Wednesday, September 23, 2009
Wednesday, August 12, 2009
It is good to see an update, I hope you're not too exhausted and have enough energy for the final push. You can relax, after the hand in ;)
It would be a shame to lose the outside by blowing it out so much that you can't see it, and it would take away the point of having a moving camera to show off the panoramic view. Equally it makes no sense to have full sunlight falling into room and for it to still be dark. What you are trying to achieve makes far more sense when you see your photo reference from your previous post, but as I said in my reply, this lighting is physically too different to make it work with your current scene. This is definitely something to reflect upon for next time.
Here is the reference again from the first image. They both show interior and exterior and so are a good reference for your project, especially how they get it to work without the outside blowing out.
It's warm, light and bright, the lighting might not be realistic, but it is believable, and more importantly it sells the house.
We have had a few discussions on volumetric lighting, in Jamie and Lies recent posts, which you should read because it is relevant.
A depth pass can be used to add atmospheric lighting, but used on an interior, it can make a place look dusty and dingy and is often used in the thriller genre (see the Citizen Kane reference from Jamie's RE:) , and the same with volumetric lighting so I would re-think its use for a show home project. Outside you would normally only see the effects of a depth pass over large distances or as ariel perspective (colour change over large distances http://en.wikipedia.org/wiki/Aerial_perspective) or to show rain or fog.
The yellow grade you have added is warming, but it feels a little on the green side (it could be my monitor). I think that it would be better on the orange/red side and a bit more saturated. I have done a quick paint over and grade of your image, it is by no means perfect, but I want to show you a less dusty warmer colouring. I have also limited the glow/burn out to the top left side of the window, to give it a bit more direction, and contrasted up the image to get the hot highlights on the floor. When you are working in comp, be careful to not loose the colour in your feature items. The light glowing off a red object such as the sofa is what adds the pizazz you need.
As a general rule, its better to try and fix rendering problems at the source rather than trying to cover them up. If one part of the model is blotchy there is a good chance that it will occur somewhere else. Did you bake out the lighting in the end? I believe the blotchiness is because your settings are not quite right yet, I have seen it before. Look on the forums for help, I think these might be useful.
There is no problem using your comping fix, but just be aware that this is not the ideal solution and if you had the time it would be better to change the settings.
I hope this helps, good luck finishing off.
Tuesday, August 11, 2009
I much prefer your new renders, I like the fact you have used reds and greens and yet avoided looking christmassy. I think I understand the nostalgic look your after and hopefully I can give you some tips, to help soften it.
Normally I wouldn't recommend changing your reference 3 times :) but I think that you have come up with something much more unique for your showreel. I hope that you are happier with the end result. It is much more endearing than what you started with.
The direction of your key light is not immediately obvious. You will not see a shadow that faint and that sharp in the real world. Faint shadows tend to be soft occlusion type shadows rather than the key light shadows. I like to think of the fill light as the base lighting, you add the key light on top to add depth to the image. If you do have shadows but can't see them, it's because your fill light is too overpowering, and your key light is not intense enough. This is a common mistake by lighting beginners, possibly because students automatically think that lighting is about being able to see everything and therefore make all the lights the same intensities. The result is flat looking CG images. If your not sure what a light is doing, or where a particular light is coming from, look at it in isolation. Black patches where there is no fill and no key light look ugly, perhaps use an ambient light to bring up the black spots, re-place your existing lights (or re-think the type of light) or add an extra light (perhaps a subtle bounce light).
Observe my quick renders and comp (ignore the shadow direction, this is inherited from someone else's post).
The key light adds the drama, the more contrast there is with the fill light, the more depth your image will have. As you have a flat background you have no other way of getting depth into your image other than with lighting. I have used a greener blue for the fill and a orangey yellow for the key. The blue is a warmer blue, combined with warm key, the overall impression is of warmth and softness. The strong key light also directs the eye to your subject, see the reference again for inspiration.
I took old reference, most of it over 100 years old, because although computer lighting is a relatively new subject, the concepts and traditions we use are very old and well established. There is no point trying to do something new, until you have studied the past. I recently went to the Banksy exhibition at the Bristol Museum near where I live. My favourite piece is a big stone with some words engraved on it. "The bad artists imitate, the great artist steal." below it, Pablo Picasso is crossed out, and Banksy is inscribed :)
The only way you can literally soften an image is by either softening the edges, which would only look right if your objects are moving (motion blur) or depth blur (and you have no depth), or by softening something in the images, ideally the shadows. It only needs a tiny amount of shadow blur to soften the whole image, see the image below where MR shadow blur is set at 0.002. Also make sure you use your occlusion pass to get those realistic soft occlusion contact shadows. This will help bring the whole thing together.
