Mike's Blog

Shit!

2010-05-25T03:41:00.001-07:00

Unis finished now...which means I need to find a job.......

Artefact 5

2010-05-09T02:17:00.000-07:00

The aim of this artefact was to see if an improvement could be made on the images in the previous artefact by using Mental Ray as the renderer. After a bit of research (Hakan 'Zap' Andersson's giant robot video tutorials), I found out that Mental Ray has shaders that specifically deal with CG integration into real scenes...what are the advantages (if any) of using these?...

The four shaders that were used for this artefact:

'Environment/Background switcher' - This lets you use one map as a background and another as an environment map, to provide environmental reflections.

'Environment probe/Chrome ball' - This automatically unwraps a photo of a chrome ball for reflections.

'Environemnt/Background camera map' - For a point seen in a reflection, it takes the reflected point’s 3D coordinate, convert it to its matching onscreen position (if any), and looks up the map based on this new 2D location (or returns a special value if the point is offscreen).

'Matte/Shadow/Reflection' - This is part of the Production Shaders library, and is used to create “matte objects”; that is, objects that represent real-world objects in a photograph used as the scene background (also known as the plate). The material provides a wealth of options for marrying a photographic background with the 3D scene, including support for bump mapping, ambient occlusion, and indirect illumination.

I used these shaders to produce the following images, then showed them to 10 students so they could compare them to the others in artefact 4. The question was 'Which look the most realistic?'..

This method uses 1 picture of a chrome ball to provide ambient light info on the models, then lights can be placed manually if needed (which I did in all of them)..

Landing

9 out of 10 people thought the Landing image was an improvement on the others, because of the overall colour tone and shadow.

Bedroom

8 out of 10 people thought the Bedroom image was an improvement as the statue has more illumination.

Light Garden

5 out of 10 people thought the Light Garden image looked better as it was slightly bolder than the others. The other 5 thought it was too bold and under exposed compared to the background.

Night Garden

7 out of 10 people thought the Night Garden image was better lit than the others, and commented that the shadow added to the realism.

What do these results suggest then? Well, it backs up my results from artefact 4 that scenes can be lit just as realisitcally using lights (in this case 1 mrOmni in each), rather than with HDR images. I'm definately not taking anything away from using HDR images to light scenes though, as the benefits are obvious (and discussed in last artefact)..

Has using Mental Ray improved the images? I would say defiantely, and here are the reasons..

Shadows - The shadows softness and colour can be altered much easier as Mental Ray lights have loads more options, aswel as the options in the 'matte/shadow/reflection' shader. This is why the shadow in the landing scene is improved.

Indirect illumination - Final gather and Global illumination help to illuminate the models much more naturally. The Mental Ray shaders also spill the pixel colours of the environment onto the models.

Render times - The renders with the scanline took about 6 mins for high quality...these took around 2.

Lighting the scene with a direct light - The matte/shadow/reflection shader can also recieve light, which means its possible to light real objects with a CG light. This is definately not possible with the scanline renderer! Heres a test..

Photo of table with real cup and deodrant

Heres the same photo but with a mrSpotlight added and a box..

Pretty clever! The best feature out of all of them though is the reflections. If geometry is drawn around real objects in the scene, they can be reflected. Here's an example of this..

As with the previous table image, geometry was drawn around objects that I wanted to be reflected (in this case the desk and deodrant). You can see that these are reflected in the teapot.

In conclusion, using Mental Ray is definately advantageous for these types of shots because the shaders tie everything together and blend objects in more realisitcally. It can do reflections and direct light illumination which would be impossible with the scanline, and it renders alot faster.

Here's the final image for this artefact..

Artefact 6 will be looking at how important shadows and reflections are in making objects fit into a scene..

First showreel!

2010-05-05T03:27:00.000-07:00

Finished my first showreel the other day and its about as good as I could make it with the work I had to put on it I think..

Here it is...

3D Showreel from michael green on Vimeo.

Client project on vimeo..

