Academia.eduAcademia.edu

Figure 1 – uploaded by Andrea Primavera

See full PDFdownloadDownload figure
Evaluation of the MSE between dynamic convolution and PCA based approach as a function of the number of PC. e Distortion stomp-box (Boss DS-2 turbo distortion) e Compressor (dbx 160A) e Bass pre-amplifier (Aguilar DB680 tube bass pre-amt

Table 1 Evaluation of the MSE between dynamic convolution and PCA based approach as a function of the number of PC. e Distortion stomp-box (Boss DS-2 turbo distortion) e Compressor (dbx 160A) e Bass pre-amplifier (Aguilar DB680 tube bass pre-amt

Related Figures (28)

Advanced Algorithm for Audio Quality Improvement in Musical Keyboards Instrume
Advanced Algorithm for Audio Quality Improvement in Musical Keyboards Instrume
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
Overall scheme of proposed NUPOLS for DSP implementation. An efficient DSP-based implementation of a fast convolution algorithm for rooms acoustic simulation has been proposed for OMAPL137 [9].
UPOLS workload (i). NUPOLS workload with 3 different partitionings (Ki = 64 Kz = 512 (ii), Ky, = 64 Kz = 1024 (iii), Ky = 64 Kz = 2048 (iv)). Max (a) and mean (b) workload for classic implementation. Max (c) and mean (d) workload using psychoacoustic approach.
UPOLS workload (i). NUPOLS workload with 3 different partitionings (Ki = 64 Kz = 512 (ii), Ky, = 64 Kz = 1024 (iii), Ky = 64 Kz = 2048 (iv)). Max (a) and mean (b) workload for classic implementation. Max (c) and mean (d) workload using psychoacoustic approach.
NUPOLS workload as a function of the processing cycle (IR Length=3.164 sec). (a) Global workload: NUPOLS. (b) Partial workload: small UPOLS (Ki = 64). (c) Partial workload: large UPOLS (Kz = 2048).
NUPOLS workload as a function of the processing cycle (IR Length=3.164 sec). (a) Global workload: NUPOLS. (b) Partial workload: small UPOLS (Ki = 64). (c) Partial workload: large UPOLS (Kz = 2048).
Comparison between ATH and each different EDR curve.
Comparison between ATH and each different EDR curve.
Number of complex multiplications as a function of the framesize. (a) Algorithm without psychoacoustic. (b) Approach based on EDR [16]. (c) Approach based on EDR+ATH [19]. Several tests have been carried out to evaluate the effectiveness of the approach [19] through subjective and objective evaluations considering 3 different IRs.
Number of complex multiplications as a function of the framesize. (a) Algorithm without psychoacoustic. (b) Approach based on EDR [16]. (c) Approach based on EDR+ATH [19]. Several tests have been carried out to evaluate the effectiveness of the approach [19] through subjective and objective evaluations considering 3 different IRs.
Scala EDR. Real IR (iv), Freeverb IR (v), Jot IR (vi).
Scala EDR. Real IR (iv), Freeverb IR (v), Jot IR (vi).
Workload required to reproduce the artificial effect using hybrid reverberators and fast convolution.
Workload required to reproduce the artificial effect using hybrid reverberators and fast convolution.
Block diagram of the proposed algorithm.
Block diagram of the proposed algorithm.
Workload as a function of the IR length. (a) Dynamic convolution, (b) proposed approach using 1 PC, (c) using 2 PCs, (d) using 3 PCs. An efficient DSP-based implementations of dynamic convolution has been f posed for OMAPL137 [8].
Workload as a function of the IR length. (a) Dynamic convolution, (b) proposed approach using 1 PC, (c) using 2 PCs, (d) using 3 PCs. An efficient DSP-based implementations of dynamic convolution has been f posed for OMAPL137 [8].
Proposed adaptive system identification.
Proposed adaptive system identification.
Overall scheme of the proposed morphing algorithm
Overall scheme of the proposed morphing algorithm
Case study: Morphing between a classic and electronic snare. Algorithms: Linear Interpolation (LI), proposed approach (LIP), additive syn- thesis based (SMS).
Case study: Morphing between a classic and electronic snare. Algorithms: Linear Interpolation (LI), proposed approach (LIP), additive syn- thesis based (SMS).
Proposed approach [10] both in time and frequency domain. A mixed-phase time-frequency approach has been presented in [10]:
Proposed approach [10] both in time and frequency domain. A mixed-phase time-frequency approach has been presented in [10]:
Magnitude response of measured IRs and
Magnitude response of measured IRs and
Overall scheme of the proposed approach. Based on similar procedures to iteratively estimate the impulse responses and to generate the equalization filters, they differ in the application domain. The former [12] is performed in the frequency domain, while the latter [15] works in the warped domain.
Overall scheme of the proposed approach. Based on similar procedures to iteratively estimate the impulse responses and to generate the equalization filters, they differ in the application domain. The former [12] is performed in the frequency domain, while the latter [15] works in the warped domain.
Scheme for the impulse response identification.
Scheme for the impulse response identification.
Magnitude response after equalization [15] Several tests have been conducted considering IRs recorded in different rooms.
Magnitude response after equalization [15] Several tests have been conducted considering IRs recorded in different rooms.
Magnitude response after equalization [12].
Magnitude response after equalization [12].
Related topics:
AcousticsMusicComputer MusicAudio Signal ProcessingKeyboard InstrumentsEmbedded SystemsRoom AcousticsAudio EffectsVirtual AnalogUniversity-Industry Collaboration
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved