Peugeot Onyx: Vray and Mray Model
|
3:03 AM |
No comments |
Peugeot Onyx: Vray and Mray Model by Vlady Oliver, Brazil
I'm videodesigner and creator of brands and templates for television. As such, my harvest is not very advanced modeling, which is not much required in television projects. Still, I think important to master the techniques needed to improve the finishing and see no better challenge than the automotive modeling to improve this technique. I decided to do this Peugeot Onyx, even with few references of its project, to test if I could produce a model with this complexity not even thoroughly dominating these techniques. I followed some of the best trainings on automotive modeling available on the web, that have common points and tips that I will comment here on the development of this project.
My first tip is to create an environment of the object to be modeled, to understand its size and proportions. It is important to follow the correct metric scale as adequate lighting depends on the scale used for modeling. Since I didn´t find Onyx blueprints or its exact measurements, I used a similar model, one Murcielago, to conform their approximate proportions.
Another thing that really helped me understand the process for Turbosmooth was modeling the brand. There are logos available on websites or in vector formats curves, but I decided to redo the mesh both to understand the process and to eliminate unnecessary points in the imported mesh in dwg or dxf. To apply the car itself was not necessary to such detail. Still, thanks for the challenge.
My first tip is to create an environment of the object to be modeled, to understand its size and proportions. It is important to follow the correct metric scale as adequate lighting depends on the scale used for modeling. Since I didn´t find Onyx blueprints or its exact measurements, I used a similar model, one Murcielago, to conform their approximate proportions.
Another thing that really helped me understand the process for Turbosmooth was modeling the brand. There are logos available on websites or in vector formats curves, but I decided to redo the mesh both to understand the process and to eliminate unnecessary points in the imported mesh in dwg or dxf. To apply the car itself was not necessary to such detail. Still, thanks for the challenge.
An important step in the process for preparing the "studio" for modeling are the reference images and their arrangement in subsidiary plans. Some important tips for this stage are:
- See-Through use whenever you need to see "through" the object, the better to locate it or find its components. Click the object and Object_Properties to activate this view helper.
- The reference planes can also be adjusted not to be seen in front of the model, accessing and activating Object_Properties - Backface Cull.
- The reference planes can also be adjusted not to be seen in front of the model, accessing and activating Object_Properties - Backface Cull.
- For better viewing of images in the reference planes I can activate the Material - Show Materials on Viewport As - Hardware Display with Maps. Overpriced will greatly assist in the visualization of the object to be modeled.
The best automotive modelers suggest the beginning of the fairing by two tubes, which would mark the surrounding wheel. It makes sense. With them we "adjust" the proportions of the model flattens the objects and places the origin wheels.
Another important step is to create the wheels tube with a multiple of the number of faces that the wheel shows. In this case, I used 36 faces, which is a multiple of nine pallets of the wheel to be built. 32 deleted faces, getting a cell of four. At the end of the build, just clone it 8 times, with a rotating 40 degrees to the wheel full terms, not forgetting the weld at the contact points of each piece.
As I had already made a modeling training tire for a bike that I modeled recently found unnecessary to repeat it here and just "borrowing" the set of tires that I used on the Murcielago support. I also borrowed lamps from Audi R8 because I found no details of Onyx headlights to customize it. As I stated, I had no blueprint to build properly. Photos, really aligned, have perspective distortions that a project like this, make it very difficult to correct integration of the lines that make up the object. It's approximate, but visually satisfying. You can criticize what is wrong. Another important observation is regarding the mapping. This was not my training's target. So the textures were crudely applied in projected mappings. It is a suggestion; map the model correctly. During construction could take the focus of the modeling, which is the object of my studies here.
Another step that would like to comment, although it is not part of the process itself, is that I did a preview of what I'd be modeling, by the box modeling process. It is an inappropriate process to use in this case to be very "organic" and inaccurate, but it's great for viewing and understanding the component parts of the model. With it I decided that the fairing would have a single modeling to complete the outside, only after being spun off the doors and spoilers to keep the unity of the whole. Many adjustments fairing I confess I do not quite understand, the lack of technical design. I tried to make the unibody approximate way, which results in finishing blunders. I did the best I could to understand what it meant to build.
Another interesting collaboration for the development of the project is to design "contour" to help integrate the components of the model and define the guidelines of construction. I was snapping loose parts until they allow a more accurate view of the set. The build script is always the same: two tubes in the regions of the wheels and the connection between them shaping the cowling.
