Search

We present initial results from measurements of exhaled droplets by two female singers during singing, speaking, laughing, and recitation of poetry. We also conducted measurements with a flute, a clarinet, a tuba, and only the tuba's mouthpiece. This small project is motivated by the current distancing restrictions for choirs, musicians, and actors. To be able to image and track droplets, we have developed and built a portable measurement set-up. We have detected droplets with diameters of approximately 50 μm and above. We found that for a single subject, the largest amount of droplets are produced during singing, but that laughing can produce a comparable number of droplets. Speaking and recitation of poetry do not produce as many droplets. We repeated exercises at varying distances and found that the number of detected droplets decreased rapidly with increasing distance. However, the relative height difference between people, e.g. in a stage setting, must be considered when determining how far a droplet of a given size and velocity can reach. Most droplets we detected follow a ballistic trajectory, and hit the ground after a distance of approximately 1 m. In case of the musical instruments, we did not detect droplets exiting a tuba and only few single droplets from a flute and a clarinet. A clarinet usually points downwards, such that droplets follow a downward trajectory and hit the ground after a short distance. The situation was different when the tubist used only the mouthpiece (that is, detached from the tuba) during warm-up while partially blocking the opening with a finger: in this case, many droplets were ejected from the mouthpiece. These followed ballistic trajectories at different speeds, with some droplets exiting the measurement volume. Two important limitations of this study are that (i) the sample size is very small, such that a statistical analysis is beyond our scope; and (ii) that we cannot draw conclusions about a possible risk of infection with COVID-19 when performing any of the said activities. Norges Korforbund

Presented at the International Conference on Audio, Language and Image Processing (ICALIP), 7-9 July 2008, Shanghai, China. non-peer-reviewed

Non-peer-reviewed

peer-reviewed An optical fibre based sensor has been developed to aid the quality assurance of food cooked in industrial ovens by monitoring the product in situ as it cooks. The sensor measures the product colour as it cooks by examining the reflected visible light from the surface as well as the core of the product. This paper examines the use of the sensor for the detection of blood in the spinal area of cooked whole chickens. The results presented here show that the sensor can be successfully used for this purpose. PUBLISHED non-peer-reviewed