2010-04-20T09:50:00.000-07:00

Here is the video for the client project, it works on Vimeo but not on Youtube....who knows why...

Olympic museum advert from michael green on Vimeo.

Artefact 4

2010-04-15T07:22:00.000-07:00

The aim of this artefact was to look into how HDR images are used to light scenes in 3dsmax, and compare scenes that are lit this way with scenes that are lit with regular lights.

I found about this method of lighting when writing the essay part of the assignment, and it's been used in the 3D and visual effects industry for the last 5 years or so. It's only recently that this method of lighting can now be done in the majority of 3D apps. I didn't have a clue what a HDR image was before this artefact so i had to find out more about it before i could use them in 3dsmax..

To sum it up - standard images are restriced to 8 bits of information within each colour channel. This means the Red, Green and Blue values that make up each pixel of an image are between 0 and 255. This means that the same pixel would be used to represent a white wall and, say for example, a white sky. Obviously, the white sky in real life would be much brighter than a white wall but an image would not be able to show this.

HDR images on the other hand are stored in a non-clamped format, and are capable of storing a million different different colour values between the lightest and darkest points of an image. Therefore, once the image is loaded into 3dsmax, light sources in scenes can be identified as they have much higher pixel values than regular colours and so they can illuminate scenes using this info. I assigned them to skylights in max and i got pretty good results...

This artefact has taken ages, probably about 2-3 weeks as it was all trial and error because i'd never done it before. I've taken photos throughout so i'll briefly talk about the equipment i used and how i used them, then look at the results...

In order to light a scene with a HDR image, i first had to actually make one. The method of making an HDR image for CG lighting is to photograph a chrome ball at different exposures, then feed these into a HDR program like HDR shop so that they can layered together to make a HDR image. The reason to photograph a crome ball is so that all the light from the room can be captured...

I already had a digital camera, so i brought a chrome ball (a garden ornament, which i was lucky to find)from a local garden centre. My first attempt at photographing the ball -

I realised this wasn't going to work as the relfection of the black table that was holding it up was pretty big. This would probably mess up the lighting info on it, and would definately mess the image up if was going to be used for reflections. So i had to make a stand for the ball that wouldn't be to big and was higher up..

Perfect! It looked pretty ridiculous but it was good enough for what i needed it for. I also made a stand for the camera as i didn't have a tripod..so here are the photos of the ball, the resulting hdr and the first render of a model lit with a HDR i made..

The 3 images above - 1 under exposed, 1 normal, 1 over exposed. These were then taken into HDR shop and layered together, cropped and transformed so that the image of the sphere was flat like the image below...

This is what the HDR image looks like. It's got all the lighting info stored in it. I used this image for the spherical environment in max and assigned it to a skylight. Before this, i had taken a background shot and put this into max with the statue model in it too..this is the result i got with no other lights..

Not a bad result, it had definately worked as the lighting and colour tones can be seen. Before I began my test of HDR lighting vs standard lighting, i wanted to do a few more test renders..the next one was on my landing in my house, but i noticed that something was wrong (this is where it got annoyingly complicated)..I used the exact same method apart from this..

Matching the real sphere with a CG 1 helped me match up the brightneses which was good..but it gave me this result..

I liked the shadows it casted, but you can't actually see the light variation on the model..in real life the soldiers left side would be brighter as the door to it's left is open with light coming through..(the door on the soldier's right wasn't actually open when i shot the chrome ball which slightly messed it up anyway)..so something wasn't working properly and i had to find out what it was..

After reading more about HDR lighting, most practioners advised to shoot the ball at a high variation of exposures so that a higher number of lighting info can be captured..this was where i was going wrong as my cheap digital camera wasn't capable of it..so i had to get an SLR from uni and then try again...

To make sure it worked this time, i shot the sphere at 10 different exposures...i won't show them all but here are the brightest and darkest..each had a difference of 1 Fstop..

And here's the resulting image..

It worked! So it was worth getting an SLR out..now that i knew it was working properly, i could do the HDR vs standard lights test..I made 4 different scenes, with two different versions of each (1 with HDR lighting, 1 with lights)..I then showed the images to a group of students in order to gain feedback..

