Jump to content

Airpods max modeling


Go to solution Solved by Cerbera,

Recommended Posts

20 minutes ago, Cerbera said:

Look at it from ALONG the surface, and the reason for lumpiness will become apparent.

 

Here's my version, doing the same thing. With your eye, follow each individual loop, and see how a nice a curve you have to it. Then check the same in the loops following in the opposite direction. If you don't have nice curves, lumps will start to show.

 

image.png.ca150b6663fb31ae162a5498a8b669e5.png

 

This way we minimize the work SDS has to do, and so it doesn't fight us...

 

image.png.1aca846d8444dadffff5a70e83fc6a72.png

 

CBR

 

Should I start from scretch again? or is it possoble to make it from this?

image.png.0e031b99b3bcf304b5ed030b271e40b3.png

Link to comment
8 minutes ago, HappyPolygon said:


Is it preferable to be modeled flat and them curved via a deformer ?

 

That certainly would produce a very even curve there. Bend deformer seems ideal, moreso if it was made under symmetry.

 

CBR

Link to comment
46 minutes ago, Jaee said:

 

Should I start from scretch again? or is it possoble to make it from this?

image.png.0e031b99b3bcf304b5ed030b271e40b3.png

 

That is more polys than you need. You can lose the ones either side of the horizontal centreline I reckon. Or, use the double symmetry as recommended, in which case, the starter points you need are simply these, which I got from a 4 x 4 plane I then deleted most of !!

 

image.png.c16d9f0845ef8427ea325659025f6914.png

 

...which in symmetry(s) and SDS give you this, with just 4 polys !

 

image.png.587f2dd75235849661b5eaa0d0c580d3.png

 

Now you can just slightly move the centre poly out a bit to get the most perfect curved surface there...

 

image.png.6e554db335bb4412d3c7e3000d36b21c.png

 

Now we can make SDS do all the work, and redirect the edge flow, by just dragging a copy of the perimeter loop back a bit to get this...

 

image.png.37ce797b44759d232741e9f5eb724cfb.png

 

Look how magnificently smooth and even it is, with almost zero effort - I have moved just 1 poly ! Such is the power of JUST ENOUGH polys to describe the form.

 

Continue, in that vein ! 🙂

 

CBR

 

 

Link to comment

So, if we continue that round the back, we can finish off the base mesh for that bit, with equally few polys, like so...

 

image.png.48e9620715202098b955bc3b01b17bee.png

 

Now, of course that isn't enough mesh resolution to add the smaller holes and details, so next we apply 1 or 2 levels of subdivision to get perfectly even geo, again, without having to move anything ! This is called 'establishing curvature'; baking an initial level of subdivision into the mesh which allows us to continue with the smaller details.

 

So having checked that for ultimate smoothness of curvature with a soft metal texture...

 

image.png.4718da0e3b7590e83a33d529edad3fb2.png

 

...here's the same mesh, but now with L1 SDS applied, which is eminently more suitable for making those holes at the top, which is presumably the next stage of this bit. I am also down to 1 way symmetry now, as the bottom of these cups is slightly longer / more rounded than the tops.

 

image.png.9676dfa78d7f4eee176092bc4d45b630.png

 

CBR

Link to comment
3 hours ago, Cerbera said:

So, if we continue that round the back, we can finish off the base mesh for that bit, with equally few polys, like so...

 

image.png.48e9620715202098b955bc3b01b17bee.png

 

Now, of course that isn't enough mesh resolution to add the smaller holes and details, so next we apply 1 or 2 levels of subdivision to get perfectly even geo, again, without having to move anything ! This is called 'establishing curvature'; baking an initial level of subdivision into the mesh which allows us to continue with the smaller details.

 

So having checked that for ultimate smoothness of curvature with a soft metal texture...

 

image.png.4718da0e3b7590e83a33d529edad3fb2.png

 

...here's the same mesh, but now with L1 SDS applied, which is eminently more suitable for making those holes at the top, which is presumably the next stage of this bit. I am also down to 1 way symmetry now, as the bottom of these cups is slightly longer / more rounded than the tops.