Water holding capacity (WHC) is among the most important quality traits of meat. However, the mechanism behind WHC continues to be poorly understood and online measurement has yet to be achieved in the meat industry. The overall objectives of this thesis were to advance the understanding of WHC in meat (specifically the changes of different water domains during drip production process) using nuclear magnetic resonance (NMR) proton T2 relaxometry, and to investigate the suitability of NMR as a reference method for faster, online spectroscopic methods to measure WHC. Visible/near infrared spectroscopy (Vis/NIR spectroscopy) and X-ray spectroscopy were investigated as potential online methods. NMR was compared with the traditional reference method for WHC in pork, EZ-DripLoss method, and the error magnitudes and sources were discussed. NMR was also investigated as a method to predict purge in vacuum packages in early post mortem (p.m.). The results show that NMR is an accurate and quantitative method for measuring small changes of water content in a controlled system (H2O and D2O mixtures). NMR can also separate the T2 values well in another controlled system (CuSO4 solutions of varying concentrations) consists of three different known T2 values, that are similar to the T2 values found in meat. The complexity of sample inhomogeneity and sample handling introduces errors in NMR measurements of meat, and standardized procedures need to be considered. Regarding meat samples, the three, decomposed spin-spin relaxation time components corresponding to water domains of different mobility were seen to change during drip production. For relatively shorter periods of dripping (45 hours), the migration of water and larger molecules from the meat to the drip domain was irreversible, and governed by molecules from the slowest relaxation domain (contains free water). In addition, it was found that NMR T2 relaxometry could be considered as an improved reference method for spectroscopic techniques when compared with EZ-DripLoss method to measure WHC, i.e. the relaxation time of the slowest relaxation component (T22) correlated better with both Vis/NIR and X-ray spectra than EZ-DripLoss values. Two different approaches of model fitting (discrete and continuous model) were applied to the NMR data and directly compared. It was found that the two fitting methods gave different results for both relaxation rates and intensity for all three components, which might cause different interpretation of water activity in meat during drip. Finally, the ability of predicting purge from pork muscle after 9-day vacuumpacked storage using NMR parameters measured early post mortem was explored. Results show that NMR had limited prediction ability. This was investigated further and it was found that it could be an effect of muscle structural changes during storage, which affect WHC, but could also be due to the substantial errors in NMR and purge measurements relative to the variation in purge. In summary, it was concluded that NMR proton relaxometry is a very informative method for WHC measurement. However, careful and standardized sample handling is required, and errors caused by this issue should be further assessed. Furthermore, it seems that there is a need for NMR instrument can be adapted for WHC measurement in meat samples, with a larger sample holder size than the common ~2.8 gram for intact meat measurement. This can increase robustness towards sample inhomogeneity and reduce sampling errors. Kjøttets evne til å binde vann er blant dets viktigste kvalitetsegenskaper. Mekanismen for vannbinding er langt fra forstått og online målinger er fortsatt en drøm for kjøttbransjen. De overordnede målene for denne avhandlingen var å fremme forståelsen av vannbindingsevnen (VBE) til kjøtt (spesielt endringer i domener av vann under dannelse av drypp) ved hjelp av kjernemagnetisk resonans (NMR) proton T2 relaksasjon, samt å undersøke NMR sin egnethet som referansemetode for å gi for raskere, online spektroskopiske metoder for måling av VBE. Synlig/nær infrarød og røntgen spektroskopi ble undersøkt som potensielle online metoder. NMR ble sammenlignet med den tradisjonelle EZ-DripLoss metoden med tanke på at denne kunne bli en ny referansemetode for vannbindingsevnen til svinekjøtt. Størrelse og kilder til målefeil ble diskutert. NMR sin evne til å forutsi drypp i lagrede vakuum pakninger ble undersøkt tidlig post mortem (p.m.). Resultatene viste at NMR var en nøyaktig og kvantitativ metode for å måle små forandringer i vanninnhold i enkle system (H2O og blandinger D2O). NMR gir tre ulike T2-verdier i CuSO4 løsninger av varierende konsentrasjoner, og disse tre T2-verdiene ligner på de tre som finnes i kjøtt. Reproduserbarheten til NMR-målinger av kjøttprøver kompliseres av prøvenes inhomogenitet og av prøvehåndteringen, og standardiserte måleprosedyrer er derfor nødvendig. Når det gjelder kjøttprøver, så gjennomgår de tre spin-spin relaksasjonskomponentene endringer i tidsrommet hvor dryppet produseres. Etter relativt korte perioder med dryppdannelse (45 timer), vil migrasjon av vann og større molekyler fra kjøttet i drypp domene være irreversible, og styres av molekyler fra det langsomste relaksasjons domenet (inneholder fritt vann). NMR-T2 relaksasjon ansees som en forbedret referansemetode for spektroskopiske teknikker relativt til EZ-DripLoss metodens mål for vannbindingsevne, dvs. relaksasjonstiden til den tregeste relaksasjonskomponenten (T22) er korrelert bedre med synlig lys / NIR og røntgen spektra enn til EZ-DripLoss målinger. To ulike tilnærminger til modellbygging (diskret og kontinuerlig modell) ble sammenlignet, og resultatene indikerte at de ulike modellene ga forskjellige resultater både for relaksasjons hastigheter og intensitet for alle tre komponenter for vann. Valg av modell kan således gi ulike tolkninger av vannamobiliteten i kjøttet under dryppdannelsen. Til slutt, ble evnen til å predikere drypp fra svine muskel etter 9 dagers lagring i vakuumpakning forsøkt predikert fra NMR målinger som ble utført tidlig post mortem. Resultatene viste at NMR hadde begrenset evne til å predikere drypp 8 dager fram i tid. Dette kan være forårsaket av strukturelle endringer i kjøttet som påvirker vannbindingsevnen under lagring, men også de betydelige feil i NMR og drypp-målinger som eksisterer i forhold til variasjonen i drypp. NMR proton relaksasjon er en informativ metode for vannets status i kjøtt. Imidlertid må man være forsiktig ved prøvehåndtering, og feil forårsaket av dette problemet bør vurderes nærmere. Det ser ut til at det er et behov for NMR-instrument med en større prøveholder enn den som vanligvis bruker ~2,8 gram intakt kjøtt. Kjøttprøvenes inhomogenitet og utvalgsfeil kan da reduseres og NMR-metoden tilpasses måling og prediksjon av vannbindingsevnen til kjøttprøver. Norges forskningsråd ; Nortura SA