Here are the images - the first image of each version is the HDR lit one...

Darkroom

Landing

Garden

Night garden

I asked them which version of each scene looked the most realistic and received mixed feedback..(i didn't tell them how each was lit)

The HDR versions of the Darkroom and Landing were definately thought to be more realistic, but both the garden scenes had mixed opinions as some thought that the standard lighting looked better. It's interesting to hear as i originally thought that all the HDR versions looked better, but i can now see why some people prefered the standardly lit scenes.

What does this tell me then? That i've wasted alot of time and money!..haha not really. It does tell me that lighting a scene with your own eye can be as good as HDR lighting, especially if you took them through compositing (like i did in artefact 3). But, when lighting a scene with a HDR, you know that the lighting is going to be 100% accurate (if done properly). There is no messing around with positioning lights and getting them the correct brightness etc...it took me quite a while to set up lights correctly in the scenes in comparison to using an HDR image.

Obviously it's quite abit of hassle setting up the ball and photographing it, but i'd say the results from this artefact show that it's worth doing. Also, the shadows seemed to be much nicer using HDR images in the skylight, compared to using an omni with the scanline renderer..i even had to composite the render of the standard lighting Landing image into the HDR version so that the shadows could be used as they just looked awful and fake -

What's next then? In artefact 5 i'll either be looking at comparing renderers for CG in live action shots, focusing on shadows and reflections or looking into creating mattes and masks (which i've never done)..so i'm going to ask David which one he thinks would be best..

If I had to pick the best image from this artefact i'd say it was this one -

People commented that the shadows look excellent, and the tones of lighting on the model look accurate (this is the version lit with HDR). This is also a reason why I think it might be interesting to look into the importance of shadows and reflections, and the best ways to produce them in the next artefact...we shall see..

Video not working on youtube..

2010-03-25T03:08:00.000-07:00

Sorry about the video below, not sure why it's not working on youtube. It could be because it's an mp4 video and not an avi..i have a think about how to upload it..

Too busy to update, but finished client project..

2010-03-24T13:52:00.000-07:00

At last i've finished the client project, and i'm happy with it. It's the Olympic museum, and here it is...

I dread to think how many hours i've spent on it, but i'm glad it's done. Theres a few things that i'd possibly change, but nothing thats going to make me open up 3dsmax max and after effects again this week...or maybe even next week! I might even retire now

Compositing...is good!

2010-03-11T01:48:00.000-08:00

Over the last few weeks (from PRP and Client project) i've come to realise that compositing is something that i'm now going to focus on mastering as i've seen what it can do. Compositing in after effects makes it easy to add stuff to final renderings that would take ages to do without..and it makes it look much better.

I've now sorted out how the sky is going to look and be added, how the cloud shadows will effect the ground, how the ground will change brightness like in the time lapse videos, fire for the torches, dust for the concrete.

The dust and fire are pre matted royalty free pieces of footage which are going to look good i think when composited in properly...i'm in the process of downloading some cloud time lapses and i will post again when i've got them.

After speaking with Simon on Tuesday, he said interior fly throughs will not be essential as long as the exterior animation is good..so this relieves abit of presure as i don't have to rush making these if i don't have time. It's 50/50 at the moment i think..

Also, i'm no longer thinking about rendering the advert in HD as there is just no point. Jools said footage on our show reel has to be 720x576 (PAL) so thats what i'm going to stick to...and to be honest after doing a couple of tests, there wasn't a huge difference in quality anyway.

720x576 = less rendering time, less hassle, still good quality

1440x1080 = more rendering time, more hassle, no point, shit, crap

What's next to do?

1.find royalty free footage of firework display
2.re-animate the entire advert - make it more intersting and snappy
3.light the scene for night shot - and prepare footage for this
4.i'm sure there will be more things to do..

www.mikegreen3d.co.uk

2010-03-07T11:21:00.000-08:00

I'm pleased with my portfolio website..though i just need some good work on there now! Hopefully that will come..

