Download DAFx Challenge Introduction & Results The 1st DAFx Parameter Estimation Challenge is an open initiative to advance the state of the art in parameter estimation for acoustic modeling. Stated as a system identification problem, this first edition focuses on plate reverberation—an archetypal dense, modal and weakly damped acoustic system. Participants tackled two tasks: (A) estimating the physical parameters of a vibrating plate from its impulse response, and (B) recovering the modal parameters of the same system. Both rest on a simulation framework based on the damped Kirchhoff–Love plate equation, and both are posed and scored entirely on synthetic data produced by that framework: no measurement of a real plate is involved. Two participants solved Task A down to machine precision by different strategies: one a neural network trained on a very large dataset, and one gradient-free optimization with many inexpensive evaluations. Task B proved considerably harder: the best submission attains a relative error of 0.33 on a [0, 2] scale, and every method recovers modal frequencies and decay rates far more accurately than modal gains. A complementary frequency-domain evaluation reorders the ranking and exposes a systematic gain bias to which the per-mode metric is blind.
Download Evaluating Tokenization Strategies for Expressive Classical Piano Performance Generation Expressive piano performance generation needs symbolic pitch, timing, and dynamics. We evaluate six tokenization strategies for a Transformer that generates classical piano performances. Our tokenizations add velocity, beat annotations, and sustain pedal, from note-only to full representations. We pretrain on MAESTRO, then finetune on ASAP with beat-level annotations. The model uses anticipatory-style note encoding with cross-attention on composer and genre. FAD on the ASAP test set shows that note + velocity + pedal and full modes achieve the lowest mean FAD (1.76 and 1.97). Both beat the note-only baseline (3.10). Beat tokens show mixed, category-dependent effects and do not improve the best modes on average.
Download PAEDB: A Synthetic Primary-Ambient Dataset Generation Pipeline for Automatic Upmixing Using Deep Neural Networks Automatic blind upmixing aims to convert audio from a smaller channel format (e.g. mono or stereo) into a multichannel format using estimates of direct and diffuse spatial statistics within the signal. Current approaches rely on primary-ambient extraction (PAE) algorithms, which lack real-world context through limited processing windows. Deep learning music source separation (MSS) models have been applied in voice-primary-ambient extraction (VPA) upmixing systems for handling direct components, but still rely on DSP methods of surround channel generation. This work further investigates utilizing source separation within VPA upmixing, focusing specifically on the task of stereo decorrelation and ambience extraction for 5.1 surround. We also release PAEDB (Primary–Ambient Extraction Dataset), a high-quality music dataset derived from MUSDB18-HQ and MoisesDB, comprising 1,809 primary–ambient stem pairs totaling over 550 hours of audio. The performance of selected DNNs trained on PAEDB is then evaluated using signal metrics and a listening study. Our findings indicate that DNNs can effectively model the behavior of PAE algorithms, establishing PAEDB as a strong foundation for ML upmixing systems and underscoring the need for higher-quality multichannel data to advance beyond conventional methods.
Download ALAMODE: Automated Learning of Acoustical Modal Parameters via Differential Evolution This paper is a technical report on the methodology submitted for Task A of the 1st DAFx Parameter Estimation Challenge. The goal of the challenge’s task is to invert the multi-dimensional physical and geometric parameters of a virtual plate reverberator given a target reference impulse response. To achieve this, we present a multi-stage gradient-free optimization framework. This three-stage optimization is computed using an efficient physics-based simulator, starting with an optimization of only mode frequency-determining physical parameters, followed by a 6-DoF parameter optimization with position-determining ones and a final phase for frequency- and position-independent mode amplitude estimation.
Download Band-Count Dense Modal Estimation with Fixed-Frequency Differentiable Resonator Refinement ★ Task B of the 1st DAFx Parameter Estimation Challenge requires estimating the frequencies, decay rates, gains, and number of modes in a dense plate-reverb impulse response. Weak and overlapping modes make sparse peak detection prone to severe undercounting. We train an ExtraTrees regressor on simulator-generated data to predict mode counts in four frequency bands. These counts define dense frequency grids, after which a differentiable all-pole resonator model refines decay and gain while keeping frequency fixed. On two separate synthetic validation sets, the system reduces a local challenge-style error by about 66% relative to the official default peak-picking baseline. The improvement is mainly associated with lower mode-count mismatch, while decay and gain remain the largest error sources. These findings support separating modal-density estimation from continuous parameter fitting.
