Thanks so much for this great tool. I have been messing about with it to try to further my AI assisted research and a couple of issues have come up. For computing mode overlap integrals I need the raw nodal displacements before either step - correct relative amplitudes between modes, and hard zeros outside the body rather than the smoothed extrusion. would it be possible to include an option to give the unprocessed field?
I'm trying to run comparisons between standing waves which are computed at the boundary, and once the smoothing has blended it all together its impossible to recover the information.
Also the native fem-accelerator 404s/403s on download. the JS fallback works but its slower.
also the wasm backend is broken apparently (TypeError : e is not a function) So everything silently lands on the slow JS path.
that will probably make more sense to others than it does for me, but that is apparently a limitation I've hit.
Whaaaaouuuh! What a crazy tool bob, a thousand thank's!
I have a lot of ideas that I would imagine for it but I realise this represents a lot of work...
But the ability of combining mutltiple frequency onto multiple mediums would be amazing. Maybe being able to setup sources for the perturbations as well as an overall medium like air encompassing the whole grid, or water contained in a hollow cavity. Like the lattice of the singing bowl and the water it is containing. Or those egyptians vases filled with air and/or water. Thank you so much for that amazing tool!
Hi Stan, glad you like the beta. It had a lot of bugs. I've been working hard on those over the last couple of days, and I've just released a huge update, so check it out! Delete your cache and do a forced refresh. Enjoy!
What I need is the field before both steps. Specifically, per mode:
• the raw nodal displacement eigenvector (3 components per node), in whatever units the solver works in — no per-mode rescaling
• the node ordering or node positions, so values can be tied to locations
• ideally the post-cleanup occupancy, since the largest-connected-component and despeckle passes can change which voxels are actually in the body
A flag on the solve message — something like rawField: true — would be perfect. Frequencies are already fine; it’s only the fields.
Why: I’m computing mode overlap integrals (∫φᵢ·φⱼ over the body) to get nonlinear coupling coefficients between modes. That calculation needs true relative amplitudes between modes, and genuine zeros outside the boundary. The normalisation I can undo myself given the geometry — the smoothing I can’t, because once real boundary values are averaged with fabricated exterior ones the information’s gone. And the boundary is exactly where the physics I’m after lives.
So I basically need the non-smoothed boundary information for computing as opposed to 3D printing which it is optimised for - I believe?
Hi Bob - Just trying to get more eyes on your work.
9/11 Bombshell: The Towers Were Resonated Into Dust – New Science Proves It
Bob Greenyer’s WTC Dust Analysis and Standing Wave Physics Finally Explain Dr. Judy Wood’s “Dustification” Evidence
Article: https://911revision.substack.com/p/911-bombshell-the-towers-were-resonated
Thanks
Thanks so much for this great tool. I have been messing about with it to try to further my AI assisted research and a couple of issues have come up. For computing mode overlap integrals I need the raw nodal displacements before either step - correct relative amplitudes between modes, and hard zeros outside the body rather than the smoothed extrusion. would it be possible to include an option to give the unprocessed field?
I'm trying to run comparisons between standing waves which are computed at the boundary, and once the smoothing has blended it all together its impossible to recover the information.
Also the native fem-accelerator 404s/403s on download. the JS fallback works but its slower.
also the wasm backend is broken apparently (TypeError : e is not a function) So everything silently lands on the slow JS path.
that will probably make more sense to others than it does for me, but that is apparently a limitation I've hit.
I'm trying to extend it to do torus knots!
So cool,
thanks so much.
I just downloaded the Mac, Win and Linux build files for the native accelerator, no problem.
I think I fixed the other bug. Let me know.
If you can describe what you want in detail, perhaps I can help.
Whaaaaouuuh! What a crazy tool bob, a thousand thank's!
I have a lot of ideas that I would imagine for it but I realise this represents a lot of work...
But the ability of combining mutltiple frequency onto multiple mediums would be amazing. Maybe being able to setup sources for the perturbations as well as an overall medium like air encompassing the whole grid, or water contained in a hollow cavity. Like the lattice of the singing bowl and the water it is containing. Or those egyptians vases filled with air and/or water. Thank you so much for that amazing tool!
Hi Stan, glad you like the beta. It had a lot of bugs. I've been working hard on those over the last couple of days, and I've just released a huge update, so check it out! Delete your cache and do a forced refresh. Enjoy!
I'll standing your wave!
This is what the AI is after.
What I need is the field before both steps. Specifically, per mode:
• the raw nodal displacement eigenvector (3 components per node), in whatever units the solver works in — no per-mode rescaling
• the node ordering or node positions, so values can be tied to locations
• ideally the post-cleanup occupancy, since the largest-connected-component and despeckle passes can change which voxels are actually in the body
A flag on the solve message — something like rawField: true — would be perfect. Frequencies are already fine; it’s only the fields.
Why: I’m computing mode overlap integrals (∫φᵢ·φⱼ over the body) to get nonlinear coupling coefficients between modes. That calculation needs true relative amplitudes between modes, and genuine zeros outside the boundary. The normalisation I can undo myself given the geometry — the smoothing I can’t, because once real boundary values are averaged with fabricated exterior ones the information’s gone. And the boundary is exactly where the physics I’m after lives.
So I basically need the non-smoothed boundary information for computing as opposed to 3D printing which it is optimised for - I believe?
https://youtube.com/shorts/RixEq4T3UbE?si=cpMgunRJ8ueNQt4C
Down Bob's lane
This is outstanding. Been thinking for a few weeks about doing this with electric and magnetic fields