The lack of appropriate references made me commit the first mistake of the project, the headlights cutting through a ProBoolean (the one I used in the process) that was too small. Although very useful, boolean operations leave much "dirt" in the components obtained with excess points and polygons misfits. Cleaning the model is often more laborious than the approximate methods, but the visualization can compensate the cleaning job that must be done.
I had intended to model the interior of the car and went positioning the references to do so. Until then, I did not know if I should be one model for the interior and another for the outside, due to the time of surrender of the details that would not be visible. Here I had the best experience on GI rendering (indirect ilumination) and without the attribute, because the Onyx's glass sunroof is critically dimming to the interior of the model.
Another step that would like to comment, although it is not part of the process itself, is that I did a preview of what I'd be modeling, by the box modeling process. It is an inappropriate process to use in this case to be very "organic" and inaccurate, but it's great for viewing and understanding the component parts of the model. With it I decided that the fairing would have a single modeling to complete the outside, only after being spun off the doors and spoilers to keep the unity of the whole. Many adjustments fairing I confess I do not quite understand, the lack of technical design. I tried to make the unibody approximate way, which results in finishing blunders. I did the best I could to understand what it meant to build.
Another interesting collaboration for the development of the project is to design "contour" to help integrate the components of the model and define the guidelines of construction. I was snapping loose parts until they allow a more accurate view of the set. The build script is always the same: two tubes in the regions of the wheels and the connection between them shaping the cowling.
The lack of appropriate references made me commit the first mistake of the project, the headlights cutting through a ProBoolean (the one I used in the process) that was too small. Although very useful, boolean operations leave much "dirt" in the components obtained with excess points and polygons misfits. Cleaning the model is often more laborious than the approximate methods, but the visualization can compensate the cleaning job that must be done.
I had intended to model the interior of the car and went positioning the references to do so. Until then, I did not know if I should be one model for the interior and another for the outside, due to the time of surrender of the details that would not be visible. Here I had the best experience on GI rendering (indirect ilumination) and without the attribute, because the Onyx's glass sunroof is critically dimming to the interior of the model.
Despite being discouraged by the render time consuming, I decided to mount the "studio production" model in VRay, already testing the materials that would be used in the project, prioritizing the visualization of what was produced. It ended up being a wise decision, because the surrender was not too heavy and the finish could be tested directly in the final product. I added a light and VRay Plane, which has made a considerable gain in the project environment.
"Cut" the shape before finishing proved to be a wrong decision, because of mismatches that cause the alignment of the plans of each section. Some details, such as the junction between the base of the front bonnet rear view cameras remained unsolved until the end of the modeling, for lack of more precise references. Still defects were unnoticed in the final product.
The back of the model deserves special attention, because it determines the "closing" of the aerodynamic lines that start at the front. I think there was a relative accuracy, since certain joints are only suggested in the imagery of the project. I had to provide for a new cut in the shape, whereas the back has a retractable spoiler, which is only visible in a detail video presentation of Onyx.
At this stage of the model, I ended up having unexpected help from the internet. I met Dawid Żuk, a Polish who speaks Spanish and modeled the Onyx front grid, providing images on the internet. I spoke with him and exchanged views on the project. Until now I did not understand why he modeled only the front of the grid model, but okay. Clay render images he provided helped me understand what I should do. I am grateful to the new digital friend. Things in the web.
Finalized, the grid was very detailed. Many modelers adopt the refinement Turbosmooth in several successive steps, and the mirror and do not modify the model. Not used neither one thing nor the other, because I wanted a very light model to be refined gradually. The usual collapse is finalized object, but I didn't want to lose its construction history, just to explain the steps of modeling properly. The grid, however, was not grouped by a set symmetrical due to the Peugeot logo. Still with all its elements are retained their modifiers to enable understanding the modeling steps. It is a model for studies, not presentation. Best that can be "taken apart" if necessary.
Returning to the refinement of the shape, I was complementing the back and keeping the unibody uncut, just setting the seams as tail lights and exhaust fan support.
Continuing the rear of the model, includes new contours to better understand its closure by adjusting the cut for the LED panel, logos rear bumper and exhaust set.
The back of the model deserves special attention, because it determines the "closing" of the aerodynamic lines that start at the front. I think there was a relative accuracy, since certain joints are only suggested in the imagery of the project. I had to provide for a new cut in the shape, whereas the back has a retractable spoiler, which is only visible in a detail video presentation of Onyx.