Download WaveNet-Style Guitar Amplifier Model Pruning for Real-Time iOS Deployment WaveNet-style convolutional networks emulate tube amplifiers and distortion pedals with high fidelity, but their computational cost has confined them to desktops or dedicated DSP hardware. We present a sparse-enabled WaveNet inference engine for iOS that runs heavily pruned neural guitar amplifier models in real time on iPhones. Aggressive iterative magnitude pruning removes 90% of the network weights with no perceptible loss in quality. A custom sparse C++ engine turns this sparsity directly into compute savings, sustaining low-latency real-time operation on a CPU-only iPhone implementation where the dense model cannot. On-device output matches the trained model to within int16 quantization error. At the demonstration, visitors will play a guitar through the app on iPhone hardware and A/B the on-device pruned model against the physical pedal it emulates. Source code and audio examples are available online.
Download CLEAN2FX: Label-Conditioned Modeling for Clean-to-Effect Guitar Audio Transformations We present Clean2FX, a study and demo of label-conditioned clean-to-effect transformation for electric guitar audio. Given a clean guitar input and a target effect label, the task is to synthesize the corresponding effected signal while preserving the musical content. Training and evaluation pairs are constructed from EGFxSet real, single-tone recordings by assembling matched clean/effected chords, melodies, and mixed timelines. This allows for controlled comparison across effects. We evaluate four neural approaches under a common spectrogram-based transformation setting: two variational autoencoders and two U-Net models that differ in whether they operate on linear or log-magnitude representations. Performance is measured using linear-magnitude spectrogram MSE and Fréchet Audio Distance. The U-Net models outperform the variational autoencoder variants. Per-effect results show that distortion effects are most readily improved, whereas delay and reverb effects exhibit weaker FAD gains despite substantial spectral-error reductions. A conditioning-sensitivity diagnostic provides evidence that the best model responds to target labels rather than collapsing to a single transformation. Our demo website compares two models applied on real-world guitar performances outside training and validation data, providing audio and spectrogram examples of the practical clean-to-effect behavior.
Download FDN Sandbox: Real-Time Experimentation and Analysis of FDNs This work presents sfFDN, a modular and real-time-capable C++ library for Feedback Delay Networks (FDNs), together with the companion FDN Sandbox application designed for interactive experimentation, analysis, and parameter optimization. The library implements the canonical FDN as well as several recent extensions, including filter feedback matrices, velvet-noise decorrelation filters, and two-stage graphic equalizers for attenuation and tone correction. The Sandbox application exposes these features through a graphical interface, providing a suite of real-time visualizations, as well as an optimization framework supporting nine algorithms from the ensmallen library, with built-in loss functions for both colorless reverberation and room impulse response matching. Both the library and the application are open-source.
Download Pulsetable Synthesis of Wind Instrument Tones We revisit pulsetable synthesis, an efficient technique for generating plausible and expressive wind instrument tones. Based on the principles of pulse forming theory, this method models sound production as the periodic repetition of shaped pulses characterizing the target instruments' spectral envelope. In this approach, single-cycle waveforms, referred to as pulses, are stored in pulsetables indexed by their corresponding fundamental frequency. During synthesis, the pulses are read from these tables to form a periodic waveform, which is further shaped by time-varying low-pass filtering, amplification, and reverberation. These processes are guided by control signal contours that describe how fundamental frequency, brightness, and loudness evolve over time. Through case studies with real-world wind instrument recordings, we show how the interplay between these control signals gives rise to articulations such as attack transients, vibrato, and growl. Finally, we discuss the potential of this framework for integration into Differentiable Digital Signal Processing (DDSP) models, where neural networks could learn synthesis parameters directly from training data.
Download Adapting Diffusion-Based Music Synthesis to Speech and Singing Voice Conversion Recent diffusion-based generative models have achieved strong results in domain-specific audio generation tasks such as speech, singing, and instrumental music synthesis. However, these models are typically specialized and do not generalize well to mixed or intermediate audio types. In this work, we adapt a diffusion-based model originally designed for multi-instrument music synthesis to voice conversion, covering both speech and singing within a unified framework. Specifically, we extend musical note-based conditioning to include phonetic posteriorgrams (PPGs) and pitch contours, and reinterpret timbre conditioning as speaker or singer identity via feature-wise linear modulation. Experiments show that the adapted model matches or surpasses a dedicated voice conversion system in terms of naturalness and performer similarity, while maintaining accurate pitch control across speech and singing. At the same time, we observe limitations in phonetic fidelity and a degradation in vocal quality when incorporating instrumental training data. Furthermore, we demonstrate that off-the-shelf feature extractors provide effective conditioning signals, enabling large-scale self-supervised training without manual annotations. These results highlight the potential of cross-domain model transfer towards unified audio generation systems capable of handling speech, singing, and music.