RECORD
SpecSinGAN: Sound Effect Variation Synthesis Using Single-Image GANs
Sound · Submitted 14 Oct 2021
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim A listening study comparing our model to real recordings and to digital signal processing procedural audio models in terms of sound plausibility and variation revealed that SpecSinGAN is more plausible and varied than the procedural audio models considered, when using multi-channel spectrograms.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Single-image generative adversarial networks learn from the internal distribution of a single training example to generate variations of it, removing the need of a large dataset. In this paper we introduce SpecSinGAN, an unconditional generative architecture that takes a single one-shot sound effect (e.g., a footstep; a character jump) and produces novel variations of it, as if they were different takes from the same recording session. We explore the use of multi-channel spectrograms to train the model on the various layers that comprise a single sound effect. A listening study comparing our model to real recordings and to digital signal processing procedural audio models in terms of sound plausibility and variation revealed that SpecSinGAN is more plausible and varied than the procedural audio models considered, when using multi-channel spectrograms. Sound examples can be found at the project website: https://www.adrianbarahonarios.com/specsingan/
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Single-image generative adversarial networks learn from the internal distribution of a single training example to generate variations of it, removing the need of a large dataset. | background 1.00 | 0.06 | 0.04 | asserted 1.00 |
| s1 | In this paper we introduce SpecSinGAN, an unconditional generative architecture that takes a single one-shot sound effect (e.g., a footstep; a character jump) and produces novel variations of it, as if they were different takes from the same recording session. | contribution 0.98 | 0.88 | 0.08 | asserted 1.00 |
| s2 | We explore the use of multi-channel spectrograms to train the model on the various layers that comprise a single sound effect. | method 0.98 | 0.88 | 0.03 | asserted 0.89 |
| s3 | A listening study comparing our model to real recordings and to digital signal processing procedural audio models in terms of sound plausibility and variation revealed that SpecSinGAN is more plausible and varied than the procedural audio models considered, when using multi-channel spectrograms. | result 1.00 | 0.96 | 0.80 | asserted 1.00 |
| s4 | Sound examples can be found at the project website: https://www.adrianbarahonarios.com/specsingan/ | contribution 0.49 | 0.14 | 0.02 | asserted 0.55 |
What was examined, and what was not
- arXiv listing abstract v2
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Single-image generative adversarial networks learn from the internal distribution of a single training example to generate variations of it, removing the need of a large dataset. In this paper we introduce SpecSinGAN, an unconditional generative architecture that takes a single one-shot sound effect (e.g., a footstep; a character jump) and produces novel variations of it, as if they were different takes from the same recording session. We explore the use of multi-channel spectrograms to train the model on the various layers that comprise a single sound effect. A listening study comparing our model to real recordings and to digital signal processing procedural audio models in terms of sound plausibility and variation revealed that SpecSinGAN is more plausible and varied than the procedural audio models considered, when using multi-channel spectrograms. Sound examples can be found at the project website: https://www.adrianbarahonarios.com/specsingan/
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 1 sentence without a confident label
Uses the method or tool of 1 paper
-
We use WGAN-GP [ref] for the adversarial loss, in combination with a reconstruction loss with a weight of 10.
Improved Training of Wasserstein GANs · Architecture · 1.00
Cites as background 7 papers
-
There are multiple architectures and applications that have been explored, such as autoregressive models for waveform synthesis [ref] , GANs for unconditional waveform synthesis [ref] or for frequency-domain conditional synthesis [ref] [ref] or diffusion models for conditional and unconditional waveform synthesis [ref] .
WaveNet: A Generative Model for Raw Audio · Related Work · 0.99
-
There are multiple architectures and applications that have been explored, such as autoregressive models for waveform synthesis [ref] , GANs for unconditional waveform synthesis [ref] or for frequency-domain conditional synthesis [ref] [ref] or diffusion models for conditional and unconditional waveform synthesis [ref] .
Adversarial Audio Synthesis · Related Work · 0.99
-
There are multiple architectures and applications that have been explored, such as autoregressive models for waveform synthesis [ref] , GANs for unconditional waveform synthesis [ref] or for frequency-domain conditional synthesis [ref] [ref] or diffusion models for conditional and unconditional waveform synthesis [ref] .
DrumGAN: Synthesis of Drum Sounds With Timbral Feature Conditioning Using Generative Adversarial Networks · Related Work · 0.99
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.