Modeling approach?

[fixeighted]

I have yet another oxygen chamber to build. I've made probably 20 different types so far but this one had me stumped for quite a while, in terms of how to go about modeling it.

It appears simple but isn't quite so, as it needs to be one contiguous shell and there are some rather awkward curves/cutaways within it.
There is scant reference available apart from dimensions and so basically the attached is it, mostly.
The question is; how would you go about creating the main shell?, as accurately as possible and of course with good topology.

Although I did eventually find a rather cool method I'd be interested in seeing any other approaches.

Naturally I considered CAD (there exists a model used in their advertising, though it falls far short of how it looks in reality) but these days
I don't have access to anything - the last I used was PowerNurbs but have since lost the license.
Plasticity looked like an option but I didn't want to buy that without being certain it could do the job convincingly, so if anyone uses that and
can prove the concept I'd be interested in the workflow.

If anyone sees this as a challenge then I'll post my method to compare techniques

[fixeighted]

I thought I'd update this post, as I completed the model and the technique might be helpful to someone

Due to how awkward the main shell was, being one moulded object with many compound curves, variable fillets and other tricky stuff, I needed to
come up with a method to be able to build/edit it in realtime with the ability to easily adjust it.

Ultimately I built it from a plane, using the simplest of solutions to edit the base object with few vertices whilst seeing the fully smoothed end result.
The key is the edit poly mod which allowed this workflow and which was a happy accident whilst trying various things along the way
It took me a couple of days of other failed approaches to get there but the result was ideal and as relatively simple as it could be

[glorybound]

Nice modeling... I mosel houses, which is easy . How did the final turn out?

[Joelaff]

Nice approach. I love it when you find a trick to achieve your goals. It so often comes down to simplifying things in your head like you did here.

[fixeighted]

Yes this was a good hard surface exercise and it turned out almost 100% perfectly, apart from one area, where the cylinder here is booleaned from the top.
It proved impossible to fillet the top edge and have it maintain proper curvature in the base level, given all the compound curves going on in that area, which is annoying.
Luckily a large radius edgetex took care of it, so job done

I've always built extremely diverse and complex models, since the dawn of 3D...had to, as it was never like it is today with, mostly, everything available online
and a tutorial for everything you can imagine
Doing this you tend to amass all sorts of creative approaches to modeling.

Oddly, years ago I remember something seemingly simple, a high res tennis ball, had me completely stumped for a loooong time lol...just remembered that.

[Joelaff]

I recall the tennis ball conundrum as well, back in the LightWave days I think, maybe even before then. I booleaned out the seams, which worked great once I figured out the shape to use. Some Sasquatch (plugin) fur and it looked great.

[fixeighted]

I can't remember how I did it at the time but I couldn't figure it out when I just tried...you forget what you've forgotten
I had to look it up and of course it's so simple in a couple of different ways.