Research Artefacts put on hold?..

2010-03-07T11:10:00.000-08:00

The deadline for the research project artefacts is mid May..so surely i can put these on hold until i've finished my client project which as to be handed in 26th March..? Plus, i've got assignment 2 of spp to do for 20th March.

I've completed three artefacts so far and would much prefer to just get the client project done first before starting another, instead of keeping swapping from one to the other. When doing this before, i find myself losing track of where i was at and lose abit of motivation so there might not be an artefact update until late this month now..

I'd still have 6 weeks to finish the artefacts..which is easily enough time as there's only three left to do. I'll have to talk to David Downes about it as it surely makes sense..we shall see..

Cloud time-lapse and how ground changes brightness..

2010-03-07T10:47:00.000-08:00

I've been looking at cloud time-lapse videos and am definately going to include this in the animation of the museum..

I know the video is VERY cheesey with that music though..pretty funny. But, i think its going to be important to animate changes in the brightness of the ground aswel as this is what happens..

So far, i've done a rough animation of the museum and the order in which it's going to be constructed together and also looked at using particle effects for parts of it..the video is big though and will be uploaded to youtube when i've got time..

What do i need to do now?

1.Look at how the scene is going to change to night time
2.Attempt to light the museum so it looks good
3.Model 2 of the interior rooms for the flythrough - i'm going to try to keep it fairly simplisitc due to time, but make sure the lighting makes up for it to look good
4.Look at the best way to composite cloud time lapse footage to replace the sky
5.Look for footage of a firework display to composite in for the end shot
6.Is the final render going to be HDTV or standard PAL?
7.Go over all materials and make sure they look good
8.Probably use raytrace for the reflections of the windows..what is best way though?
9.Cry because i've got looads of work to do!

So i've got quite abit of work to, will i get it done in time? Only time will tell.............

SPP assignment 1 - done..and my new portfolio website

2010-03-03T04:27:00.000-08:00

Since last Wednesday the assignment 1 for SPP is all ive been doing and was finally handed in yesterday..and i'm happy with my job and course choices.

The job was a 'CAD 3D modeller' at a company called Archiform 3D..in America! I explained in my rationale though that although this is the type of job position i'd like, i'd need to be working in England. The way they show their work on their website is impressive..here's the website..

http://www.archiform3d.com/employment/

The course was '3D Computer Generated Imagery' at Kingston Uni. I thought this was a pretty safe choice as it'd be carrying on with what im doing at the moment..I also listed some visual effects courses that have 3D modules that looked interesting.

Oh yeah, and ive just set up a portfolio website with the help of my old course mate Kam which is nice. Although, i don't really think my previous work is that great so i need to get a move on and update it with some better work (my client project hopefully)..but it's a start in the right direction i think..

My new portfolio website www.mikegreen3d.co.uk

Artefact 3

2010-02-24T07:25:00.000-08:00

The purpose of this artefact was to see if an improvement in the render of the statue in a garden scene could be made by rendering seperate elements of the image, then compositing them together in Combustion and altering them to acheive higher realism. I was keen to learn how to use Combustion as it meant that i could tweek individual elements of the render such as diffuse, lighting and shadows, and try to improve the overall realism of the scene. So to start off with, i created 1 render in 3DSMax with basic lighting, and then made another 2 by rendering seperate elements of this to Combustion and tweeking each pass. I then showed these three images to 10 people and asked them which was most realitic and why..

This is the first render that i started off with..

(No compositing)

And from this i made the next two images by rendering the elements seperately and altering them..

Image1

(diffuse brightness increased, shadow opacity increased, lighting slightly increased)

Image2
(diffuse slightly increased, lighting slightly increased)

The lighting in each of these is slightly different, as are the shadows. 7 out of 10 people suggested that Image 1 had the best overall realism, as the lighting was clearer and lighter which better matched the surroundings. 3 out of 10 said Image 2 was more real as the lighting was more subtle.

