The Perception of Band-Limited Decorrelation Between Vertically Oriented Loudspeakers

Abstract

Two experiments have been conducted to investigate the perceptual effect of band-limited interchannel decorrelation between vertically oriented loudspeakers. The perceived vertical image spread (VIS) and tonal quality (TQ) of phantom auditory images have been subjectively assessed in multiple comparison trials. The aim of the article was to find a lower decorrelation boundary that provides a significant increase of VIS, whilst maintaining TQ close to that of the original source. For test stimuli, decorrelation was applied to natural sound sources and pink noise in groups of octave-bands, where the lowest band was varied between 63 Hz and 8 kHz and the upper band was fixed at 16 kHz, resulting in eight decorrelated conditions for each source. Unprocessed octave-bands below the lower boundary were reproduced simultaneously through the lower main-layer loudspeaker only, and a monophonic main-layer only condition was also included in the comparison alongside the decorrelated stimuli. Results reveal that vertical decorrelation of the 500 Hz octave-band and above tends to significantly increase VIS, similar to that of broadband decorrelation, with little impact on TQ. In some cases, decorrelation of higher octave-bands and above can also produce similar increases of VIS with less impact on TQ, however, this is shown to be largely source-dependent. These results suggest that vertical decorrelation of lower frequencies has little perceptual benefit, and band-limiting vertical decorrelation to higher frequencies is likely to reduce low frequency phase cancellation. Applications of such an approach include 2D-to-3D upmixing and binaural audio rendering, with additional implications for 3D audio recording.

3D Microphone Array Comparison: Objective Measurements

“…On the other hand, the other vertical pairs of PCMA-3D (FL-FRh, FL-RLh, and FL-RRh) still had at least 1-m spacing between the microphones and therefore their ICCs were comparable to those of the other spaced main arrays in general. Gribben and Lee [61,62] found that in a nine-channel loudspeaker reproduction, the effect of vertical ICC on vertical image spread (VIS) was largely insignificant for low frequencies but significant for frequencies above about 1 kHz, albeit only slight. The current results show that the ICCs of the vertical pairs for all of the spaced arrays apart from PCMA-3D were very low (about 0.1 or below) for the middle and high frequency bands.…”

Allpass Feedback Delay Networks

“…For the opposite direction let us assume, there exists diagonal matrix P satisfying (23). Thus, A J is diagonally similar to pV {Dq -1 .…”

“…Both SISO and MIMO allpass FDNs were applied to a wide range of roles including: 1) increasing the echo density as preprocessing to an artificial reverberator [2,16]; 2) increasing echo density of in the feedback loop of reverberators [17][18][19][20]; 3) decorrelation for widening the auditory image of a sound source [21][22][23]; 4) as reverberator in electro-acoustic reverberation enhancement systems [15,18,24,25]; 5) linear dynamic range reduction [26,27] ; and 6) dispersive system design [28][29][30]. In the broader context of control theory, allpass FDNs are strongly related to Schur diagonal stability [31], e.g., stability properties of asynchronous networks.…”

“…Both SISO and MIMO allpass FDNs were applied to a wide range of roles including: 1) increasing the echo density as preprocessing to an artificial reverberator [2,17]; 2) increasing echo density of in the feedback loop of reverberators [18][19][20]; 3) decorrelation for widening the auditory image of a sound source [21][22][23]; 4) as reverberator in electro-acoustic reverberation enhancement systems [16,19,24,25]; 5) linear…”

Localisation of Vertical Auditory Phantom Image with Band-limited Reductions of Vertical Interchannel Crosstalk

“…For blanket reduction, the median thresholds that were derived in [9] were used (i.e., −9.5 dB for 0 ms and −7 dB for 1 ms). The 0 ms condition is equivalent to a vertical coincident microphone configuration [26], whereas the 1 ms condition represents the spacing between the main and height microphones being 0.34 m. This would be a practical spacing for a 3D microphone array, since, beyond this spacing, there would be no perceived increase in vertical image spread and the tonal quality would decrease [27]. In addition, the author's previous study on the VLT while using the same sound sources also tested the ICTD of 10 ms, but there was no significant difference found between 1 ms and 10 ms in the results.…”