In the interest of teaching those that wish to learn as well as continuing what some feel was a valuable discussion, I am interested to know specifically what steps you took to arrive at the result you show in the image below if you have the time to list them out:
This was an image you posted in this thread:
Thanks for your time and input.
There have been a number of posts w.r.t. Studio Surface quality. Measuring the continuity to surroundings. Lots of try-this, try-that to get close to an acceptable result.
This post is to illustrate How-To get a high very quality results (against a simple case)
For post readers: its typical of the quality for A-Class automotive exteriors..
Start with a low order surface, increase the order to suit whats needed.
Single convexity "C" (preferred) not "S"
The technique can be extended for different cases.
This Case: The requirement is: To create a high quality middle surface between 2 * High quality base surfaces.
Results and explanation: (image below)
Lets consider the G2 number = 0.1637 and why...
The Upper Curve has a curvature definition @ the ends...
The Surrounding Surfaces also has their Curvature Definition
The guide Curvature and the surface curvature ~has a small mismatch / ~are considered differently.....
You can see the influence the lower "S" edge has...
Inflections in base geometry like this: might be considered undesirable (Hence Example 2)
BUT to achieve the desired quality use Match Edge
see the setup below.
Continuity now shows as very nice...
Now we look at Example 2:
To avoid the "S" the base surfaces should be directed to a common definition.
The surfaces should be shortened: Enlarge or Snip is good to do this..
Studio Surface (G1) is added here...
The spread of poles looks nice.
Match Edge (as shown in Example 1
do this for the left and right edges..
The spread of poles looks very nice.. Read: No hint of an "s"
Very good continuity
Section Analysis: Shows perfect G3 (Flow) across the boundary..
Thank you very much for the response. Pretty slick technique.
I have to ask a couple of things, and they're genuine questions, not intended to sound stand-off-ish - we're past that from now on:
What's been done to the middle surface's guide curves prior to all of this? To me, the poles don't look anything like the originals and seem to have either been recreated or adjusted, particularly the guide curves because it appears like the poles have been bunched up towards their ends (larger gaps between the guide's 3rd & 4th poles). The image below is showing the original curves (divided at intersection), with zero pole movement, only increased in pole count to match your surface degree on the ends and the guide spline degree based on the poles I counted from your 3rd image:
Given the above guide spline pole structures, I don't get the same results with Surface Continuity Analysis - other than G3, my numbers are lower on the right side (G0: 0.0, G1: 0.0050, G2: 0.0028, G3: 0.9315) and the same on the left side with G3 being much higher (G0: 0.0, G1: 0.0331, G2: 0.0236, G3: 1.8528). As you can see in the above image, I did use Divide Curve to break up all the continuous splines and then adjust the splines that affect the end surfaces to be 6 degrees (7 poles) & G2 on both ends - should I have just either Rebuilt or X-Formed those into 6 X 6 surfaces (7 X 7 poles)? I assume this might be affecting the continuity analysis - your thoughts or what is different here? Refer to 3 surfaces in wireframe display with surface poles & knots turned on (no segments/knots in my splines or surfaces) - surface continuity analysis G0 thru G3 is there but fonts are small (need to know how to make them larger if possible - if you know a trick, please let me know. I'm sure it's related to a Windows 7 Personalization -> Window Color & Appearance setting somewhere).
Again, this might be due to different curves or how I made the end surfaces 6 X 6 degrees, but once I apply Match Edge and pick G0 thru G3 continuity and increase the middle surface to 6 X 7, it takes the poles nearest to the shorter guide curve and pushes them to the right making the surface hull very undesireable. This causes a deviation on the other end of the Match Edge - I'm using your settings but getting way different results.Something is definitely different between what you're showing and the file that was originally posted or I am doing something on my side that's causing this extreme difference in the Match Edge results.
Ignore the last question regarding the shift in poles - I missed the 2 Lock Poles options in the Match Edge dialog. Once I did that, the poles stopped shifting. Now I've got zeros for all continuity analysis results.
I find it's interesting how the hull structure isn't more like the results you get when you use Project Movement option in Match Edge - disregarding depth of the poles (normal to the face) the hulls seem better aligned but that tends to result in higher continuity analysis results.