Very close! :o The flow is pretty much correct by now. Seems like it is not the perfect one if you look at the curvature analysis. As usual, I'll most probably rebuild it later.
The windscreen area *should* be easy because it is part of the plan. I hope it goes as planned or else it's gonna be ... tedious.
just wondering whenever i go to match a curve with g2 or g1 the curve shape totally screws up , how do you match them and why does it take so long for you to do it? remember im new to this if sound ignorant i just dont get what is right or wrong i mean how do you get g1 or even g2 single span (basically a-class) without totally changing the shape ? do you insert knots kinks or something coz i get creases with knots, but what bugs me the most is why do you spend so long tweaking a little part when i seem to do it 5 minutes i must be missing a s#$%load! of stuff in between!
ps that ferrari is looking really really really good congrats!!:buttrock: :buttrock: :buttrock: :buttrock:
@anthony27:
I don't use knots on main surfaces because they will create multi-span curves. Basically the thing that takes time is to get the following things right:
- Achieve good continuity/flow
- Keeping single-span surfaces
- Retaining the proper shape
As you said, when you match curves or surfaces, it kinda get screwed up. So ideally what you want is to match them but not screwing up the shape. To do that you need manual adjustment along with <Match> or <MatchSrf>, which is the real tedious process.
Try to enable control points display when you do the matching. This way you see how things are being "matched". It really takes me time to get things into shape, perhaps pro gets it done faster.
If you want Class A results, I believe you'll need to make sure the flow and continuity is good (on the main surfaces). You can easily detect bad flow with curvature analysis. Try to do curvature analysis on a 5 minute part, I bet you'll see why.
I can't help it. I feel bored so went on and work on the car again after getting some rest. The hood is good now, all is left is to match the area near the windscreen. The 2 surfaces highlighted in attachment #1 are mock-up only. It's gonna be tough dealing with this area.
I'll study some references later for accuracy. In case you're wondering, I look at reflection of the original car to figure out roughly how they (Ferrari) modelled the car (roughly estimating surface boundaries).
It's 8:30 PM now and I'll go for my dinner. Continue later. :chinese:
Windscreen area isn't as tough as the rest. No single-span surfaces there though. Couldn't keep it single span because of the need of continuity matching with trimmed area of the hood. I *might* rebuild these tomorrow with cleaner surfaces.
Next to do, windscreen and the area behind the doors (air intake). I'll revise the windscreen area before I continue. The hood is pretty much final, won't be tweaking it much unless there is something really wrong.
One thing to note, I never join my surfaces. When you join surfaces in Rhino, the edges will be merged with the surrounding surfaces, creating an approximation of the edge. As far as I know, approximation is like trimming, it does not give you accurate result. I find it a good practise to not join or alter the edges.
I will join these surfaces together after the filleting stage when they are all final.
11:30 PM now, getting ready to sleep already. Hope you guys enjoy the images below.
I'm preparing for the windscreen now. Rebuilt the hood area to flow with the A pillar, all single span. Although I had to use degree 9 surfaces for those parts.
@andrei:
I'm using Rhino 3.0 SR5 at the moment. Yes it has proper curvature continuity under "tight" tolerance.
I am not sure what you meant exactly, whether Rhino has the tool for curvature check or do I have accurate tools to show the continuity of my work.
To answer both, yes. Rhino has <CurvatureGraph> to do that although not as robust as other softwares. And the answer for the next, yes I think I've nailed the curvature continuity quite well. Definitely not "perfect" but under low tolerance.
Here are some screenshots with curvature graph attached, #1 is curvature graph of U direction, #2 is of V direction. On some surface boundaries, you see one curvature graph because both edges share the exact same curvature and CV's position.
Attachment #3 and #4 are zebra shots with very fine line on both different direction. I have to admit the area near the A pillar on the hood isn't the prettiest because of the matching with trimmed surface edge. I will be working on that when I work with the A pillar.
Note that the surfaces are not joined together therefore mesh are not forced to join at the edges. They are all individual meshes. This way you can't really hide continuity errors.
Let me know if you notice anything wrong. We can't really get perfect curvature continuity at trimmed surface edges. It is hard, so what I do is to try and get it curvature continuous under very low tolerance.
Just to be curvature continuous isn't everything, it is much harder to get the shape and flow correct.
Also let me know if there is anyway I could do it better, regardless of what software package you're using. This is basically NURBS with plenty of manual modelling which I'm sure the same across different softwares. Accuracy of NURBS is not even a matter here, as I work under low tolerance.
Some related off-topic notes for everyone:
I know that with StudioTools you can display continuity check at surface boundaries which really ease the whole process. This is definitely a feature I would like to have in Rhino, given that it doesn't bloat the software. Perhaps this is also the tool mentioned by andrei earlier.
I find Rhino much easier to use compared to StudioTools, it is less bloated and I get more control with it. I have tried StudioTools myself for few days in the past but I find it way more intuitive to go with Rhino instead; things such as command line really makes a lot of difference.
I do know that StudioTools has plenty of tools that Rhino didn't have, but so far things can be done in Rhino, just in a different way.
It is easy to start modelling with Rhino that most people can easily come up with visually good-looking models. I believe people have bias thinking Rhino is less accurate and not suitable for serious modelling because of the amount of these quick models from Rhino.
I have yet to learn what is less accurate or less serious about NURBS in Rhino. Don't get me wrong, I am still learning the differences of both softwares to find out what I am missing. At the moment I'm assuming it is a biased opinion regarding Rhino. Unless of course, if someone shed me some light on that area.
nice work the car is looking very smooth, i know you use the zebra to check for continuity and stuff like that but what are the other ones (the thermal graph, etc)for? i dont know much about rhino or even taking a look at studio tools.
The work you are doing here is absolutely incredible, I love the attention to detail. A couple years ago I modeled this part of a Modena using almost exactly the same method so I can feel your pain, I remember making 5 or 6 different hoods before finally getting one I was somewhat satisfied with. As the pic below shows it looks like you've pretty much nailed the curvature so far.
Appreciate the thorough posts, for someone like me who is at an intermediate level, writeups like these are a rare and valuable resource. Looking forward to seeing how you tackle the rear intake scoops. Later
Wow! Thanks Nate! I've got plenty of references but nothing is as CLEAR as that one you've posted. All I know is that even Ferrari is not perfect (check attachment).
Hope you'll get some time to work on a car with Rhino and show it up here. :buttrock:
Back on topic: Currently working on the A pillar. A bit tricky here. I'll post some updates later.
to be honest with u i love ferrari , but when i went to the motorshow i saw the 430 close up and took some pics of it and i was less than impresses to be honest in some parts it just didnt look right at all
Thanks Nate! I've got plenty of references but nothing is as CLEAR as that one you've posted. All I know is that even Ferrari is not perfect (check attachment). 