I personally wouldn't use an incidence shader as a rim light, but it does a good job of making the character stand out from the background. I would try to keep it a little more subtle, and perhaps tinted a cooler colour to contrast with your warmer key light.
Don't feel that you need to have a central composition, I think it is more interesting slightly off center. You will need the stronger lighting to direct the eye, and then add a vignette to frame the image. As a final note, re-read the other posts, ones to Lies and Holly, as well as the shadow extract, should be particularly useful to you.
Good Luck finishing off, I hope have time to make some changes before the final deadline.
Thank you for all of the feedback in your last post, it was very helpful.
Since then, I have been working around the clock to get my passes rendered out.
I am now in my final stages of rendering and the main problem that I have encountered is a strange ‘blotchy’ effect on the wall above the windows which seems to be a result of the final gather calculations from the appended file:
I am now creating a new pass that I can ‘comp in’, to hopefully reduce this effect:
I have now started to pull together a few composited scenes in Nuke. I started by trying to make the sunlight appear to be glowing around the window frames:
I’m still trying to show how the scene would look if the interior lights were also switched on. My initial attempt resulted in the room actually appearing to be darker:
I’m much happier with my latest attempt, where the interior begins to look much brighter & whiter than before:
I have also created a Z-depth pass. It is applied to this frame… However, I have not yet tweaked it enough for it to look very convincing:
Here, I have added ‘volume rays’, to simulate sunlight piercing through the scene… I like the effect that it has on the TV screen but not so sure about the other areas. Perhaps it just needs colour adjustment or to be more subtle:
This will probably be my second to last post because I have a very limited amount of time to make any changes. I’m looking forward to your thoughts on what you have seen so far…
Sorry for the late reply.
Great work! I think you have the lighting side pretty much wrapped up, I will only say that you should watch your hot spots on the horses, they are clipping a bit too much. The white horse is reflecting light 'like metal' in places, you loose the detail when it is burning out plus it is a bit distracting from the overall model. I would consider wrapping the key around a tiny bit more to accentuate the lighting on the model.
I really like the idea of using atmospheric lighting for this kind of project. The way you have it set up at the moment looks more like a lens flare than god rays, I actually think that in this instance, a lens flare would be a perfectly good way to achieve the same effect, so maybe do some research into this. The position of the bright hole in the clouds means that the god rays have nothing to contrast against, before it meets the character, see the reference for examples. Also it's not really working because of that particular angle of the light, and the fact that it is probably a bit too subtle. If the background and framing were different, perhaps for different shots, I am sure you could get it to work, to achieve the mood you are after.
When I look at your image, the overall impression is great, its atmospheric, it has drama, but it is being undermined by the strong depth blur on the distant trees. It gives the illusion that the model is small, rather than the true scale of men on horseback, and I think that it is compromising the epic nature of your pose and lighting. That amount of blur would work on a close up, but not a wider shot. If you are going to use depth blur on your images it is best that you have a basic understanding of cameras. I am no expert but I have put together an example, to show you what I mean.
I used a Canon 350D, focal length of 23-mm ( I am using a wide angle lens on purpose) and f-stop 4.5 to take the pictures below . The only variable is the distance of the subject, my WALLE toy, to the camera. The nearer the focusing object is to camera, the shallower the depth of field (DOF is the area in focus). Notice how close the object is, to get the background to blur extensively [that is my arm and I am looking through the viewfinder at the same time]. Try using this calculator to see it in numbers ...
With these settings, the subject only has to be 6.5 meters away for the depth of field to be infinite. Therefore the blur would be over a very big distance, and would be minimal. Try imagining a couple of horses there, how close would I have to get to match your depth blur. You would need a very wide angle lens to fit the horses in the shot and be close enough to reduce the DOF to that extent. With a narrower field of view you would need be substantially further away from the subject and have a super fast expensive lens (the f-stop would be smaller) to match your blur.
Of course depth blur can be used to give an epic feel, but only when the character is supposed to be tiny compared with the background.
