Refine
Year of publication
- 2021 (66) (remove)
Document Type
- Article (31)
- Part of a Book (16)
- Conference Proceeding (7)
- Article trade magazine (4)
- Book (3)
- Collection (3)
- Doctoral Thesis (1)
- Workingpaper / Report (1)
Has Fulltext
- no (66) (remove)
Keywords
- Forena (5)
- BestHeatNet (4)
- Soziale Arbeit (2)
- Aerospace Engineering (1)
- Alertness (1)
- Altern (1)
- Antriebe (1)
- Applied ethics (1)
- Arbeitswelt (1)
- Auktionsverfahren (1)
Department/institution
- Fachbereich - Maschinenbau und Verfahrenstechnik (25)
- Fachbereich - Sozial- & Kulturwissenschaften (21)
- Fachbereich - Medien (7)
- Fachbereich - Wirtschaftswissenschaften (6)
- Fachbereich - Elektro- & Informationstechnik (3)
- Sound and Vibration Engineering (3)
- Creative Media Production and Entertainment Computing (2)
- Fachbereich - Architektur (1)
Musicians and music professionals are often considered to be expert listeners for listening tests on room acoustics. However, these tests often target acoustic parameters other than those typically relevant in music such as pitch, rhythm, amplitude, or timbre. To assess the expertise in perceiving and understanding room acoustical phenomena, a listening test battery was constructed to measure the perceptual sensitivity and cognitive abilities in the identification of rooms with different reverberation times and different spectral envelopes. Performance in these tests was related to data from the Goldsmiths Musical Sophistication Index, self-reported previous experience in music recording and acoustics, and academic knowledge on acoustics. The data from 102 participants show that sensory and cognitive abilities are both correlated significantly with musical training, analytic listening skills, recording experience, and academic knowledge on acoustics, whereas general interest in and engagement with music do not show any significant correlations. The regression models, using only significantly correlated criteria of musicality and professional expertise, explain only small to moderate amounts (11%–28%) of the variance in the “room acoustic listening expertise” across the different tasks of the battery. Thus, the results suggest that the traditional criteria for selecting expert listeners in room acoustics are only weak predictors of their actual performances.
Differentiable Compound Optics and Processing Pipeline Optimization for End-to-end Camera Design
(2021)
Most modern commodity imaging systems we use directly for photography‐-or indirectly rely on for downstream applications‐-employ optical systems of multiple lenses that must balance deviations from perfect optics, manufacturing constraints, tolerances, cost, and footprint. Although optical designs often have complex interactions with downstream image processing or analysis tasks, today’s compound optics are designed in isolation from these interactions. Existing optical design tools aim to minimize optical aberrations, such as deviations from Gauss’ linear model of optics, instead of application-specific losses, precluding joint optimization with hardware image signal processing (ISP) and highly parameterized neural network processing. In this article, we propose an optimization method for compound optics that lifts these limitations. We optimize entire lens systems jointly with hardware and software image processing pipelines, downstream neural network processing, and application-specific end-to-end losses. To this end, we propose a learned, differentiable forward model for compound optics and an alternating proximal optimization method that handles function compositions with highly varying parameter dimensions for optics, hardware ISP, and neural nets. Our method integrates seamlessly atop existing optical design tools, such as Zemax. We can thus assess our method across many camera system designs and end-to-end applications. We validate our approach in an automotive camera optics setting‐-together with hardware ISP post processing and detection‐-outperforming classical optics designs for automotive object detection and traffic light state detection. For human viewing tasks, we optimize optics and processing pipelines for dynamic outdoor scenarios and dynamic low-light imaging. We outperform existing compartmentalized design or fine-tuning methods qualitatively and quantitatively, across all domain-specific applications tested.