Already the research showed that an improvement could be made through compositing however, there were also criticisms of the composited images. People suggested that none of the images were realistic as there were no shadows on the ground, or on the statue itself. Also, people suggested that the statues appear unrealisitc as there wasn't enough light variation on them which the sun's location/direction would create.

The difference with an extra light (acting as the sun) can be seen in the next two image..

Without extra light

With extra light

(No compositing, just lights in 3DSMax)

The image with extra light added was shown again to the people in comparison to the others, and the majority of them suggested the light had improved as it gives the light a variation of intensity (lighter on left where sun is). This also created a slight shadow on the floor. 7 people suggested that the statue was still too dark and that the shadow wasn't strong enough. I rendered seperate passes of this for compositing and here they are..

Diffuse

Direct lighting

Shadows

So in response to the feedback i had to make the statues main lighting brighter, and make the shadow on the ground more visible. The next three images are the result of tweeking the brightness and opacity of the diffuse, lighting and shadows to get the effects i needed..

.....AND HERE'S THE FINISHED IMAGE FOR THIS ARTEFACT.....
WAHOOOOOOOOO

To be honest im loving the power of compositing and think it's easily the best way to acheive better renders. The best thing about it is that it's quick and you dont have to sit around all day waiting for something to render as it's instant, and this artefect has shown me that big improvements can be made on renders that initially look good in 3DSMax. Artefect 4 is either going to be looking further into compositing or looking at HDR lighting. Im going to decide very soon...

Particles

2010-02-21T12:21:00.000-08:00

This has took me most of today! Glad i've done it tho as now i know how to do it i can use it in in some way for the animation..it could look pretty good with a bit of tweeking i think. Got to now figure out how im going to fit this type of animation in to the advert..

Building construction time lapses..

2010-02-19T08:21:00.000-08:00

Been watching building time lapses to get an idea of the order in which buildings are put together. Im going to try and follow this order with the museum..from what i can see the order is..

Foundations - Main frame - Roof - Inner frame - Outer panels - Windows - Smaller details.

object width="425" height="344">

Animating the museum..

2010-02-19T07:26:00.000-08:00

I've decided to get started on the animation so that i can get an idea of the best way to go about it..after a few little tests of putting the building in the environment and adjusting the perspectives, i think the best way to animate it would be to slice the building into the parts that i want to animate seperately then animate them in the same 3dsmax file. I could then alter the perspectives and sizes after when i want to composite them in to the different shots. Doing the animation now is a good idea so that i can get an idea of how it's going to look aswel..

So far, im thinking about using the pro-boolean reveal method (which i looked at in a post about a week ago) to animate the building and im also going to look at particle flow animation. The backgrounds that i've used in the renders below are pretty low quality so i need to re-take them once this snow clears up.. also, the building looks on the small side so think i need to scale it abit.

Musem frame..

2010-02-18T12:10:00.000-08:00

Getting there with the frame it's starting to look better..got to now decide whether to work on the texturing and lighting or make a start on the animation.

Might make a start on the animation soon as i could be wasting time on the lighting as im still unsure whether to light it up at night..

Holland metro station..

2010-02-17T10:01:00.000-08:00

Found this, would love it i could light the museum in a similar way..

Frame and outer building..

2010-02-17T08:06:00.000-08:00

The frame and outer bit of the building is done with just the smaller details to add..also got to do an entrance and more importantly..the interior. This is going to take most of my time for the next coming days i think..going to try and keep it fairly minimal and simple as ive still got texturing, lighting, animation, filming, compositing to do! This is going to take forever..

Museum frame..

2010-02-17T06:34:00.001-08:00

There must be an easier way to model the frame than the one i did as it took ages and was VERY repetitive..anyway here's the first render of it..

Chongqing library in China..

2010-02-15T04:26:00.000-08:00

Stumbled upon this library building design in china, and it's really similar to my museum..this gives me a better idea of how the frame of the building could look which is helpful even though this ones alot bigger! It looks like the main frame of the building is offset from the frame of the windows then its got supports that connect the two..this is what im going to do now..

Lighting the museum at night..