At this stage of the model, I ended up having unexpected help from the internet. I met Dawid Żuk, a Polish who speaks Spanish and modeled the Onyx front grid, providing images on the internet. I spoke with him and exchanged views on the project. Until now I did not understand why he modeled only the front of the grid model, but okay. Clay render images he provided helped me understand what I should do. I am grateful to the new digital friend. Things in the web.
Finalized, the grid was very detailed. Many modelers adopt the refinement Turbosmooth in several successive steps, and the mirror and do not modify the model. Not used neither one thing nor the other, because I wanted a very light model to be refined gradually. The usual collapse is finalized object, but I didn't want to lose its construction history, just to explain the steps of modeling properly. The grid, however, was not grouped by a set symmetrical due to the Peugeot logo. Still with all its elements are retained their modifiers to enable understanding the modeling steps. It is a model for studies, not presentation. Best that can be "taken apart" if necessary.
Returning to the refinement of the shape, I was complementing the back and keeping the unibody uncut, just setting the seams as tail lights and exhaust fan support.
Continuing the rear of the model, includes new contours to better understand its closure by adjusting the cut for the LED panel, logos rear bumper and exhaust set.
At this stage, the external modeling almost closed, I started numbering the different polygons to receive materials by the Multi / Sub Objects. It wasn't the target of my studies here and I ended up making several materials that could be grouped in the same main material. Homework for an upcoming training; texturing.
Good old luminant material, which does not allow using Mental Ray, works very well in VRay, even as part of a luminant object as lanterns and panels. There are great tutorials on the web that explain how to make neons and flashlights with these renderers. Jeff Pathon has excellent automotive materials in their designs, but most were developed for Mental Ray, I'll use in the final version. I adopted the same material of the Audi R8 Madham, which are excellent for the study in VRay. Work perfectly.
A very common process in complex objects like this is to duplicate some of their ways to model the remaining detail, how to use the edge of the glass to model the internal structure that supports or rear hood to model the retractable wing. The side window may also be positioned considering the boundary between the glass ceiling and the top of the door. The object takes shape, like knitting a sweater made by our grandmother in her rocking chair. The process is very similar.
The outside of the model is finally completed, adding side air outlets, the rear view cameras and glasses. The ceiling, which I did not comment until now, was the aftermath of the initial reference model. I returned to model it properly now, starting from the finished shape. Also discarded the mirror, because the object is glass and even zeroing all touch points the dividing line between the two sides symmetrical visible. The solution was to weld the two pieces and finish the ceiling as a single object, without amendments.
At this stage of the project I had been able to cut the shape without worrying about the amendments. For this grouped objects and I linked door all in a row-axis, in order to open the door on this line, the axis location. The door has a small angle, which I adopted at 5 degrees to open without hitting the rest of the stuff. From the pictures, it looks like it slopes slightly upward, which dismisses the idea of having an axis parallel to the main axis of rotation and use in view. It's a detail, but they say the devil lives in them, hehehe...
Finalized the shape and opening the door, we can focus on internal modeling. One of the details that was not remarked the need to extrude the polygons of the two tubes to make an initial "lock" on wheels. I do not know how it was solved in the project, but the wheels and the internal cell are dangerously pasted, the references that I got. Will require some tricks to hide this problem. Again the lack of blueprints compromises the accuracy of the project.
My first move was extrude some lines that I did as level curves to better visualize the fit between the various components and their volumes inside the vehicle. It is easier to model each piece from their sizes and positions previously defined. It took extra time to understand what is every thing inside the model. As I am not an automobiles lover, modeled certain things exactly without knowing what they are, which compromised its features. Even so, a complex project like this, the level of detail needs to be properly solved, because no one will see outside if the car has a cigarette lighter with a brand embossed or not. My limit was the arrangement of parts within the model so as realistic as possible. As a professional video, I'm not familiar with giants render processes that consumes all the coffee of the environment. This is my first.
Mistakenly, I began the interior by refining the steering wheel, with a match in single camera image that had the object. It was the wrong decision most of all, because the steering wheel is aligned with the principal axis, which would facilitate its modeling absurdly. Just went out this detail after examining the very few images that I had. I think that the right answer is always the simplest. And the Onyx seems clumsy to drive. Paper seats ? I just can believe it sitting on them.
Cockpit was slowly taking shape, always based on contour drawing that was understood as a form in the images. I refined the structure under the glass ceiling, seats and hoods added.
The car roof has an internal structure under glass cockpit of the same material as an extension just like a sandwich of several layers. His model casts a shadow inside the cockpit that can only be overcome with a render in GI, which doubles its running time. So I adopted one Onyx convertible kkkkkk. It is the only way to see what I'm doing without losing absurd time.
