every time I wanted to make a major change I made a new camera, since you can't undo a camera move).

So once I had a relatively close match up (yes that is a screenshot I've added in in Week 4-I forgot to take it in Week 2, which is something I'll get onto later in this post) I left it for the time-being (I'll edit it again once I have the actual assets in Engine, as they'll change the positioning and scaling slightly) and moved onto concepting my models.

So since none of our references are particularly high quality (the pain of doing a room from a TV show I guess), I decided that for the couch and armchairs I'd spend some time just analysing them and breaking them down to make modelling them easier. It's something I'm learning to do more and more when modelling something I haven't come across before. I give myself an hour or so to break it down and work out how I'm going to model it, and if I need to do anything special with it to make it look good (e.g. cloth modifier for the cushions, or Z-Brush later on), or Identify any tricky areas such as where the couch arm connects to the back (something I learnt to do whilst modelling the pick-up truck last year).

So here are my rough concepts for the couch and armchair (as they're my largest assets and the most complex).

Couch:

Armchair:

Something I'd do next time which I didn't do this time, is actually draw topology over my concepts in tricky areas to work it out then, or at least get a good idea for how I'm going to do it.

Ok so first here are the assets I have to model:

• Couch

• Armchairs

• Pillow

• Couch Cushion

• Blanket

I started with the couch, which turned out to be a bit trickier than expected as there are literally NO REFERENCES of that style of leather couch (hence the need for a concept/best guess at what it looks like from all angles). I made this rather messy mood-board of as many different angles of the couch that I could find from the master-board to refer to whilst modelling.

Here is a shot of the initial high poly couch I tried to model alongside the lower poly one. I also struggled a bit with modelling the arms as shown in the two models, but I worked it out eventually:

And here is the final Low poly couch:

Something I've realised this year that I hadn't quite grasped last year is there are lots of different ways to model something, and you don't always need to make it all one model. Last year I'd have driven myself mad (madder) by trying to model the couch as one object whilst still keeping topology clean. This year I modelled it as seperate objects that slotted together. It's safe to say though that I was very rusty with modelling (I did model over summer, but I was modelling an organic mesh so my asset creation skills were more than slightly neglected).

I also initially made the cushions on the couch way too high poly (thanks to misuse of turbo-smooth and cloth modifier). I was advised to fake a lot of the information with normals, which made sense since I was planning on using normals for the couch creases anyway. So I remodelled the cushions, taking care to chamfer them in a way that didn't create triangles (as Z-brush doesn't like triangles).

A new work flow

I then took the couch into Z-Brush (EEEK). I'd looked at the programme over summer to familiarise myself with it's interface, but I'd never actually used it on a model I'd created before.

Thankfully I took to it quite quickly, seriously once you know the brushes you need to use to do something it's actually pretty smooth sailing. (I want to learn how to do hard surface stuff in Z-Brush too).

When sculpting the couch I decided it would be easier for me to sculpt the couch cushions and arms separately.

So here is my sculpt of the couch cushions:

And here is my sculpt of the couch arm:

The good thing about baking normals is to bake them you need your low-poly model unwrapped, so this sort of kicked me into getting my unwraps done as I go, which is something I'm not normally that good at, I tend to unwrap all my models in mass which is quite stressful.

So here is the unwrap of my couch:

Important note:

I initially unwrapped the couch cushions and arm as separate unwraps to bake them separately. But was then made aware of the fact that that wouldn't work very well when texturing the couch, as it would be silly to have three different texture sheets for it rather than one. This was a pretty easy fix though as it just meant combining the unwraps in Max by attaching the model back together as one object, and re-packing my unwrap. (I actually did this re-pack in week 3, but decided it made more sense to put it in this blog post as it means my modelling process can be seen more clearly).

I also learnt to start using marmoset as a quick way to see how your normals look on your model and fix the map in Photoshop. (I'm planning on fixing the maps next week once I've modelled and unwrapped everything) so for the time being I just used it to check that overall the bake had worked.

Here is my couch in marmoset with just normals applied (again I was silly and forgot to take progress screenshots so this is the couch with fixed normals-this feels like it's going to be an action plan thing):

So with this done it was time to model The Armchair.

Here is my reference board for it:

This model nearly killed me. I found the arm and back bit very very tricky, and after a few botched attempts(tried modelling it in various ways):

Decided that the best course of action would be to model the arm side and back bit seperately like so:

I then put it into Z-brush and wanted to scream again since smoothing the model would pull sections up to be shorter (because there wasn't enough topology I think. It also created a big seam between the two arm sections as they weren't attached so to fix this I divided the model with smoothing off, and then smoothed manually using the smoothing brush. I also masked the line where the two arm pieces meet before dividing so they didn't grow apart. It also meant I could smooth quickly manually without causing the same issue.

The sculpt for the armchair was quite a quick one but unwrapping it took a while, as again I'm pretty rusty still.