I have a pretty good idea why you want to blur the background. To get the painterly feel you are after, to bring the focus onto the model (your lighting and central composition do that already) , to hide things in the background that you don't want people to see. I realise that you have your reasons, but I want to show you how it is going to be seen by professionals and audiences alike. They might think that it is a lack of your understanding guiding you rather than artistic reasons.
My final comment is on the textures on the ground. The grass seems out of proportion with the horses, a little small I think. It's best if the textures match as much as possible to really convince the audience of the scale of your model.
I hope you feel that I have helped you through this mentoring process. You have shown that you respond well to feedback and that you have an artistic flare, very hirable qualities.
Good Luck with the final hand in!
Sunday, August 9, 2009
Friday, August 7, 2009
Some information on shadows, that I started writing for Holly, but think it would be of interest to everyone, not too in-depth but enough to get you started.
How you blur your shadow depends on the type of shadow, and the software. Both ways have their pro's and con's. [here's the science bit]
Shadow map shadows work by first of all producing a depth map texture (the shadow map itself) as if looking through the light then the software interprets by projecting this map back through the light to produce the shadow. Ordinarily this doesn't take into consideration transparent objects, but 'Deep shadows' do (I think this is called detailed shadows in Mental ray), sometimes it is an additional setting on the shadow map. Beware these textures have much bigger file sizes compared with ordinary shadow maps, although you can make the resolution smaller to compensate, so only use them if you need to see shadows with fine transparency detail such as when using fur.
The first pro is that they are quick, as a depth pass is a relatively simple calculation. Normally they are created first (sometimes in a pre-pass), temporarily stored, and then used in the render. You can also save out and re-use your maps. This is useful when you have a static scene with a moving camera, as you only have to do the shadow calculation once (sometimes know as once per job shadow maps). If you have a static set but a moving character, you might want to light the scene with one key light, and two shadow maps, one that is used once per job (and therefore is being reused) for the set, and a much smaller map for the character, which you will have to create for every rendered frame. In most software you can now specify a shadow camera, to use instead of the light, simply for the purpose of producing your shadow maps. This means that you can reduce the size of your shadow maps, and have more efficient renders, by focusing in on the thing that is casting the shadow, rather than the entire stage. Also you can adjust the angle/position of your shadow, without changing the lighting, a useful trick for faking lighting.
I prefer to use shadow maps, as it gives me the control I want with fast results. Some software producers have incorportated extra efficiencies, especially those that do not rely heavily on raytracing. Renderman shadow maps are written out as .tex files which is a type of mipmap. This is very efficient when rendering the frame as it will use the most appropriate texture size for the shot taking into consideration the distance of the objects from camera. It's too much to explain here but look it up if your interested. http://en.wikipedia.org/wiki/Mipmap
The Con's are that they can be tricky to work with, especially if your talking big, complex maps. The software calculations can get things a tiny bit out, and self shadow objects that should be in full light. The bias setting is there to compensate for this small amount of understandable error, so you can put the projection of the shadow map back, or bring it forward to correct the results. Bias tends to be very small numbers, but is also normally dependent on the size of object in your scene so it might take a bit of experimenting to see the results. I have seen people brush off using shadow maps because it darkens their lighting, they don't realise they have self shadowing objects (or perhaps they trying to use one small map for too big an area) and that adjusting the bias will fix it. I believe can only make an informed decision when you understand some of the technical aspects behind lighting.
The accuracy of the shadow map is limited by its resolution, and to blur the map you can use the filter size setting (in maya) or the softness in Mental ray, although the softer you make your shadows the more samples you will need, otherwise you will quickly lose quality. Samples and Resolution sizes are normally set at numbers like... 2,4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096. They are to the power of two. It is way more efficient to use these numbers (I think for memory accessing reasons), so stick to the rule. If your not sure where to start when using these settings, start low, and work your way up until you see the quality/time balance that will work for your project. If your using a small shadow map, say 64 pixels square, on a PAL render of a character, the shadows might give the impression that they are soft. But watch out, you may only have the appearance that they are soft because the map is too small, there is a good chance that it will flicker frame by frame, due to there not being enough detail in the map, you are better off turning up the resolution to 512, and increasing the filter size/ softness. Another con is that many pieces of software do not yet have a shadow blur falloff inbuilt to work with shadow maps, so you might have to write one yourself or find a workaround if that is something you need. When it is implemented it does substantially slow down the render but it might be worth it to get the realism.