To take a break in complexity I started to face, I returned to the external refinement, modeling the "spider" that form the rear exhaust. Also refined the bumper, which deserves a mapping more sophisticated, for the complexity of the piece. I put a projected box in it. It was childish, but I think I'll be forgiven for rudeness. The focus, as always, is the modeling of object.
At this stage of the project, the amount of small errors will compromise the model in a scary way. I was making a list of corrections. I redid the door several times, because it would need to fit both the cockpit and the model, and has a refinement which is an internal extension of all the dynamic lines of the car. Even so, it was only satisfactory. The junction between the door and the inner lining deserve some more pictures of detail, something that had not. This does not compromise the final product. It's what matters. Many modelers are lost in the details and whether or not to refine them. Adopted not do it, not to lose focus of the project.
In a first stage, I created the cockpit volume, linking their contours. The interior of the vehicle begins to take shape.
Adjusting the cockpit, I could refine the doors a bit more, making the two parts engageable over one another. It took work.
Gradually the model is gaining more detail, by approximated processes. I added plates for the front panels, detailing the docking port on the shape, dashboard and steering axle, all more or less embedded in each other. Lacks much.
Details like the hood latch and central internal doors have been added. Some maps were needed and I stopped to produce them.
I must confess that the production of this project had never seen a tutorial to produce seamless textures, or without amendments. I will need them to map the cockpit and other details with Onyx finish "paper wood" as they defined in industrial design. I left one of the pictures I found of that texture and distorted the image texture for a more planned in Photoshop. applied a layer - Flatten image - because I wanted to use only the visible part of the image.
I started with the same idea that I use when I want to do a looping any pictures if I crop the image in four, from the center, each piece will be the continuation of the next and may be amended without problems. For this we can use the Offset tool in Photoshop, which makes the image "moves" by the frame. How do I use HDTV format (1920x 1080) in my productions, my initial texture had this format. Ctrl + J to copy the layer, Filter - Other - Offset to make position adjustment in the frame.
Applying an Offset of 960 and 540 in this particular image whose size is twice these values, we have a cropped image at the center, where the edges are continuous. As we only want the edges of this result, just delete the "diamond" formed by the undesirable partitions and have a texture "tileable" or continuous.
Details like the hood latch and central internal doors have been added. Some maps were needed and I stopped to produce them.
I must confess that the production of this project had never seen a tutorial to produce seamless textures, or without amendments. I will need them to map the cockpit and other details with Onyx finish "paper wood" as they defined in industrial design. I left one of the pictures I found of that texture and distorted the image texture for a more planned in Photoshop. applied a layer - Flatten image - because I wanted to use only the visible part of the image.
I started with the same idea that I use when I want to do a looping any pictures if I crop the image in four, from the center, each piece will be the continuation of the next and may be amended without problems. For this we can use the Offset tool in Photoshop, which makes the image "moves" by the frame. How do I use HDTV format (1920x 1080) in my productions, my initial texture had this format. Ctrl + J to copy the layer, Filter - Other - Offset to make position adjustment in the frame.
Applying an Offset of 960 and 540 in this particular image whose size is twice these values, we have a cropped image at the center, where the edges are continuous. As we only want the edges of this result, just delete the "diamond" formed by the undesirable partitions and have a texture "tileable" or continuous.
Returning to the interior of the model, we can now refine it and map it properly, and produce cuts and fittings required for the internal equipment.
Next step in the modeling are the internal equipment and devices such as the center console and seat belts. I did a search of this equipment as the project showed the name of the manufacturer, as the four corners of the seatbelt and turbo. Detailing unnecessary. Some cylinders and bricks solve the problem, since the materials are convincing. Adopted the policy of least effort, because not intend to render images of the inside of the vehicle in a first step.
Here, one more stop to external refinements, such as cutting the front plate and the ONYX logo golden vehicle displays at the bottom of the shape. I took to the railings of the side air outlets and slightly improve their collectors, still provisional.
This time I decided to take the steering wheel and speedometer for real, from cylinders properly deformed and fitted together as envelopes of the final object.
In parallel, I decided to "comb" the fairing, including lateral fins and ending the region of the junction of the windows and doors.
Rework. Again. The door and its breakdown caused numerous internal remakes in this stage, because the object has to fit properly with the internal and external lines of the projected object, with little references on how this was solved by its designers. The result was approached and has structural problems, but it serves to a scene without much detailing. It's what matters right now.