Here is the (initial)unwrap:

And here is the sculpt:

I changed the model and its unwrap quite a few times, and consequently it's sculpt. So my initial normal map for this armchair required a lot of fixing. I had a few problems with changing the pivot in max, so then my high poly and low poly models would be in different places in X-normal, then I had to change the size of a few things on my armchair, so I had to import my high poly into max to try and scale them as closely as possible to match etc. Then I changed how it was unwrapped, then how it was modelled, and so how it was unwrapped, essentially I changed it a lot, but because I was so caught up in the process whilst doing it, I didn't take multiple screenshots. The whole armchair is a bit of a blur to be honest. So here is the final armchair in Marmoset with normals applied:

The pillow and cushion were my easiest assets to model. You can see them in this still:

I used cloth modifier to get their basic shapes. (you can make them relatively low poly with cloth modifier, you just have to be careful with how many loops you start with). I then edited their shapes further in Z-brush using the move brush. Ordinarily I know you aren't meant to edit silhouette in Z-brush because that means you can no longer use your max model as your low poly one, however it was fine because I could simply use my lowest subdivision on Z-brush as my low-poly model. (since I didn't need to use dynamesh).

So here are the models:

Here are the low-poly versions in Z-brush:

Here are the high-poly sculpts:

Here are the unwraps:

And here they are in marmoset(their normals required minimal fixing, so this is how they looked pretty much from the beginning anyway, so in this case lack of before and after isn't so bad):

(masking was my friend for those pillow and cushions seams)

The blanket.

So at first I was a bit worried about how I would make this, but a third year gave me some good advice for it. Essentially I blocked it out in max, using very basic loops to get the overall silhouette done, then exported it and started sculpting it in Z-brush.

Here's the original low-poly:

The problem was I didn't get the silhouette right in max, which was fine whilst I still had my lower sub-divisions. However I had to use dynamesh a few times on the blanket due to holes getting created and weird line things that I don't understand, but that dynamesh fixed. Using dynamesh meant my lower subdivisions were deleted though, which meant, how on earth do I get my low-poly version of the blanket? The answer, Z remesh and polypaint.

What I did was use polypaint to paint on the areas of my blanket that needed the most detail (with the most saturated colour) so the front as thats what you see in the still, and then painted on areas of medium detail like the side (before it hits the couch).

here is the poly paint/sculpt:

I then used z-re-mesh to get a new low-poly model.

Here is the new-low poly model in max:

(as it came in quite high-poly from Z-brush I stripped it of loops and faces that wouldn't change the silhouette or be seen).

Unwrapping it was relatively simple, it just involved quick planar on the front faces and then pelt on the lower-poly back part.

Here is the unwrap:

and here is the blanket in marmoset, again, (normal map will need to be edited next week-this image shows a relatively cleaned up normal map, as again...progress shots..):

So thats my models all modelled, sculpted and unwrapped. Miraculous I know, I actually stuck to my timetable for once. And whilst I know it's not all down to one thing, this week I started to use a task book (to go alongside my course notes) and all I can say is, how have I lived without it?! It's helped so much with staying organised, on task and giving myself more manageable tasks to do.

This week in game production we went over shaders and PBR which I found fascinating. I love learning about the more technical side of the course, and in my opinion understanding something helps you to execute it better. So the more I understand about how the different maps work for example, the better my maps will look. (In theory).

Visual Design was about feedback on our projects which you can see applied above in the Engine section of this blog post.

In critical studies we looked at critical analysis of an image. This involved talking about an image personal to us and breaking it down into different points. For example looking at the backstory (if it's relevant to the critique), looking at the different elements of art, for example how the images uses composition or different artistic techniques, and finally interpretation and themes that you could draw from the image. Going over this definitely has value as it will help me critique my own work, and others, which is something I am trying to get better at. I certainly suffer with wanting positive affirmation as opposed to critique, however this year I want to start taking critique on board a lot more, and have the patience to re-do things (i'm bad at this, when it's done, I want it to stay done, but I'll never improve that way, so suck it up I guess).

So the part you've all been waiting for. The...

Action Plan

Project wise, next week I need to:

• Fix my normal maps.

• Check my models scale in the whitebox

• Export them to be put into engine (this involves all the normal checks; n-gons? X-form, pivot point, centre to world, light maps, and smoothing groups).

• Plus I want to have all my models named appropiately with clean max scenes.

• I will also clean up my files as this will make hand in a LOT easier.

• I want to have a team meeting to update the asset list and get everyones work into engine.

• I'm also going to be texturing, and if I have time start work on the stretch assets.

This week I also noticed that I struggled to do my blog as I hadn't been great with progress screenshots. I also don't work that efficiently and I burn out quite quickly. So I decided that to tackle these issues I will adopt a 'do things as I go' approach. Essentially this means taking progress screenshots regularly as I work. Saving my work neatly as I go, cleaning up files as I go, and staying more on task. This means coming into labs, blocking out distractions and getting my tasks done so that I can leave at 8 rather than 9 (Baby steps to healthier working hours). I also decided I want to stop work at 7, do an hour of personal work (for sanity reasons) then go home. I also want to give myself the evenings off, something I didn't do last year. This should help prevent me from getting sick from stress anymore.

I'm also going to try and write my blog in personal tutor sessions rather than on the weekend. I find writing around other people difficult but I want the weekends for myself, so it's something I'm just going to have to get over.

I also need to get back into the workflow of do it, move on, if I have time go back and do it better. Rather than wasting time polishing one asset and then having to rush the others.

Finally back up the back ups of the back ups. (I'm buying a new harddrive).

Here's to a more productive next week: