Been practicing my modeling skills more lately and was inspired to try to model this speaker when I saw some nice images on my Facebook feed. Been modeling a bit on Blender using HardOps/Boxcutter, KitOps, MeshMachin3, Machin3tools, Cablerator and Fluent: Power Trip then rendering and applying materials in Keyshot.
For my hero image, I went with a Right Quarter view of the speaker, with the cover plate removed to show off the intricate design under the cloth cover material by back lighting the speaker and plate and tweaking the settings on the material density and specular color to show it realistically. Another beautiful aspect of this speaker is the woodgrain the manufacture used that meshes right up to a metallic plate that houses the speaker components. The side panel with the dials for sound adjustments is also pretty interesting with chamfer cuts on three sides with the right side of it being a gradual wave design.
Modeling the speaker plate, I decided to approach this by modeling the inside design and outer plate separately to make it easier to apply materials in Keyshot.
I also did a frontal image of the speaker without the speaker plate to show off the speaker internals and also the hidden Edifier logo when the plate is installed.
Spent a bit of time getting the details of the speakers correct, from the ridges of the sub, the tweeter design with the bar that stretch over top of it. Finished the front side of the speaker by kitbashing screws that match the actual product using KitOps.
Hope you all enjoy some images I put together to illustrate my journey into 3D modeling with Blender and putting my professional CMF skills to make my model look realistic.
Nice! If I may do a suggestion, use a bit more polygons (or do some sub-division) next time on the bigger round speaker parts. Especially enlarged you notice the polygons.
Thanks a bunch Oscar! Now you mention it, it certainly does appear to not be as smooth. Blender sometimes tricks the eye cause they have a feature that shade smooths lower poly models to make them appear high poly. I try to keep the poly count lower though probably one of the biggest issues I run into in Blender is shading issues, which often requires me to creatively retopo the mesh.
The rear panel I made particularly took me a while to finally get it all smooth due to a bunch of the cuts I made into it for the parts that go into it. The tweeter was also a bit of a pita because of the ngon shapes inside of the circles was pretty tough as well haha. Hopefully I will get better at solving these with time .
Well I really had the same countless of times that I noticed during close-up renders I could have used more polys. Nurbs are not as crazy as the name sounds.
If I create models myself I try to be a good boy and use as much as quads as possible mainly because you can make adjustments so much easier if you can make nice selection loops etc. And I love (hard surface) Sub-D modelling which really needs quads.
I mainly use MODO and it has a great ‘fusion’ tool which basically can create really nice booleans including a little rounded edge if you want and that also needs quads. But with such a surface like that rear panel I wouldn’t care too much. I do notice sometimes you just need some more edges to prevent weird artifacts if you load it into KeyShot.
For me it’s also still a nice journey and it’s great to follow your steps in modelling, motivates me as well to keep modelling. I’ve quite a few weird fantasy things I modelled which still needs some nice renders. But every time I get another idea and things start to collect virtual dust.
Did you also get the UV-unwrapping done? That was something I was struggling with a lot with my own models and while not always needed it took me really a long time to figure out while MODO is in fact pretty friendly for UV-unwrapping (well they say).
I try to keep things as quads as often as possible as you said, it makes boolean cuts much easier and keeps shading manageable.
I decided to make a big push into modeling this year as a personal goal cause I feel my Keyshot skills are decent, but just missing the other half of the 3D in asset generation. In terms of work, my company generally provides models, but I have already been seeing a pay off in my investment into this skill as I can create novel product designs from scratch or even tweak existing models to better work in Blender. Since the CAD I generally work off of is from SolidWorks due to them being manufacturing 3D assets, this often time cause strange artifacts in Keyshot because they are executed as tri’s with millions of faces for what could be much more simple mesh.
I did UV Unwrap this, all though I am still pretty novice at doing it within Blender. I am still doing UV unwrapping in Keyshot as it is much easier for my skill level. I have started getting to understand with Keyshot how to make the software not produce the dreaded all white faces. This is further made easier cause I mark the edges during modeling in Blender for shading reasons, but this ultimately makes UV Unwrapping in Keyshot much easier. It probably would be pretty easy to do in Blender, but it has been pretty fast in Keyshot and I do all texturing currently in Keyshot, I just haven’t found a need to do it in blender yet.
All though talking to some coworkers who are much advanced in modeling than I, Blender’s way of UV Unwrapping is the more proper workflow for industry standards - so I am sure I will eventually have to try to tackle that. Blender has some great render engines with Cycles and Octane, but that is another rabbit hole I will have to begin going down at a later time when I probably delve more into more complicated animations.
Cool! Yeah what I like so much about modelling is that you can visualize your own ideas in a way they look real. Sometimes a struggle because my modelling skills have also loads of room for improvement but once you get the basics a bit it gets more and ‘natural’. I tend to think too complicated sometimes while you better just start doing and if it doesn’t work, start again.
I’m now finishing a little project where I also used some images created by my PC running Stable Diffusion AI, pretty handy if you can’t really draw like me or need a certain photo. Getting what you want is still a lot of trial and error but it’s quite interesting as a tool. Wonder what direction that will go.
And it’s fun there are so many cool programs these days, I love Substance Painter as well, so never a dull moment if you want to learn stuff. Keeps the mind young.
I’ve had my own foray into the Stable Diffusion rabbit hole. I ran a test and put portraits I photographed along side AI generated portraits to see if people could tell the difference. When done well and photoshopped, they ended up being indistinguishable.
The Adobe Substance suite seems pretty darn awesome. I would love for a chance to play with it and invest sometime into creating materials, but alas, it is yet another sub that I really don’t want to shell money out for
.
