Seventh Eve

Seventh Eve: Final assessment video

2010-05-27T15:53:00.002+01:00

As the final project has now been handed in, here is the final assessment video for Seventh Eve:

BLOG NAVIGATION

2010-05-25T10:32:00.000+01:00

The contents of this learning blog cover two major projects- Seventh Eve, a 2D puzzle gamer based in ancient China, and a selection of conceptual paintings for the professional practice module.

To read posts relating to the game Seventh Eve, please click here.

To read posts relating to the professional practice module, please click here.

Animations

2010-05-25T10:02:00.000+01:00

Referring back to my previous post (Re-rendered animation), I have been using an old 3D environment as a background for rendering my animations, to provide a better environment than simply rendering out on a grey presentation background. At this point in my project, I haven't got the time to animate any further models, and need to be focusing on finishing my post-production document and the final assessment video, so I am simply neatening up and rendering what animations I have, and will return to finishing a few more small animations if I have time afterwards. These are the animations so far:

Re-rendered animation

2010-05-20T23:07:00.000+01:00

Whilst looking through what animations I had already rendered to compensate for the damage to my home computer (refer to this post), I found the file for the 3D environment that I had made for my original pre-production animatic when my idea had still been to create the game in a mix of 2D and 3D gameplay.

Though this idea later changed to encompass an entirely 3D game, now that I have no way of rendering my work in the final environment I wanted to at least be able to show a small part of the final video within an environment, and all I needed to be able to render a video within this one, as I already had the basic modelling and the camera animation, was to import in the animated Zhinu model and add the sky.

In the final presentation video, there will be mock menu buttons overlaid over the video to simulate gameplay.

Computer crash

2010-05-18T10:57:00.004+01:00

As of yesterday, my home computer has stopped working due to a virus that's shut down the system. I've managed to back up my entire project up to this point on an external harddrive- stopping myself from losing any of my work- however, this does cause a large issue when it comes to finishing my projects.

Though I covered the potential issue of technical failure in the risk assessment in my pre-production document (In case of technical failure: I will make sure to back my work up in a number of places- on an external hard drive, on two separate hard drives on my personal computer and on the network at university. In case of my graphics tablet breaking, I have a spare available and also have 3Ds Max on two separate computers at home in case of any problems. I have also changed the backup settings on 3DS Max to back up all open files every five minutes.), I didn't consider the loss of my home computer to be an issue as the university studio has plenty of computers that could accomodate my continuing work.

However, referring back to a previous post, here, the problem becomes obvious- I cannot render anything within my environment on any computer bar my home computer. I do not have access to any other computer outside of the university that has 3DS Max installed on it to see if the environment will work on them, and cannot get hold of a copy of the programme to install it on another external computer, though I am looking into trying to do so before my deadline gets any closer.

This means that, though I still have a week and two days left to animate and work on my models and project at university, I cannot render anything that involves the environment. Considering that the environment was the main focus on this project- environment design being the area of the industry that I wish to specialise in- it's clear that this is quite a major issue, but at the moment I cannot think of any way to resolve it.

I will still be able to produce a final video (I am still aiming for three minutes as outlined within my pre-production document), chances are that it won't involve anything within the environment past what I have already rendered as test animations. However, my intention is still to showcase my animations upon grey presentation backgrounds- showing at least the work that I have done even if I am potentially no longer able to create a full animatic that follows the promised storyboard from my pre-production document.

Animations in the environment

2010-05-16T16:22:00.002+01:00

Due to the way that the environment is built, it's difficult to tell how the animated elements of the level move without rendering them out fully first. As a result, here is a short video of the opening shot of the level incorporating the lanterns, the playable character Zhinu, the falling peach blossoms, the star charts and the clouds. Unfortunately, after rendering I could tell that the clouds are moving both much too fast and much too exaggeratedly, so I'll be changing this next to make them more subtle.

Pro practice: colour and light

2010-05-06T19:29:00.006+01:00

As another experiment to explore light and colour, and again aiming to create a piece within 45 minutes, I have been looking at the effects of light- reflections, light bloom, etc, to experiment with the best way to paint them effectively.

Though I've spent about 45 minutes on the above piece- the church- I like the way that it is starting to come together and will spend a little longer on finishing at least the silhouettes and shapes of the architecture and light.

Pro practice: Norse environments

2010-05-01T17:45:00.000+01:00

Trying to speed up my painting now, I allocated myself with 45 minutes in which to complete a full painting. As a loose concept, I wanted to look at Norse mythologies and the Scandinavian landscape. However, I found that I was spending too long researching into the histories and mythologies of this culture and so, rather than get distracted with elements that wouldn't have much of an impact on the visual result of my work, I ended up creating a couple of pieces based on the atmosphere of northern european environments.

Pro practice- Arabic environments

2010-04-29T12:10:00.007+01:00

In trying to speed up my painting and concepting painting, my first work for my pro practice module has been to try and create an Arabic-style environment. The main concept for the work was a lived-in environment based in the disused underground water tanks that are found commonly in Morocco. This would give me a good chance to experiment with light, water and contrasting warm and cool colours.

The above is a series of pieces showing the painting in progress. It's unfinished, but I've already spent too long on the piece and so I will move on and, if I have time and feel as though it will make a good finished piece, I'll return to it.