2010-02-14T12:25:00.000-08:00

As Simon suggested that a night shot of the museum would look good, i've been looking at pictures of buildings at night to try and get an idea for lighting..

The thing with this is that as 90% of the museum is made out of glass, it's obviously going to look bright at night and the lights are going to play a major part in how it looks. This means that the only way it's going to look decent is if the lights are postioned in the right places, the right strengths, are good colours etc...

This means that this could turn out to be a pretty massive job..and i dont even know if it would be possible to do with it being in a real scene but im sure there're ways which i need to look into.

If the scene was shot at night surely it wouldn't work as there wouldn't be enough real light to see anything, so if it was shot in late afternoon/evening this could may be solve this..

OR

The scene could be shot in daylight and the colours modified in after effects to make it appear evening..? The sky could be a seperate composite and could also be changed colour in after effects..and this could be animated throughout so it appears to be gradually become evening throughout the advert so at the end it's lit up..this would be brilliant, but it'd clearly make things more complicated but probably worth it.

Think i need to discuss this with Simon and see what he thinks..here's some of the images i found..

Ideas for animation

2010-02-14T11:33:00.000-08:00

This is awesome..

The animation for the museum could be a mix of animation like in the video above and time-lapse style animation like the video below..or is that just complicating things? Would that even look good? May be it should be one or the other..

Animation technique for museum..

2010-02-14T11:11:00.000-08:00

This is a pro boolean reveal animation technique i done as a test...this could look pretty good i think if the timing was set right so different parts were animated at different times..hmm

Going to look at others aswel, here it is..

Museum floor plans..

2010-02-12T10:02:00.001-08:00

Client project storyboard

2010-02-12T09:27:00.000-08:00

Thought i'd post these renders which were used in my client project proposal for the storyboard...it shows the museum slowly being built up, then some flythroughs of the building. After a busy period with the artefacts im going to start working on this again...probably finishing general modelling and materials, may be interior..ive also got to start figuring out what sort of animation is going to be used...anyway here they are..

Artifact 2

2010-02-11T04:59:00.000-08:00

Artefact 1 showed that a near perfect camera match could be created with a fairly slow moving medium shot of statues. In this Artefact, I put Matchmoving software Boujou4 to the test to see if the software could successfully create a camera match when using footage with different camera shot types. This Artefact has taught me that in order to create a successful camera match with Boujou (or possibly any other match moving software), the video footage used has to be of a high quality. I found that the lower the quality of the shot (e.g. low light, blurred), the harder it is for Boujou to calculate a camera match.

With low light shots and fast panning/shaky shots filmed with the mini dv, Boujou often crashed or stalled when tracking the features of the scene. The amount of tracking markers Boujou created on these types of shots was much lower than that of a slow pan/good light shot, which means that the program is struggling. The quality of the evening and night shots that were filmed was reduced due to the lack of light, and didn't work in Boujou.

The screenshots below show a good and bad example of Boujou's attempt at tracking the scene features with a mini DV camera..

The 1st shot is a fast panning shot and the 2nd is a slow extreme long shot pan. Boujou was able to track the features on the second shot more effectively

Three of the artefact videos are attached in this post. Feedback for two of the videos have been positive and one is quite obviously a failed attempt at matchmoving. In order to create a matchmove for the garden fast pan shot, a high shutter speed of 215 was used on a Sony DCR VX9000. This was so that there was less blurring on the footage, then Boujou could calculate a camera match. The 2 images below show how increasing the shutter speed increases quality of a fast panning shot (as used in the garden matchmove)..

Heres the 3 videos for the artefact....My next artefact will be focusing on compositing in Combustion, looking at different lighting solutions..

Artifact 2

2010-02-02T11:37:00.000-08:00

Artifact 2 will be taking a deeper look at matchmoving and its capabilities. Ill be taking different types of footage...so for example..A fast panning shot, an long shot, and extreme long shot, a night shot...

I think the last camera match was near perfect, so it will be interesting to test it out further and get people's opinions on it..

What next for artifacts?