Raytracing is a type of shadow rendering where the path of individual light rays are caluculated from their source (the light) to their destination (the camera) [from the manual]. Professionals can see bad raytraced shadows a mile off, so if your going to use them, make sure you get your settings right. Light Radius values control the softness of the shadows, and is dependent upon the scene size. Shadow Rays controls the graininess of the edges of a softer shadow. Increasing the number of rays increases the quality and the render times "so set it to the lowest value that produces acceptable results" [from the manual]. Basically the more soft you make your shadow, the more rays you will need to clean up the graininess.
Only touch ray depth limit if you have lots of reflection/refractions of light going on, i.e. glass, behind glass and you need it to cast shadows, as in Nigel's project, but beware of render times. If you use the area light in mental ray, shadows blur and become lighter (falloff) as they increase in distance from the object, again at cost to you in time. Raytraced shadows are much slower because of the complex calculations, and because there is no way to get around doing them on a frame by frame basis, but it is relatively easy to get them set up and working.
[extract over. Hopefully having as much information as possible will give you the confidence to use shadows in a creative way]
The character lighting much better. Just compare it with your previous post, and see how much it has improved! It's great to see your progression. Hopefully I can give you a few pointers. I also want to write a bit of technical information, mostly on shadows and their settings, which can be useful for everyone, as I know some people are in the dark about this subject and I would have appreciated it when I started out. I will post it separately after this post, otherwise your post will be too long :)
It's best not to work at your lighting too much without shadows on as it is not giving you an accurate portrayal of the final image, and making it harder for you to judge. You currently have two sets of shadows that I can see. To imitate a sun lit scene you really should only have one shadow, from the key, like your reference. Blurring shadows in comp is not an option, you would be blurring the surfaces as well, but you only want to blur the shadow. Realistic sunlight has shadows that are sharp at the source of the object casting the shadow, that get more blurred the further away the shadow gets, although its over quite some distance (look at tree shadows, and observe how we don't really notice when we are under a cloud shadows because the shadow edges are so soft as the clouds are far away). You will only need a tiny amount of shadow blur, as you are not dealing with large scale objects, so it should be quite simple to do and it will add that softness your reference displays.
[Insert extract on shadows here]
I think you are nearly there with your lighting, but would like to see a bit more contrast. The warmth, intensity and direction of your key light is not really coming across as you have too much fill evening it out. Your skin looks much better, but still needs a bit more light on it and the sweater should not be burning out, so you need to work on the balance in your shading a little (does the sweater have any specular in it? If it does I would take it out) . I think that you can go stronger with the colours of your lights, you might be surprised with how dark and saturated you can go. From what you have shown me, I think you only really need two lights, for your main rig. See the image I created from your individual light images you posted last time. All I did was change the colour and intensities of your two lights and then added them on top of each other. I also painted in some shadows on your key light, as that really makes a difference. Remember that what is not lit by the key light, is lit by the fill light, so your shadows are 'fill' lit. Later on when you are happy with the main lighting, you would benefit from a few localised bounce lights for areas where the white sweater would be reflecting light around, and perhaps a light for her teeth. Consider using a light that only emits specular, for areas such as the legs.
The eyes do look better, they are glowing in the sunlight a bit, so possibly think about lighting them separately using a key light that is less intense (using light linking) or adjusting the shading.
Your occlusion pass does not look quite right to me. It is as if it has a facing ratio image combined with it? Or is it reflection occlusion instead of ambient occlusion? Either way, see the image below, this is more what it should look like. You want to adjust the settings (max distance, samples) so that it is less spread but not crunched (when the image looks compressed). This should fix the problems you are having as the occlusion should not be affecting the sweater other than in the areas of contact like under the arms.
The life ring would look better lit like the character. Unfortunately the reference you are looking at it all rim lit, it makes no logical sense to use that lighting unless you are planning on changing the character lighting to match. I think she will stand out enough for them to not look joined together. I am still not sure about the position of the life ring, as it is not cut off by her enough to have an impact, and its not cut off by the frame either. As for your composition, look at the lines of your pose compared with the Elvgren pose, and see if you can frame it better to get the same intimate feel.
Just tracking the camera should be fine, it would only really be a problem if you are rotating around the model, as you are lighting it from one view only.