Incompatibility issues with environment and university computers

2010-04-28T11:13:00.001+01:00

In finalising working with animating the textures in my environment, I have come across an issue that may slow my work but shouldn't be too much of a major issue. For some reason, the university computers cannot render out my environment properly- I have tried using a number of different computers now, and all of them will not render the alphas in the textures properly, making the majority of the texture planes fully opaque and thus blocking out the rest of the environment and the models within it. For those that it does render correctly, it renders them at a very low quality despite the texture quality itself being exactly as it should- this is not related to the renderer used or the quality at which the AVI or JPG render is saved, as I have also tried changing these, although to no great effect.

(This issue was covered briefly in this post, though at that time I wasn't aware that the issue might be one with the computer as opposed to a setting that I was using- however, after some testing between the computers here and mine at home, I have managed to come to this conclusion.)

Though this means that I cannot render out any animation within the environment at university, my home computer handles these files perfectly well and renders them quickly and at the ideal quality, with no rendering issues- therefore, my plan is to use my hours within the university studio to animate my models and merge them into the environment, and then render them out overnight on my home computer. This shouldn't slow my workload too drastically and I'm still relatively confident about finishing or at least getting a large portion of my project finished in time for the deadline.

Professional Practice brief

2010-04-21T15:05:00.001+01:00

Though not outlined in my pre-production document, I intend to create a selection of work towards a professional practice module (a module dedicated towards creating portfolio pieces to demonstrate the level of our work in our chosen speciality). I realise that I am behind in my schedule for my Seventh Eve project- however, due to feedback received from industry professionals from Climax Studios (written in more detail in this post) I have realised that it is important for my personal employability to also have a number of pieces in my portfolio that demonstrate my abilities outside of the style that I've been working in for the Seventh Eve project, and therefore I will sacrifice some time from this project to fulfill the professional practice brief.

As also written in the above linked post (Industry visit from Climax Studios), my original intention for my professional practice module was to create a number of paintings that would serve as a series of environmental concepts for a fictional game, and to then recreate one of these in 3D.

In looking at existing portfolios of environmental artists to help assess what level of work would serve to boost my employability, I came across and focused on the portfolios of Nicolas Ferrand (website), Raphael Lacoste (website) and David Levy (website). By looking at the work of these artists, i could assess the main level of work expected in the industry- work that displayed a good grasp of perspective, light, colour theory and, presumably, the ability to produce a number of pieces that demonstrate all of these above within an acceptable speed.

Looking at these, I have concluded that I am not yet at this level of working. Though I am comfortable with painting and have a confident grasp of perspective, my application of colour and light is not yet at an industry standard and, above all of this, my speed when painting is slow and would not lend my abilities well to working within the industry.

Instead, I will first focus on creating a number of environmental concepts that will test my abilities in all of the above listed areas. By pushing myself to paint faster, and apply concepts of lighting, colour and perspective, I will hopefully be able to increase my ability at conceptual painting. Due to my being behind in my work- both through bad scheduling on my behalf and my failed extension appeal, I suspect that I may not be able to produce any work past this point, although if I do I hope to continue with my original concept of creating a series of connected environment concepts, and perhaps begin to model at least one of them.

Animated textures in the environment

2010-04-19T11:46:00.000+01:00

I've left inserting the animated textures into the level environment a little later than I should have, but it hasn't taken more than a day to get them inserted and working and now the level is fully animated. The only issue I found where I wasn't sure how to achieve the right effect was with the peach blossoms beneath the trees dotted throughout the levels. With the blossoms on a separate plane there was a visible line where the plane ended, which ruined the effect that I was after. Though I'd used vertex colours before, I wasn't sure how to use this to create a fading alpha effect. After some discussion with Adam, I have managed to create the effect that I was after, as seen here:

Blossom texture test:

Test in the environment:

As the above video was rendered out at very low quality just to show the texture in the environment, it's difficult to see the blossom- however, it does demonstrate an issue that I'm currently having with the environment textures. Though Max shows them in full detail whilst working in the programme, and the texture files themselves are full quality, when rendered out in Max they're coming out pixelated and fuzzy. I hadn't realised this until rendering just now, and I can't seem to find any way to remove the artifacting. I have tried rendering it out through different filters, changing the resolution and rendering size of the final video, but it hasn't yet fixed this issue.

Using a dope sheet and a visibility track, I've put simple animations on the stars and the glows on plants to make them fade in and out throughout the level. I've also put animations on the waterfalls to give them the impression of flowing water, and made the star charts in the sky rotate slowly. The clouds also shift backwards and forwards to give the environment a more realistic effect.

Above: a quick render showing the fading glows on the plants. You can also see in the background the turning star charts- in the final video presentation, they will turn turn slower and be less opaque to merge better with the sky behind it.

Final failed extension appeal

2010-04-18T15:44:00.001+01:00

As of this afternoon I have been notified that I've failed my final appeal for an extension. Though frustrating, as I've found it impossible to catch up with the few weeks lost after the bereavement and my illness, I had suspected that this might be the case and am simply working as hard as I can to catch up the time lost.

Magpie animations

2010-04-18T15:24:00.004+01:00

Final magpie animation- idle, taking off and flying:

Magpie animations

2010-04-17T12:05:00.006+01:00

The most difficulty I have had in animating the magpie model is finding good, legible filmed references of magpies. All birds have different flight patterns, and so to make sure that my animations are accurate I've had to make sure only to use references of magpies specifically. However, I have managed to find references for most of the necessary animations for the magpie, by working from these videos:

Magpie animations:

Idle:

Taking off:

Flying:

Standing animations

2010-04-16T17:37:00.001+01:00

Another requirement from the storyboard is for both the jing and Zhinu models to rise to standing up from either lying down or sitting. Due to the clumsy rigging, the animations for these moves are a little unsteady, but as with the other animations if I have time towards the end of the project, I will return to them and improve them.