 

image.png.9676dfa78d7f4eee176092bc4d45b630.png

 

CBR

 

the length marked as red and blue Look different. Is that natural? or not correct?

image.png.0106717e985003e368b31488fd068737.png

 

 

Link to comment

Be led by the reference. From what limited reference I have seen of that part they are fairly even, and mine are so because I scaled loops in on Y and X separately so that they were. Also I should check you are not building the white bit out of the same mesh as the rose pink bit. They are separate parts in the real world, and should be like that in the model as well, as we need radically different topology there. You can leave the outer mesh open - it doesn't even need thickness as it will be fully obscured by the other part(s).

 

Also, when you are working with staged subdivision, don't add control loops to the base mesh, as these too will get doubled in density by the subdivide, and then you'll have to go around dissolving half of them. If you look closely at reference, you'll see the curves on the inner rim of this object are incredibly smooth, and don't actually require control loops (except in one place), because the density of the polys in that area is enough to define the surface curvature alone.

 

CBR

Link to comment
35 minutes ago, Cerbera said:

Be led by the reference. From what limited reference I have seen of that part they are fairly even, and mine are so because I scaled loops in on Y and X separately so that they were. Also I should check you are not building the white bit out of the same mesh as the rose pink bit. They are separate parts in the real world, and should be like that in the model as well, as we need radically different topology there. You can leave the outer mesh open - it doesn't even need thickness as it will be fully obscured by the other part(s).

 

CBR

 

Can I see your working file? I just want to check what's difference between yours and mine...

Link to comment

While I am up with pesky insomnia I will go ahead and show you how to make the inner white plastic part of the main shells...

 

First get yourself one of these...

 

image.png.754ebc152672075091c6b5e1c694abe2.png

 

Then, Honeycomb array Clone that, sort of like this... (10 x 9 rows rather than the 10 x 7 shown below; you actually need 6 holes at the widest point).

 

image.png.13943624c269b99589185cd0aecc4bdf.png

 

When you have that right, you can make the cloner editable, off the copies we don't want, and then put the rest in a connect to join them. Make that editable too, and then delete all but one quarter of it like so, and restore 2 way symmetry, which makes the next bit lots quicker...

 

We'll also need a perimeter loop for this section, so create a very even one like I have with exactly the right amount of segments to join to the inner array.

 

image.png.d5073a80b00f680071de489542780d92.png

 

And then we just need to play 'connect the meshes', and relax (Iron / brush smooth) the polys we create between these to make a very even, regular flat mesh.

 

image.png.928ad81677d2a9267c95dbcdd8d93290.png

 

In case you wonder, we don't have to care about the minorly complex 6 point poles here because this part of the surface is entirely planar.

 

This one will not be getting staged subdiv, so we can use control loops here (outlining the holes) and for the next bit, adding depth to this piece....

 

image.png.637f56818fd818ca0bb61afc33f3b7e8.png

 

CBR

 

 

Link to comment
12 hours ago, Cerbera said:

While I am up with pesky insomnia I will go ahead and show you how to make the inner white plastic part of the main shells...

 

First get yourself one of these...

 

image.png.754ebc152672075091c6b5e1c694abe2.png

 

Then, Honeycomb array Clone that, sort of like this... (10 x 9 rows rather than the 10 x 7 shown below; you actually need 6 holes at the widest point).

 

image.png.13943624c269b99589185cd0aecc4bdf.png

 

When you have that right, you can make the cloner editable, off the copies we don't want, and then put the rest in a connect to join them. Make that editable too, and then delete all but one quarter of it like so, and restore 2 way symmetry, which makes the next bit lots quicker...

 

We'll also need a perimeter loop for this section, so create a very even one like I have with exactly the right amount of segments to join to the inner array.

 

image.png.d5073a80b00f680071de489542780d92.png

 

And then we just need to play 'connect the meshes', and relax (Iron / brush smooth) the polys we create between these to make a very even, regular flat mesh.

 

image.png.928ad81677d2a9267c95dbcdd8d93290.png

 

In case you wonder, we don't have to care about the minorly complex 6 point poles here because this part of the surface is entirely planar.

 

This one will not be getting staged subdiv, so we can use control loops here (outlining the holes) and for the next bit, adding depth to this piece....

 

image.png.637f56818fd818ca0bb61afc33f3b7e8.png

 

CBR

 

 

 

Thanks! I just created this mesh. But, I have no idea of combining it with the body part I've created earlier. Could you teach me?

image.png.609669f069ba9171eeb4ff3f542519c9.png  image.png.42081715e72d741a970aa18646e86086.png

Link to comment
×
×
  • Create New...

Copyright Core 4D © 2023 Powered by Invision Community