2010-02-02T11:18:00.000-08:00

Having tested out matchmoving software Boujou4 in the last artefact and the final video looking quite good, i want to test it out further by trying to track different pieces of footage and seeing how accurate i can match the camera. Ill then be getting feedback from friends, tutors and coursemates and seeing what they think of the footage..

The statue in the last artefact looked suprisingly well matched into the footage, as the camera was matched perfectly, the stone material applied looked ok and the lighting was quite a good match. So, how can i take this further?

Feedback suggested that the main giveaway that the statue was a CG integrated model was the lack of shadows at the base onto the ground. Also, the model looked sharper than everything else. Therefore, compositing looks like the next step to improve the realism and acheive the desired results. I could add a grain type affect to the model, and use an ambient occlusion pass to add a shadow at the base. Im sure i could improve it in other ways too...

Compositing would allow me to alter parts of the video, and in order to do this i'd have to render seperate passes of the model and use Combustion to put them back together. I used combustion the other day for the first time and got a basic understanding of how it works and it looks ideal for experimentation in the artifacts..

The final video for Simulated client project..

2010-02-02T11:10:00.000-08:00

I forgot to embed the final video into the blog back in november! Anyway here it is...

Artifact 1...continued

2010-02-02T11:06:00.000-08:00

The purpose of this artefact was to test out match moving software Boujou 4, in order to see if I could successfully match the movement of the real camera with that of the virtual camera, so that a 3D model could be integrated into the footage. I decided to film footage of statues and ornaments at around 10 seconds long so that I could integrate a 3D statue amongst the real ones. As I was new to Boujou 4, this artefact gave me the opportunity to explore the work flow needed in order to make this work along side 3DSMax.

I integrated a 3D statue into the 3 different pieces of footage, and reviewed each one afterwards by myself. In the first one, the 3D statue looks OK next to the real ones however, it stands out as the materials and lighting do not match the surroundings, and there is slight movement where the camera has not matched properly. The second one is better as the statue looks completely static within the scene, but the materials and lighting still look false. The third one was better as it was static within the scene and materials and lighting was improved.

I decided to test the third one out on other people by asking the question “Is there anything unusual in this video?”, and I received mixed opinions. 80% of people who watched the video first time around did not notice the 3D statue and even after several views still did not notice it, however, 20% did notice the 3D statue. It was apparent that the lighting on the model changed slightly halfway through and therefore made it noticeable to them. The next artefact will be looking at the same scene, but I’ll be aiming to improve it through compositing.

Artifact 1

2010-01-28T14:14:00.000-08:00

The purpose of the artifact was to test out the matchmoving software Boujou 4, in order to see if i could successfully match the movement of the real camera with that of the virtual camera, so that the 3D statue (grey stone woman) could be integrated into the scene...

better render...

2010-01-23T12:35:00.001-08:00

Heres a better render of the latest museum model....well happy with it so far..

I forgot to add in the last post, found some excellent animation techniques by Louis Marcoux on The Area website in order to make the building appear. Think ill be using these, the techniques this guys uses are pretty much exactly what i was looking for...

Client project development..

2010-01-23T12:07:00.000-08:00

My idea for the building has developed quite abit and here are some of the renders i made and it shows how the idea has developed. After a discussion with Simon, we decided that the product will be an advert for the general public that is making them aware that an olympic museum is being built in a specific area (dont know where yet)..

Since the last post, i have worked on the look of the building and also decided exactly where im going to shoot the advert. Im going to use KLV's running track as the place to animate the museum as it has flat ground and a good view. At first i didnt like this location as i thought it may look strange with the running track but then i realised it could be fairly easy to edit to make it look better.

Im pretty happy with the look of the building now, it definately looks better than the first model i made....which was so crap! They are above this post....

a rough first idea..

2010-01-11T11:37:00.000-08:00

Ive been working on the design of the building today. When designing, i find it easier to draw out rough designs in 3dsmax rather than draw them on paper as ive got the freedom to quickly change things and see the results. I used reference photos of modern architecture so i could generate my own ideas easier and have based the building on the shape of the olympic logo as mentioned in the previous post...