With the Zhinu animation (the second), keep in mind that she is sat on a ledge, hence having to pull her legs up and over it first before rising.

Idle animations

2010-04-13T22:21:00.001+01:00

In working on the animations for Seventh Eve, I have now completed a couple of idle animations for both the huli-jing and for Zhinu. Though these are not as polished as they could be, my aim is to complete all of the necessary animations to a standard that will show my understanding of the basic principles of animation (follow-through, weight, flow etc), and then if I have time left towards the end of the project I will improve these.

Huli-Jing idle animation:

Zhinu idle animation:

Failed extension appeal for deadline

2010-04-13T13:48:00.001+01:00

In referring back to this post, where I had to take a break away from university work due to a family bereavement and had been offered a deadline extension by the university, I've received an email this morning saying that my formal appeal for an extension has been rejected. Though I am currently appealing this rejection, chances are that I will not be able to extend my deadline and therefore find myself with two weeks less time than I had planned for.

Unfortunately, this cannot be helped, and though I realise now that the chances of having both Seventh Eve and my professional practice module fully finished in time for May 26th are considerably lower, I'm going to do my best to finish them to the most complete standard possible.

Mistake made in relation to DS capabilities

2010-04-13T11:37:00.000+01:00

In referring back to Sourceforge.net, the site I've been using for detailed information on the DS' rendering capabilities, I've realised that I've made the mistake of remembering wrong the texture memory size that I should be working to. Rather than constantly referring back to this source, I have instead been working from memory, and this has mislead me to believe that the DS can support 256MB worth of textures as opposed to 512KB, which is the actual figure.

There's quite a large difference between these, but as the realisation has come this late I cannot afford any time to go through all of my textures (the environment alone using 489 different texture maps) and reduce the size and quality of them. I had previously begun doing this (refer to this post), and so the work has already been started, but at this point in production I can only afford the time to fix this once I have completed the rest of my project and professional practice module.

Completed music

2010-04-12T18:38:00.001+01:00

Thanks to Phil Derham, I now have the final track that will be used for my final presentation video. The completed track is three minutes long, which will allow for an animated mockup of gameplay and then a selection of model turnarounds, concepts and the original 2D work to display the development of the game.

Environment grid issues

2010-04-11T19:11:00.000+01:00

Though I had previously decided to leave the grid issues on the environment until the end of the project and then fix it if I had any fresh ideas on how to do so, when meeting with Joe Swan (a previous graduate from the course and current lecturer) we decided to spend five minutes addressing the issue.

When I'd previously tried to fix the issue beforehand, reattaching the planes back into one large plane had messed up the textures that I'd already put down. However, with Joe's help I managed to bypass this issue, and we spent some time reattaching and welding together the planes in a bid to reduce the grid effect.

As can be seen in the two renders below, however, reattaching and connecting the planes created no visible reduction in the grid effect. In this case, I will return to my previous conclusion that the grid isn't significantly visible, and will only fix it if I can come across a suitable fix for the issue.

Final rigs

2010-04-10T14:31:00.000+01:00

I have now finalised the rigs on all of the models for the project. In regards to my previous post (Starting animation and finalising rigs) regarding the Zhinu model and re-rigging her skirt to create an impression of legs beneath the skirt, with help from Kris Doggett (a class peer) I have managed to take an alternate route to create the same effect.

As opposed to creating new bones beneath the skirt to create the effect of a limb, and then applying an IK limb solver, I have simply created a new dummy in front of the existing bones, then applied a look-at constraint from one of the middle bones of the skeleton to the dummy (see below). This means that the front of the skirt can now be pulled forwards, mimicking the effect of knees beneath the skirt. As the model is low-poly and the game itself is highly-stylised, a fully realistic rig composed of real leg bones and paddles to create full skirt movement is not necessary, and therefore this simplistic rig will suit the project perfectly.

The final rig for Zhinu:

Final rigs for the magpie and huli-jing:

I had originally intended to create the magpie rig entirely out of broken mesh, but due to the way that I had modelled the geometry in the wings this wasn't possible. However, by using the bone structure shown above it's now possible to control the wings a lot more easily, the horizontal bones allowing for a smooth curve to the wings and the two sets of vertical bones meaning that I can control the direction in which the edges of the feathers move when animating positions such as flying and landing.

Starting animation and finalising rigs

2010-04-09T11:35:00.000+01:00

Now that my environment is fully moved into Max and ready to work with, I am about to start my animation- this is quite a way off of my original schedule and the chances are incredibly likely that I won't have it finished for my deadline of the 19th of April, but this is through personal mistakes that I have made and will address in my final review of the project.

Before I can start animating, there is only one last piece of rigging that I need to address- Zhinu's skirt. I have already rigged it once- using a series of bones attached to a dummy beneath the skirt that allows it to swing backwards and forwards, and from side to side realistically, but to take into account crouching and kneeling positions, the rig contained only the ability to scale the skirt up and down so that the model could be lowered, which doesn't realistically replicate the way that a dress would act when someone kneels in it.