Next step in the modeling are the internal equipment and devices such as the center console and seat belts. I did a search of this equipment as the project showed the name of the manufacturer, as the four corners of the seatbelt and turbo. Detailing unnecessary. Some cylinders and bricks solve the problem, since the materials are convincing. Adopted the policy of least effort, because not intend to render images of the inside of the vehicle in a first step.
Here, one more stop to external refinements, such as cutting the front plate and the ONYX logo golden vehicle displays at the bottom of the shape. I took to the railings of the side air outlets and slightly improve their collectors, still provisional.
This time I decided to take the steering wheel and speedometer for real, from cylinders properly deformed and fitted together as envelopes of the final object.
In parallel, I decided to "comb" the fairing, including lateral fins and ending the region of the junction of the windows and doors.
Rework. Again. The door and its breakdown caused numerous internal remakes in this stage, because the object has to fit properly with the internal and external lines of the projected object, with little references on how this was solved by its designers. The result was approached and has structural problems, but it serves to a scene without much detailing. It's what matters right now.
The interior had to receive some new improvements and refinements, to adjust the front panel and the rear panel, making the connection between the inside and outside of the model.
After much work, finally all the pieces were in their place, just missing a few refinements. Includes taillights, which were a separate chapter in this modeling, since they were 3 buds in bevel with folds quite pronounced. As I did not have precise references of these objects but I know they are the "soul" of a project such as resemble the eyes of a living being, I decided to leave them for the end, when I had enough familiarity with everything I was modeling and adjusting. I divide flashlights and lanterns in "body", "lamp", "glass" and "fittings" by sheer intuition and analyze other similar models of cars in CG. Seems to have worked.
Steering wheel, dashboard with DVD, speedometer and dashboard properly refined and mapped. Basic maps, as adopted in the entire project. I felt the need to divide the material into the cockpit Multi / Sub-Objects, because the interior trim has a variation in texture, more color champagne, whose division was clearly visible. A very reasonable level of detail for an object whose purpose is not the detail. Will work.
Renegade means of presentation images, Onyx has a retractable top panel with push buttons and LED screens, embedded in the ceiling inside. Positions presentation of the vehicle will never favor this console, which can go unnoticed and wouldn't need to be modeled. It's too much, hehehe. It's part of the project, we will include it in the model even without ever previewing it.
Only with the final model and the first renders I was able to see some brutal mistakes I made during the modeling. Most of them were glaring headlights, small and positioned in the wrong place, taking the personality of the vehicle. The other, less visible but equally troublesome, was the lack of thickness in the side air intakes. I had to redo it. Finally, I "borrowed" a studio used by Jeff Pathon and his famous 32 Ford to implement the model for the Mental Ray. It seems to have worked well.
Peugeot Onyx MRay
Peugeot Onyx MRay Opendoors
Peugeot Onyx Vray
Textures 1 | Textures 2 | Textures 3
I hope this making of is illustrative for anyone who wants to venture into difficult models. I am far from a great modeler. Just curious, who loves 3DSMax and the challenges it presents.
Creative success to anyone here and thank you for attention.
Steering wheel, dashboard with DVD, speedometer and dashboard properly refined and mapped. Basic maps, as adopted in the entire project. I felt the need to divide the material into the cockpit Multi / Sub-Objects, because the interior trim has a variation in texture, more color champagne, whose division was clearly visible. A very reasonable level of detail for an object whose purpose is not the detail. Will work.
Renegade means of presentation images, Onyx has a retractable top panel with push buttons and LED screens, embedded in the ceiling inside. Positions presentation of the vehicle will never favor this console, which can go unnoticed and wouldn't need to be modeled. It's too much, hehehe. It's part of the project, we will include it in the model even without ever previewing it.
Only with the final model and the first renders I was able to see some brutal mistakes I made during the modeling. Most of them were glaring headlights, small and positioned in the wrong place, taking the personality of the vehicle. The other, less visible but equally troublesome, was the lack of thickness in the side air intakes. I had to redo it. Finally, I "borrowed" a studio used by Jeff Pathon and his famous 32 Ford to implement the model for the Mental Ray. It seems to have worked well.
Peugeot Onyx MRay
Peugeot Onyx MRay Opendoors
Peugeot Onyx Vray
Textures 1 | Textures 2 | Textures 3
I hope this making of is illustrative for anyone who wants to venture into difficult models. I am far from a great modeler. Just curious, who loves 3DSMax and the challenges it presents.
Creative success to anyone here and thank you for attention.
0 comments:
Post a Comment