Researchers Demonstrate Acoustic Cloning of Scattering Objects Using Digital Twins
Physicists have experimentally demonstrated a method to create digital clones of acoustic scattering objects through a two-step process involving broadband illumination and holographic reconstruction. The technique uses multidimensional deconvolution to extract an object's scattering Green's functions from recorded reverberative data, then reconstructs a hologram that scatters sound waves identically to the original. The advancement could enable realistic digital acoustic models and more efficient metamaterial experimentation.
Researchers have successfully cloned acoustic scattering objects by first acquiring their digital twins and then reconstructing them holographically. The process begins with broadband speakers illuminating the target object within a closed receiver aperture; the recorded reverberative data is then processed using multidimensional deconvolution to extract the object's scattering Green's functions. In the second step, these functions are used to create a holographic reconstruction that scatters any acoustic wavefield in real-time exactly as the original object would. The system employs low-latency feedback to reproduce all orders of interactions between the physical wavefield and the numerically defined hologram. The researchers demonstrated the technique on several rigid scatterers in a two-dimensional acoustic waveguide and showed the ability to modify cloned objects. Potential applications include creating fully realistic digital scattering models and conducting more efficient metamaterial experiments.
Limitations & open questions
The study does not discuss limitations of the current approach, such as frequency bandwidth constraints, scalability to three-dimensional objects, computational requirements, or practical challenges in real-world applications beyond the controlled laboratory waveguide setting.
What different sources said
- arXiv physicsCenter
Acoustic Cloning
Related
New AI Method and Dataset for Rapid Post-Earthquake Building Damage Assessment
Researchers have developed MSI-Net, a deep learning method for detecting building damage from satellite imagery after earthquakes, along with a new dataset (TUE-CD) created from Turkey earthquake data. The approach addresses challenges posed by short imaging intervals and varying camera angles in post-disaster remote sensing. This technology could accelerate emergency response and damage assessment in the critical hours and days following major earthquakes.
Study Examines How Aesthetic Perspectives Shape Information Systems Research
A hermeneutic analysis of Information Systems scholarship identifies four foundational aesthetic perspectives that influence what researchers recognize as worthy of study. These perspectives—imitation, sensory experience, world-making, and political doing—form an epistemic infrastructure that shapes research horizons and theoretical frameworks. The findings suggest that making aesthetic assumptions explicit can reveal overlooked dimensions in areas like algorithmic management and digital intimacy.
Study Identifies Critical Branching Dynamics in Small LSTM Neural Networks
Researchers analyzing trained LSTM networks found that small models near optimal training exhibit scale-free avalanche statistics and near-critical branching parameters, while larger models remain subcritical. The study proposes that criticality, a key principle in biological neural systems, emerges as a capacity-dependent dynamical regime in artificial networks. The findings suggest a novel connection between network size, training dynamics, and the emergence of critical-like behavior in deep learning systems.