Above: renders of the current skirt rig and a screenshot of the rigging set up.

The current 'scaling' skirt rig. Though it gives the impression of the body lowering as the character crouches, it's obvious that the skirt has simply been scaled and doesn't deform in the way that a skirt would if there were legs beneath it, as well as showing no mass beneath the skirt where the legs would be when kneeling.

Below: a mockup of the intended new skirt rig

Final environment review

2010-04-08T10:28:00.000+01:00

I have now finished transferring my environment into 3DS Max. Though the grid issue is still present, there are only a couple of areas within the environment that it is noticeable (unfortunately, these are rather prominent areas but hopefully once the Easter break is over I will be able to sort this issue out relatively quickly with help), and with the DS at such a low resolution screen (256 x 192px) it shouldn't be too obvious.

The final level render from a front-on perspective:

The same perspective with the grid issues outlined:

The level from a three-dimensional perspective to analyse the grid from another angle:

The final level comes to 490 tris worth of planes, which brings my final tri useage to 3704 of my 4048 tri budget. However, this is only if the DS were to render the entire environment, plus all props and characters on screen at once, which would never happen within the game, so even though I have strayed relatively close to my budget it would not cause an issue with the DS' rendering capabilities.

Environment rigging

2010-04-07T11:36:00.000+01:00

Now that I have finished reducing my environment textures into squares, I have begun transferring them through into Max to create the final level. As per my previous post on the subject, I am still having an issue with a very faint grid showing on the render of the environment. The main issue that I am coming across in fixing this is my mis-reading of the university Easter break- I had believed that the break finished on the 12th of April as opposed to the 19th, and so thought that I would have a week's worth of teaching time to ask for assistance on issues such as these. Due to illness, I also missed the last week of teaching before Easter and have been able to contact any peers who may be able to help me, although I will continue to build the level so that I at least have the structure of it for when I can find help on this.

A render to demonstrate the grid-line issue:

I am currently using the technique of creating a single plane of 10 x 6 squares- the equivalent of the 5000 x 3000 pixel canvas that I used to paint the environment- and then using the 'detach' tool to separate it into separate planes, which allows me to apply separate textures to each square (I had tried this without detaching the planes, but Max was unable to handle this and would not render the textures properly), and also allows me to make sure that the planes are completely flush next to each other, as opposed to trying to line the planes up by hand.

Environment textures- finished

2010-04-06T21:30:00.000+01:00

I have by now finished cutting the environment painting down into suitable, useable square textures. Before resizing down to 128x128 and changing the format of the file, the squares produced from the original painting have been 512 x 512px, and full-size quality jpgs, however even at this size the total texture memory of the entire environment only reaches 13.8MB- a consideral amount smaller than I had taken into account. Taking into consideration that the DS can handle 256MB total at one time, this leads me to believe that rather than having to sacrifice quality for size, I will be able to use the original-size textures, as well as the larger, better quality textures that I produced for my models.

This should also save me from having to render out two final videos- one at DS quality and one at high presentation quality, which will allow me more time to spend on my professional practice module.

With all of the textures that I have, this puts my final texture usage at:
Environment: 13.8MB
Characters: 9MB
Props: 370KB

Total useage: Approx. 23.5MB

Environment texture sizes

2010-04-06T12:04:00.002+01:00

Following on from my previous post regarding the DS and its technical abilities, I will have to cut my environment into a number of square texture maps to retain a suitable amount of detail without going over the limit of the DS' 256MB texture memory.

In cutting the environment into squares, I have managed to solve this issue in several ways. For example, the very back layer of the environment, the back part of the sky (see below), has only a couple of areas of intense detail as opposed to the rest of the texture, which is mainly just darker patches of colour, cloud and stars.

This means that in cutting the texture down into separate texture maps, there will be only a number of maps that will require a higher quality to retain their detail. My conclusion is to create the majority of the texture- the areas with less detail- as 128 x 128px texture maps with in 16-colour indexed format, which reduces the average file size of these maps to 550 bytes. Then, in the areas of texture that require more detail, I will save them as 256 x 256px texture maps, in 32-colour indexed format, which allows them more clarity but still keeps the file size at under 5KB.

An example of a low-detail texture map and a high-detail one:

In this manner, I should be able to retain the detail and definition of the majority of the environment whilst keeping the required texture space beneath the 256MB limit, which has been one of my main concerns in regard to keeping this project within the DS' capabilities.

Transferring the environment into 3DS Max

2010-04-05T17:55:00.006+01:00

I've started converting the painting of my environment into a format that will be suitable to transfer into 3DS Max whilst still taking into consideration the DS' technical capabilities.

As mentioned previously in my blog and my pre-production document, the main issue with transferring the environment will be the DS' texture rendering abilities. The largest texture map that it can handle at once is 1024 x 1024 pixels, and the largest amount of texture memory that it has is 256MB; this means that, as my original environment is approximately 5000x3000 pixels, there will have to be some considerate rethinking involved. Shrinking the entire environment down to 1024 pixels wide would be the most suitable answer in terms of time and effort- however, due to the size and detail of the level this would only cause a significant loss of detail- something that would only detract from the final visual impact of the game.

My solution is to construct the environment in Max out of a number of square planes. Then, by cutting the painting of the environment into square textures, I can retain the details of the painting without having to sacrifice them in favour of the DS' running abilities.

