arXiv subjects

Konstantinos Tsamis

Publications and source records attributed to Konstantinos Tsamis.

4 recordsLinked to original sources

arXiv2026

Seconds-Aligned PCA-DAC Latent Diffusion for Symbolic-to-Audio Drum Rendering

Symbolic-control drum generation requires preserving explicit event timing and dynamics while synthesizing acoustically plausible waveforms. We present Sec2Drum-DAC, a conditional latent-diffusion model for symbolic-to-audio drum rendering. The model conditions on event features sampled in physical time at codec-frame locations and predicts standardized principal-component coordinates of frozen DAC summed-codebook embeddings rather than waveform samples. In the evaluated DAC configuration, 72 principal components capture the observed training-frame summed-latent subspace under the stated SVD threshold, yielding a compact continuous denoising target with a deterministic reconstruction path to the 1024-dimensional DAC latent space before waveform decoding. Across 1,733 held-out four-beat windows, PCA diffusion improves paired spectral and transient metrics over deterministic PCA regression and a symbolic rendering baseline, while direct regression remains stronger on phase-sensitive waveform L1. Auxiliary RVQ cross-entropy improves short-step diffusion on mel error, onset-flux cosine, and waveform L1, with the most favorable trade-offs occurring at 6-25 denoising steps depending on the metric.

cs.SD↗

arXiv2026

Drum Synthesis from Expressive Drum Grids via Neural Audio Codecs

Generating realistic drum audio directly from symbolic representations is a challenging task at the intersection of music perception and machine learning. We propose a system that transforms an expressive drum grid, a time-aligned MIDI representation with microtiming and velocity information, into drum audio by predicting discrete codes of a neural audio codec. Our approach uses a Transformer-based model to map the drum grid input to a sequence of codec tokens, which are then converted to waveform audio via a pre-trained codec decoder. We experiment with multiple state-of-the-art neural codecs, namely EnCodec, DAC, and X-Codec, to assess how the choice of audio representation impacts the quality of the generated drums. The system is trained and evaluated on the Expanded Groove MIDI Dataset, E-GMD, a large collection of human drum performances with paired MIDI and audio. We evaluate the fidelity and musical alignment of the generated audio using objective metrics. Overall, our results establish codec-token prediction as an effective route for drum grid-to-audio generation and provide practical insights into selecting audio tokenizers for percussive synthesis.

cs.SD↗

arXiv2024

Filtral pretoposes and compact Hausdorff locales

The category of compact Hausdorff locales is a pretopos which is filtral, meaning that every object is covered by one whose subobject lattice is isomorphic to the lattice of filters of complemented elements. We show that any filtral pretopos satisfying some mild additional conditions can be embedded into the category of compact Hausdorff locales. This result is valid in the internal logic of any topos. Assuming the principle of weak excluded middle and the existence of copowers of the terminal object in the pretopos, the image of the embedding contains all spatial compact Hausdorff locales. The notion of filtrality was introduced by V. Marra and L. Reggio (Theory Appl. Categ., 2020) to characterise the category of compact Hausdorff spaces within the class of pretoposes. Our results can be regarded as a constructive extension of the aforementioned characterisation, avoiding reference to points. If the ambient logic is classical, i.e. it satisfies excluded middle, and the prime ideal theorem for Boolean algebras holds, we obtain as a corollary the characterisation of compact Hausdorff spaces in op. cit.

math.CT↗

arXiv2020

Regular and effective regular categories of locales

We examine the analogues for the respective categories of locales of two well-known results about regularity and effectiveness of some categories of spaces. We show that the category of compact regular locales is effective regular (=Barr-exact). We also show that the category of compactly generated Hausdorff locales is regular, provided that it is coreflective within Hausdorff locales. We do not appeal to the existence of points (which would render the two results trivial) but rely on the treatment of the subject by methods that are valid in the internal logic of a topos. On the course to the result about compactly generated locales we arrive at a generalization of a result of B. Day and R. Street, deriving regularity for a cocomplete category containing a dense regular subcategory closed under finite limits and colimits and satisfying a certain compatibility condition of pullbacks with appropriate colimits.

math.CT↗