Its always useful to have extra passes for defining the skin, and clothes, so that you can make last minute adjustments in comp. Make a new pass, assign constant shaders in red green and blue, to different body parts. Even if you don't use them in the end, its good to have options.
Good luck finishing off, feel free to make ask any questions. I am excited about seeing everyones updates.
PS hope these images work, I have done some investigating.
Thursday, August 6, 2009
As there has been some discussion of Godrays I thought I had better post!
I decided to do a trial comp of one of my shots emphasising nostalgia and glory. I couldn't resist the godrays and actually quite like the result, I think with the theme of my project (the king rising to the challenging opponent and emerging victorious) it is an occasion where I think I can get away with it. I will be careful not to overdo it though and probably only use these effects on the shot where my narration tells of the heroism and resulting victory.
Oh, and I made the decision to move away from Final gather. All the reflection information I liked from the battlefield photos I took are being cast by a hidden constant dome now, and the soft indoor feel that i was getting just felt inappropriate and I decided to use lights instead. It just felt generally more appropriate and render times are obviously much better too.
Here's a render of the actual lighting without the volume effect, so any improvements you could suggest here would be great. I'm having a slight issue with getting my shadow pass out of xsi so I wasn't able to mask the occlusion within the shadows yet as you suggested earlier.
It has painterly romantic feel I think, which I'm quite comfortable with but I still have the option of cooling everything down and using more cool 'Scottish' colours. I did a few tests of how this approach may look. I still think I prefer the overly dramatic top picture, it has an almost fantasy over the top feel which I like.
Anyway, any thoughts you may have will be really helpful at this stage as there is still a bit of time to implement changes.
Wednesday, August 5, 2009
Have posted you a previous post to see if it opens, it does, so I will continue on from there.
Forgot to mention that I have put a simple blue background in the comp, but I don't intend to use this one, I will take reference from your first suggestion of how to paint the background so as not to lose the shape of the model, and also to put a horizon line in
I wanted to say thank you for all your advice on compostion. I still think this needs abit of work, but I think it's getting there. Have tried to tighten in the frame, she probably still looks too central. I'm hoping that the life ring and the suitcase weight her like a triangle. I really like the way her legs cut into the life ring, have taken inspiration from Elvgren on this one, my only concern is that she is not separate enough, colourwise, from the life ring and hence looks like she and the life ring are the same mass.
I have also tried to light the life ring by taking inspiration from Elvgren's painting, by using a strong rim light on it. Do you think this is working with the lighting on the model? I realise that the life ring is looking abit dark from the front, however I'm not sure how to light the life ring from the front without taking away the effect of the rim light.
I have lit the set and model separately:
You advised me against lighting set-ups for an animated camera, however I would really like to track the camera up close to the model's face (please see first image of the lead up to the close up of the face) - I don't intend to rotate around the model - does this sound simple enough?
I really like the lighting set-up that you advised me and I feel that the warm lights have really added a cheerful feel to the model, however I feel as though the face needs to be lit better, but because she is facing away I am struggling to get the flow of light which is traveling up her front to reach the whole of her face. I tried lighting the face from front on, but this seemed to take away the nice shading which is on the side of the sweater.
Do you think that the eyes look better - one still has two highlights, will work on this. I tried to add SSS to the eye, but I just couldn't get a singular strong highlight, and it looked like it was glowing, so I gave up on that one and just used a sphere a transparent sphere as the cornea. I still think these are a bit too white, but will keep adding to them.
You probably don't need this but I thought I would include images of the separate lights I have used for my model. Have just realised that I have no rim light, sorry I thought I did
Finally you mentioned previously that the skirt's shadows need to be more blurred, not too sure how to do this, should I blur them in comp?
Looking forward to your response - really love the way the lighting is bringing my model to life
Many thanks for your previous response,
Just testing this image out to see if it opens. It's my first comp - it looks ugly, but that's fine for a first attempt, sure I will improve.
I tried doing the Ambient Occlusion pass and inverting it so that I warm up the shadows(as you advised). I think this is beginning to work for the skirt, I feel it is giving the skirt a tactile feel. However I don't like what it's doing to the sweater - perhaps I should do a separate AO pass just for the sweater and refine the AO settings so I can really warm up the shadows on the sweater.
Hope the images open,
Welcome back! I am happy that the other posts have helped, and it is good to see that your using Nuke to do some compositing. I can see lots of work has gone into your excellently detailed model, it deserves some lighting to match it.