In Max, the final construction would appear something like below:

By creating a quick test construction in Max, I managed to render these results:

The only issue that this technique creates a very faint grid around the edge of the planes where they meet and overlap ever so slightly. In the same test I then created one large plane that was then separated into a grid, and applied the same textures to each part of the grid to create the same effect without the interlapping issue; however, this technique is only rendering solid colours as opposed to the texture and its correct alphas and therefore I need to recreate the test in a number of different ways before I can work out the reason for this.

The test with one solid plane (left) next to a construct of four separate planes (right) that demonstrates the current issue:

Posed model renders

2010-04-04T18:42:00.008+01:00

In testing the model rigs with the textures applied to see where there might be issues with texture stretching, I have saved a few posed renders of the finished models.

Huli-Jing:

Magpie:

Zhinu:

Seventh Eve music

2010-04-03T22:07:00.002+01:00

I've spent the majority of today with Phil Derham discussing the music for the final video of Seventh Eve. As the previous track was too heavy for the game we've raised the key of it, changed a number of instruments used and will now be recording a layer of vocals that will replace elements of the initial track that didn't work.

The track should be completely finished by April 10th.

Magpie rigging tests

2010-04-02T12:38:00.004+01:00

My main concern in the rigging for the magpie model is the flexibility of its wings. Out of the three models, it is the rig that poses the most issues in terms of making the wings believeably poseable in the three key positions- fully outstretched, folded against the body, and stretched right back.

Though it is not the most graceful of movements, I have managed to create a rig that can plausibly fold the wings flush against the magpie's body, and as the model will occupy such a small space on the screen in the final animations, hopefully it will not be an issue. The videos below are a simple test, and I intend on the finished animations to spend more time on the movement to make it smoother and more plausible.

Zhinu textures

2010-03-31T11:31:00.003+01:00

I've finished texturing the model for the main character, Zhinu, today, which should be the last model that requires any texture work for this project.

Out of all of the characters, Zhinu's model has the most detail and as such I've made the decision to leave her texture sheet at 512 x 512 pixels so as to preserve as much detail as possible. As none of the other characters' textures have been huge in terms of file size, this shouldn't be an issue in keeping to the DS' texture memory of 256MB. However, if the environment (which is the element which will take up the most texture memory) turns out to require more space, I will amend this and make Zhinu's texture smaller.

Magpie textures

2010-03-30T15:54:00.001+01:00

After a couple of weeks break that I've had to take off due to illness, I've finished off the texturing for the magpie, as seen below:

Below are the texture sheets for the model. The entire model has been done on one 256 x 256px sheet, with an alpha channel for the wings.

The separate alpha channel for the wings:

Illness and schedule delays

2010-03-29T18:49:00.004+01:00

Since March 21st I have been bedbound with the flu. This has made progression on the project difficult, and I am now working solidly towards having it finished on the rescheduled deadline of 19th of April.

Huli-Jing textures

2010-03-18T18:04:00.009Z

In order to keep to my schedule, finishing the textures for my characters has been put behind more pressing priorities- aka rigging and finishing the environment painting- for a while. However, I have finished the texturing for the Huli-Jing now, as can be seen below:

Below are the texture sheets for the model. The first is the texture for the face and body, and second is the textures for the hair- these have required a separate UV map to provide the alphas that are necessary to give the same silouhette to the hair that is in the concepts.

The separate alpha channel for the hair:

Final texture sizes are both 256 x 256px.

For reference, the original concept turnaround:

Huli-jing rig and animation tests

2010-03-17T19:49:00.000Z

The rig for the huli-jing is now fully completed. I haven't yet included all of the controllers needed to make animation quicker, but the model is fully posable and I have rendered some early animation tests to be able to see where there may be issues with the rig:

Zhinu rig and animation tests

2010-03-16T20:08:00.000Z

The rig for Zhinu is almost fully finished now. The only issue I am having is with her skirt- though a basic dummy that can be scaled has been linked to it so that a believable bounce can be added to her walk, and the bottom of it can be moved back and forth to recreate a realistic swing to the skirt, I am having trouble adjusting the rig so that she can kneel and sit with a realistic bend in the skirt. I'm thinking of taking out the current bones in that part of the rig and replacing them with a simple, three-bone rig that will mimic a single leg, which can then have an IK limb solver applied to it.

However, the rig for the upper body is fully working now, with only a little stretching in some of the geometry, but with such a low poly model I haven't yet been able to find an alternate weighting on the bones that will alleviate this problem.

I have rendered some early upper body animation tests to get a better idea of the stretching issues:

An issue here is the stretching in the underarm area. I am still working out the finer elements of the weighting across the back, shoulder and underarm area, but as the texture here is plain I hope that any stretching that may remain won't be too obvious.

Another issue here is again with the underarm area- where it folds there is some geometry intersecting with the torso. The only element of the sleeves that are controlled by a separate dummy are the ends of the sleeves, and therefore the upper arm is controlled entirely by one bone- however, this makes it very hard to solve this problem.

Once texturing is finished for the model, I will return to this problem and assess how obvious it is and how best to weight those bones with minimal texture stretching. When rigging is fully complete I will post up screen captures of the final rig.

Environment rigging- texture issues

2010-03-15T19:22:00.000Z

Now that I've finished the painting of my environment, I've been using the past few days to convert all of the different texture layers into the appropriate formats for rigging the environment.