This building is going to be an olympic museum with two floors in each of the parts of the building. When you start to design the building, you begin to realise important things such as.....How big does the building need to be? What will be inside the building? What materials will be used on the exteriors and interiors? Will the building be harder to integrate into a real scene because of the large amounts of glass?...and the list goes on...

Im talking to Simon tommorow so to show him the initial idea and see what he sais....

Live client expanded brief...

2010-01-10T09:57:00.000-08:00

Here is Simons original brief that he sent me..

Within an urban (town/city) environment composite a futuristic building encapsulating the spirit of the 2012 Olympic games. The shot should show a time-lapse transition from the site as is to the new building. Creative flare can be taken in how the building animates together as well as it's overall design again appearing as a 21st centaury dynamic piece of architecture.

Here is my expanded version of the brief..

This production will be a one to two minute advert for the general public, which aims to make them aware of a new Olympic Museum that is going to be built in Kettering, ahead of the London 2012 Olympics. The advert will show the exterior of the museum being constructed so that it is interesting to the audience, and will also show a fly through inside the museum.

Client Project..

2010-01-08T07:49:00.000-08:00

Simon Wallet has sent an expansion of the CGI integration and compositing brief. I have to design a building that encapsulates the spirit of the 2012 olympics and integrate in into live video footage. The building then has to be animated in with either construction or destruction...

Im going to animated it being constructed as i think this will be more appropriate. At the minute i feel like im at the start of a hike up a mountain to be honest but im excited about this project and think its got huge potential.

The research project taught me the areas of importance when integrating CGI into video footage in order to make it look realistic. Lighting plays a huge part and ill probably use mental ray as the renderer, materials and the way they interact with the environment are important, and obviously the scale and accuracy of the model will be important.

I am yet to decide on the exact methods i will be using in order to composite the model into the footage and ill look into that later. The question is now.....How am i going to design the building? What is the purpose of the building? What features will it have?...

Simon has asked the building to be of a modern architectural style....so this basically gives me loads of freedom when designing the building, which is cool. Here is a quick design i made for a building which is based on the logo of the 2012 olympics...No materials have been added its just a rough design for the shapes..and id be pretty shocked if i seen that in the field outside of B&Q.

Simulated client project...

2010-01-08T06:58:00.000-08:00

Its been a almost 2 months since ive updated the blog and in that time i have completed both the simulated client project and the first part of the research document and am pleased to say both have gone well...

I changed the title of my research project from 'How realism is created in architectural visualisations' to 'How is CGI used in film to create visual effects?' and am very happy that i did as i found it interesting. The decision to change this title meant that i would also be focusing on this during the client project and Simon Wallet has given me a 'CGI integration and compositing' brief to expand and work on.

From now on ill be updating this blog pretty regularly as it will help me look back on the progress i have made with the next two projects...

Here is a link to my Anti Piracy Advert for the visualisation brief...

www.youtube.com/watch?v=TTWg8RMHRjM

Simulated project Personal research project

2009-10-15T10:03:00.000-07:00

For my simulated project, i have to visualize peer to peer filesharing over the Internet. It has to be a 20+ second animation/visualization that portrays to the audience the illegal side of this. Illegal filesharing over the Internet was brought even further to everyones attention when website owners of the 'Pirate Bay' were taken to court.

I want to create something that is visually simple to follow, but gets the message across effectively. Ill use the rest of this week to draft out ideas ready to get feedback from Simon Wallett (if he is my tutor, think theres been a cockup somewhere as my timetable suggests otherwise!). hmmm..

For my Research project, so far im thinking about exploring how to bring accuracy and realism to Architectual Visualization projects. Think this would be quite intersting to look into, and it'd widen my knowledge of CAD drawings and MAX alot..no dought this will change but this is what im thinking so far..

1st Blog..

2009-10-15T09:49:00.000-07:00

This is my first blog on this page, and of the year. Im going to have a celebration...