Horror movies tend to use muted (desaturated) or monochromatic colours, duller lighting and atmospheric fog. I think this is what you mean when you say your image looks gothic. In some ways this is a more realistic portrayal of a night time lit scene, in film this might be used to show the harsh reality of a situation, but I don't think that is what your looking for. To get a sickly look they tend to use greener blues for mid-tones and yellows or pale greens for highlights. They often use black, either by having strong key lighting, casting black shadows, using black sets, costumes or props, or just to have darkness surrounding to frame the action. Darkness is the unknown and it can be scary! Any other more vibrant colours are used to attract the viewers attention, red is particularly good in horror, especially is it contrasts brilliantly with green, and it is the colour of blood!!! You need to avoid some of these filmic devices to move away from the gothic feel you have at the moment.
Volumetrics are often used to create atmosphere, but make sure that you think about what it is saying, and use it for a reason. In the Gladiator reference there is a shot when the protagonist is just about to (but consequently doesn't) kill his enemy in the colosseum, I think the use of god rays is to show the almight and righteous nature of his role, as he is also defying the wicked emperor at the time. In the Lord of the rings poster, it adds mystery, along with the green cast and puts emphasis on Gandalf, who was previously killed, which in my mind glorifies him (is he the saviour?) In Citizen Kane, the volumetrics are used to add atmosphere and intensity to what would have ordinarily been uninteresting. You can sense the pressure in the small enclosed cinema, where these men have been trapped working and smoking. In the Michael Jackson 'Thriller' film, there is an eerie blue fog, we associate fog or mist with scary films, possibly because, like darkness it represents the unknown.
In an outdoor lit scene you would normally only see volumetrics if there is fog, or city smog, another unsavory thought. For your project I think that a glow would be more appropriate, as glows add warmth to a light rather that mystifying it.
For a warmer, more romantic feel you might want to try using the contrasting colours orange and blue. I have done a quick rough paint over, to show you the instant difference it can make. I have not been subtle, I want to show you a dramatic change, and I think it really does add happy warmth and 'Disney' innocence to it. The more saturated you make the lighting, the warmer and happier it's going to feel. Although this is not very realistic it does feel believable to me and I think this is much more like your reference. You can still use darkness, but contrast it with warm vibrant key lighting. As this is a modelling project, and you are going to want to see the model from all angles, you probably want to light it so that you can see the detail on the model.
To really feel the key light source you will need to make it stronger, and then you can pick out the details of the dress with a well placed directional fill light. I would bring your rim light around a tad more to accentuate the shape of the model. You probably want to light each shot separately taking one rig and then tweaking the lights on a shot by shot basis, remember what I said about lighting to camera. You might feel that you need some context for the rim light, as currently the black background implies there is no other light source. Even just a hint of a moon should do the trick. When lighting keep everything exposed (that means no big areas where it is burning out, and no black areas) so that you have some control in comp, but also try to keep the light/dark contrast.
As you have so few lights you might want to consider rendering your lights out on separate passes, and colourising them in comp. I find it quicker but it is not for everybody, so don't feel you have to. Also render out an occlusion pass to remove any unwanted fill or rim light, another option is to have those lights casting soft shadows.
Your floor is looking flat at the moment but when you a put displacement map on it, to define the slabs and add variation, the light should pick it out nicely. You also might want to consider using a specular map on the floor so that you get a different quality of shading on the slabs compared with the cracks.
I would think about rendering the lantern separately, so that you can experiment with it a bit. I would also look at turning off shadow casting on the lantern, and just use occlusion to make it sit better on the ground. Modern street lighting tends to be bright and when you take a photo all you see is the glow of the light. Old victorian gas lamps may have created more of a sillouette, you might want to look for some reference. Possibly the scale of the texture is a little big in places, making the lantern base look a bit like a miniature compared to the character.
One final note on composition. I think you will create a better frame for your models if you move them slightly to the left. Because you have two objects of interest, the lantern and the character, you want to make clear which is more important, and I think that keeping your character more central works better.... although it is best that she is not actually central as that can seem unsettling (see "The Orphanage", and 'There will be blood" posters on the earlier reference)
I know you intend to use more that one camera angle, so think about making an interesting composition for every shot. As you have such strong lighting you do not need a central composition, as you are directing the audiences eye to the character. Use your imagination and see what you come up with.