However, I've come across a major problem that has been an incredibly bad oversight on my behalf- the DS engine can cope with a maximum texture file of 1024 x 1024 pixels, and has a maximum texture memory of 256MB- this includes not only the environment but also prop and character textures as well.

Unfortunately, I didn't take the environment into consideration when first considering these restrictions. It is possible for me to crop the textures into 1024 x 1024px grids, but my problem will be in keeping the file size below the 256MB budget as all of the layers, thus far, will come to approximately 500MB.

A possible solution would be in rendering at the resolution of the DS screen. Each DS screen can only render at 256 x 192px- if I take this into consideration and shrink all of my texture files to acommodate this, it may be easier to keep to the memory budget. However, as this would make any final video render difficult to assess, I could then create two final renders- exactly the same content, but with one displaying the higher-res, more memory-heavy textures for assessment and then one that adheres to the DS specs to prove that it was possible to keep the game to my original technical specifications. All this would require is creating the final animations and video, render it at high spec and then swap the textures out afterwards for the lower resolution ones and render that at the DS specs of 256 x 192px.

Final environment

2010-03-10T23:44:00.000Z

Though quite a bit behind schedule, I've just finished the painting of the final level for Seventh Eve. Due to time constraints, my own mistakes in allowing myself to get behind schedule, and personal matters I've had to cut out the final part of the level- the garden beyond the gate- but there is enough of the level still remaining that I will be able to demonstrate all of my learning outcomes and show off the majority of the environment as it was intended.

When animated, much of the environment will move- the star chars in the sky will rotate, as well as the clouds, which will shift, the stars will fade in and out and the waterfalls will also have animation on them.

Click the thumbnail below for a full-size shot of the final environment.

Industry visit: Climax Studios

2010-03-03T16:23:00.002Z

Today, students on the course were visited by prominent members of Climax Studios who would be able to give us a quick review of our projects, suggest weaker areas that we might be able to improve upon and give us an idea of life working in the games industry.

Visiting were Stuart Godfrey, the recruitment manager, and Glenn Brace, the art director for the studio- and I managed to have a few minutes to show them my project and get some feedback from them in terms of how my project presented itself in terms of employability and lending itself well to a professional portfolio. The feedback that I received, in general, backed up what I had already concluded about my project- that my 2D environmental painting was strong but the project, in its finished stages, wouldn't exhibit much in terms of showing off any 3D modelling skills.

The main issue was that, as a highly-stylised game, the project would only display my abilities in one niche area. From the beginning, my concepts had focused on a minimalist, simplified style and as such I had no work proving my skills in concepting or modelling in higher detail or higher poly. Though my painting and concepting was strong for the project itself, in terms of building up a solid portfolio I needed more pieces that would show off my skills in other areas.

However, they agreed that the work that I have produced for the project is a good representation of work for a game of its genre and for its intended demographic, and though their main concern would be the mixing of 2D environment and 3D models they agreed that if it is done correctly it should work well with the original concepts of the game.

Upon mentioning the professional practice module of this year, it was suggested that, in this module, I should focus on creating a series of works based on realism and high detail, to showcase a talent for more than just stylised work. From this feedback, I intend to create a series of environmental concepts for realistic environments, and then convert one of these into a fully modelled and textured high-poly environment. As I have little intention on entering the industry as a character artist, I will be focusing my entire professional practice work on environments and concept painting.

Break and schedule extension

2010-02-18T00:35:00.003Z

Due to a family bereavement, I will be taking a short break from this project. In lieu of this I have applied for a two week extension, also, and so rather than aiming to finish on the 26th of March the project will hopefully be complete on the 19th of April instead.

Final models

2010-02-13T22:03:00.001Z

The majority of the modelling for the final characters has now been finished. Though there's still room for improvement in them, I ultimately need to be moving onto rigging and texturing them to catch up on my schedule, and rather than fall behind I believe that they are at a stage where I am happy to leave them.

The largest problem with Zhinu's model is getting the animation loops right in her arms- as they have large sleeves the elbow loop is distorted and I am having to take into consideration the fact that though the arm will be controlled by bones, I will probably use a separate dummy to control the end of the sleeves so that they can distort when her arms are down by her side, stopping the geometry from intersecting with the rest of her model.

The huli-jing has caused a lot of problems also in his modelling- not only has it taken some while to work out how to model his hair without causing a huge increase in his poly count, but it has taken a while to work out the best way to model it that it won't get in the way when posing him at a later stage. Also causing a problem is the fur on his legs- I have ended up using the same technique employed on the magpie's wings (see previous post, here), but as these are positioned over the joints of his legs I have had to take into consideration distortion when animating.

The main issue with the magpie has been its wings (see here)- however, once that was concluded modelling has proven relatively simple and uncomplicated.

Prop renders and wireframes

2010-02-08T20:16:00.001Z

Peach: 80 tris
Texture map: 128x128px

Lantern: 192 tris
Texture map: 128x128px

Closed scroll: 56 tris
Texture map: 128x128px

Open scroll: 30 tris
Texture map: 128x128px

Flat scroll: 12 tris
Texture map: 128x128px

Further work on magpie model

2010-02-06T15:19:00.001Z

After considering all of the techniques in my previous post regarding the magpie's wings (here), I have settled for using solid geometry and texturing an alpha along the edge to create the required silouhette.

I have also hugely cut down the amount of polys on the tail, and will texture the feathers in as opposed to modelling them.

