Vesa Välimäki, Jussi Rämö and Fabian Esqueda
Companion page for a paper in the 21st Conference on Digital Audio Effects (DAFx-18)
Aveiro, Portugal, 4–8 September, 2018
The article can be downloaded here.
This paper proposes signal processing methods to extend a stationary part of an audio signal endlessly. A frequent occasion is that there is not enough audio material to build a synthesizer, but an example sound must be extended or modified for more variability. Filtering of a white noise signal with a filter designed based on high-order linear prediction or concatenation of the example signal can produce convincing arbitrarily long sounds, such as ambient noise or musical tones, and can be interpreted as a spectral freeze technique without looping. It is shown that the random input signal will pump energy to the narrow resonances of the filter so that lively and realistic variations in the sound are generated. For real-time implementation, this paper proposes to replace white noise with velvet noise, as this reduces the number of operations by 90% or more, with respect to standard convolution, without affecting the sound quality, or by FFT convolution, which can be simplified to the randomization of spectral phase and only taking the inverse FFT. Examples of producing endless airplane cabin noise and piano tones based on a short example recording are studied. The proposed methods lead to a new way to generate audio material for music, films, and gaming.
Airplane noise outside the aircraft. Original sound sample from freesound.org.
The synthesized sounds are created based on a 1 second clip of the original airplane sample (Original).
Original sound sample from freesound.org.
The synthesized sounds are created based on a short clip of the original piano sample (Original).
The idea is to generate an extended signal using the inverse FFT from the original magnitude and a random phase spectrum. The IFFT-based method produces a long segment of the output signal at one time. Furthermore, the segment produced by the IFFT method can be repeated by concatenating copies of itself without the need of windowing or crossfading. This property comes from the fact that the fast convolution, which is the basis of the proposed IFFT synthesis method, implements a circular convolution in the time domain.Airplane noise:
Video illustrating the click-free looping of exended sounds created with the IFFT method. Notice that the time axis is circular.