Magpie: wing modelling

2010-02-04T11:54:00.001Z

As the smallest character on screen at any point, my aim is to make the magpie the most low poly of all of the characters. However, as a bird, this is starting to cause some problems- mainly, in the wings and the tail.

Looking at the concepts (see here), the tail and wings require distinct edges that will help the silouhette of the magpie read better from a distance- this being of high importance, as the magpie will only occupy a small area of screen space and therefore readability is key. However, creating a distinct shape like this will most likely raise the poly count considerably, and this is why I've taken several days to consider the problem.

In looking at pre-existing models of low-poly birds, and in speaking with peers and industry professionals on Solent Gamers (a membership is required to view the content of this forum), there have been several suggestions in how to tackle the problem.

The main suggestion has been to model only the body of the magpie, and to treat the wings and tail as alpha planes. This would give me the ability to create only three planes, and then create the silouhette with texturing- however, this would only create a fully successful silouhette from one angle- the front- and as the model needs to be seen from all angles, as well as being animated, this technique doesn't quite fill what is needed.

Above: wings modelled with only an alpha plane.

Building on this idea, I've given some thought into padding the wings and tails out with several more alphas- this would still keep the model lower poly than if they were modelled as full geometry, but give them more bulk than if they were simply one plane.

Above: wings modelled from several alpha planes.

Though this is giving me more of a shape than before, I'm still having the issue of the body only looking right from one or two separate angles, as opposed to giving me a solid 360º turnaround. Also, the alpha plane still requires a certain amount of polys to make it animate smoothly, and by tripling or quadrupling the number of planes to give it a three-dimensional effect, I've ended up using more polys than I would if I were to model the wing in solid geometry.

Thinking back to a previous project undertaken last year, I modelled a higher poly bird that employed a similar technique of mixing alphas and geometry. As keeping things low poly was no issue in this project, rather than making the entire wing an alpha plane, a plane was simply run around the edge of the wing to create the desired feather effect. However, there was still the issue of visible joins between the wing and the alpha, though with the magpie model being so small this may be a viable solution.

Above: a previous project showing another way of joining geometry and alphas.

After several attempts at further solutions, I've come across one which solves the problem- it's relatively low in poly-usage, gives the desired silouhette and gives a believable shape both from a distance and close up.

By using solid geometry, but keeping the edges very narrow, I can get away with applying an alpha to the texture of the entire wing- which will give me the desired shape and bulk without leaving any joins that would be unavoidable when combining single planes and geometry. There is still a very small gap where the alpha makes it possible to see inside of the wing, but this is so small and the model will occupy such a small amount of screen space that I can't see this causing a visible problem in the final project.

3D placeholders

2010-01-28T11:02:00.001Z

Creating a series of very simple 3D models for each of the characters and props has helped to give me an idea of how many polys I will be needing to dedicate towards each character. In general, the DS can render 2048 polys per second at a frame rate of 60fps (as of here); as a point and click 2D scroller, it isn't necessary for the frame rate to be so high and by lowering it to 25fps, the engine efficiency can be raised to approximately 4000 polys per second.

Placeholders for the huli jing and the magpie. If I am to keep this within spec to run on the DS, I can afford between 700 and 1000 tris for each character, which gives me 1000 tris remaining for smaller props, such as scrolls, peaches and other items.

The model for Zhinu, developed past the placeholder stage. As the main character she will probably be higher poly than the magpie, but around the same as the jing as the both of them share approximately the same space on screen.

Some props developed to give myself a better idea of how many polys I will be looking at for smaller items. The lantern is 224 polys, the closed scroll is 88 and the open scrolls are 44 and 62 polys respectively.

Environment development

2010-01-21T14:21:00.000Z

From the final concepts that I created for my environment, I've needed to compose a final environment that lends itself to a 2D world. The level itself will be made of a series of planes, each at varying distances from the camera so that they will give the impression of being in a three-dimensional environment.

A very early test of this paralax effect can be seen here:

Taking into consideration the elements of my environmental concepts that feedback from both peers and industry professionals, the final environment of my level is as below.

Following is the storyboard for the final animatic, followed by the level blockout again, but with the storyboard annotated over it to display the critical path of the level.

Final video music

2010-01-20T21:35:00.007Z

For my final assessment video, I have asked Phillip Derham to compose a short, original track that will match the atmosphere and style of Seventh Eve. The initial track that I have received is three minutes long- the length of my final assessment video- and incorporates piano, panpipes and Chinese flutes.

Initial track concept:

Final concept works

2010-01-19T12:33:00.000Z

As pre-production on this project has ended, I have the final concepts ready to be converted into the final environment and the character models. Click on the images for larger files.

Concept work for the menu screen:

Mock-up of final product:

Concept development

2010-01-17T12:30:00.000Z

Before proper production on Seventh Eve can begin, I've needed to go through all of my concepts so far and finalise them. Taking into account appropriateness, semiotic design and design in terms of how easy it will be to rig and animate, this is the development that I've done thus far.

Click on the images for larger files!

Zhinu conceptual development:

The next, and most important, design decision has been Zhinu’s colour scheme. As a Chinese character the colour of her hair and skin is relatively unchangeable, and that leaves the main impact to be her clothing- something that, as she will be relatively small on the game screen, is important. Most of the colour pallettes that I drew up for her were based on existing Chinese outfits, but others had more of a solid concept in regards to the story of Seventh Eve.

The first colour scheme I took a serious look at was a pallette based on the colours of jade. Jade, as a gemstone, has had important ceremonial and artistic roles in Chinese culture, and in the tale of the Cowherd and the Weaver Girl, Zhinu was said to be either the daughter or the granddaughter of the Jade Empress.

The other colour scheme was one based heavily on the use of red, as it is not only the colour of the Heavens in Chinese culture but also symbolises good fortune, luck and is used heavily during festivals.

By mocking up another rough environment not only helped me get an idea of what the final game would look like, but also helped see the final colour scheme against an environment to make sure that it would both stand out against and compliment the game.

Another problem when it comes to the rigging stage of the model will be the long ponytail, so by looking at traditional Chinese hairstyles- most of them pinned up- I've managed to change the design to get around that problem.

Huli jing conceptual development:

As Seventh Eve would be set in the Chinese Heavens, I wanted to base any characters past the main playable character on Chinese mythological creatures. Luckily, Chinese mythology is full of creatures that served all different purposes in both religious and earthly positions, and there were plenty of resources and paintings to research to find the right character to fit the game.

At first, I selected a number of creatures that could fit the role of a useful assistant character (see above for a selection of preliminary concepts). My main idea was the Qilin, or the Chinese unicorn, which, in legends, was said to have perfect good will, gentleness and benevolence to all living creatures and therefore would be a useful guide or helper character.

I also considered the Four Aspects, which were the four mythological creatures that resided in the stars and each held sway over one of the four seasons and the four elements. The two concepted were Zhu Que and Qing Long, the Azure Dragon and the Vermillion Bird, as I thought that these would read best with the chosen demographic.

Also concepted was the Peng, or the Chinese Roc, a giant bird, which symbolises ambition and great accomplishments but had no alignment to either good nor evil, which would have given me free reign with its use in the game mechanic.

However, ultimately, I already had another bird character within the game, so I wanted an earthbound character that would give a little more variation. Ultimately, I found the huli jing; fox spirits which could be either good or bad, tricksters or loyal companions. This I found fit the game perfectly- it would be a character which could provide aid and hints to the character, but do it in such a way that required solving puzzles or riddles first to entertain its whims before it would assist the player.

Above: an illustration of a huli jing from the Shan Hai Jing, a classic Chinese text from approx. 200BC.

The original concept for the huli jing. At this point I had no solid visual style for the game, and was considering a generally more 'childlike' style, similar to the Zelda: Windwaker game, as the game at this point was still to have been a fully explorable 3D world.

However, as I concepted the game further and began thinking of it in terms of a 2D world with more solid ideas of gameplay, and began to research the huli jing further as a creature, I began to realise that a more curved shape wouldn't do the character justice and re-concepted the character, making it more slender, more ambiguous in alignment and gender.

Thumbnails quickly sketched up to get a more solid grasp of the shape and silouhette of the huli jing.

This was the concept that was done after I finalised on a style for the game. Wanting to give the huli jing a personality- rather than creating a creature that was a carbon copy of old paintings of the jing, I looked into giving it a mask- adding to the ambiguity of tha character.

The addition of a Mask led to thinking about the Chinese opera, where characters have heavily painted faces- the designs and colours representing the character's personality and their current mood. When this art form was taken across to Japan, to become Noh, or the Japanese opera, they used masks to represent these, and by looking at different colours and styles and their meanings, I came up with a suitable, if simplified, design for the huli jing's mask.

Above: markings and designs on for the Chinese opera and various Noh masks.

Before I could produce a turnaround I needed to concept the character in various poses to get a proper grasp of its body shape, as its body is composed of so many sharp edges and so abstractly proportioned that it took me some time to get a proper idea of how it really worked.

The final turnarounds for the huli jing.

Magpie concepts

In comparison to the other characters, the magpie didn't require much concepting, as it is based on a real life animal- however, I needed to work out how to create it in the style of the game, and in a way that would make it easy to model and animate in low poly.

References used for the magpie- the formosan blue magpie, red-billed and yellow-billed blue magpie were used for the body and tail shape, and the Yellow-billed magpie used for facial markings. All of these species are common in China.

Original concepts for the magpie.

Final turnaround for modelling the magpie.

Other items and props:

Unused concepts

2010-01-14T14:30:00.000Z

Throughout the pre-production period, there were a number of concepts that were started but then cut from the final game due to the strict schedule that I would be following.

The largest concept that was removed from the final project was the introduction of a 3D world- as the legend of Seventh Eve is based around two characters, one mortal and the other Heavenly, the original concept of the game was that it would take place in both Heaven and on earth- and that the style and colours would change depending on the levels. Earth would be depicted in 3D, and Heaven the 2D-style that ended up as the final style. However, I did begin a number of environmental concepts based on earth:

There were also a number of environmental concepts that I began with the intention of developing further, but in terms of composition, effect and gameplay potential were scrapped before they were finished.

Also never taken to completion were a number of object and character concepts that again, would have either been too difficult to impliment into the game due to engine restrictions, or served no purpose to gameplay, or were specific to an environment that wasn't taken past the original concept.

Pre-production document

2010-01-14T13:48:00.000Z

Before beginning production on Seventh Eve, a pre-production document is required to finalise everything about the project, from the concepts, to the demographic and visual style and level layout. This document has been finished today, and is available for download here:

Pre-Production document via RapidShare