 Hello everyone, my name is Chen and welcome to our presentation. Today I'm going to present our recent study about smartphone-based hearing tests for virtual hearing clinic influence of ambient noise on the absolute threshold and longer scaling at home. As you all know, any home environment typically you don't have a sound attenuated boost and it might have the ambient noise and of course the ambient noise might influence the results of the measurements for example it might directly masking the test stimulus and on the other hand it can also distract the participants and make some boosts their attention. Therefore we might want to avoid the negative influence of the ambient noise. So from the earlier studies we know that we can measure, monitor and control it as the noise levels and to eliminate the influence we propose our first research question is it visible to perform smartphone-based hearing tests remotely at participants' homes if the background noise is considerably low. And the second research question would be for the loudness scaling test would be the influence be smaller at a higher loudness category. Here is our experimental design. We designed repeated measurements in 15 subjects. So the subjects are required to repeatedly do the measurements in two different environments. One is the inputs environment as a reference and second is the home environment. While in the home environment the subjects were required to check and record the real-time noise level using the decibel X app. So as you can see here this is the user interface of the app and the red you can just read the values here for the real-time level and you can also record it. And the results after that also another point is that the subjects were needed to ensure that the values were lower than our recommended limits. So to make sure that the environment is over silent. The results measured in two different environments were then compared and in our we expect that the difference is to be small and to be clinical tolerant. And here are the results of the pure-time geometry. We mainly plotted the Bart Altman plots, a statistic tool to assess the differences. So we have here the plot here. We have different shapes to indicate different frequencies. We have two or three different frequencies ranging from 0.25 to four kilohertz. And the Y axis differences between the HDR and the Harris result. The Harris result measured in two different environments in both environment and home environment. And then the X axis we take the main values of the two of the Harris result at two different environments in DBHL. The red dash lines here indicates the upper and the lower limits. So they were calculated as 1.96 standard deviation. And also finally we would provide a statistic measures, root miscal error correlation coefficient, bias and root miscal error. And the black line here indicates the line of bias. And the shadow there is the 95 percentage of confidence interval. So from this plot we can clearly know that first of all there was no significant differences between the two environments for the pure-time geometry. It tells us that we can simply use the environment. We can simply do it in a home environment and it can be it have equivalent results to the in-boost environments. And also the most of the points line within the upper limits and lower limits which means that again validated that the environments are replaceable to each other. And here are the results of the category column scaling. Again we use the same statistic plots about hashman plots. And we use the different shapes of points to indicate different frequencies. At three different frequencies. And for low-nose category we use different colors to represent different low-nose categories. For example the red indicate not heard and the blue indicates the too loud. The y-axis is the medium levels assigned to each category units into different environments. So for the in-boost environment for the at-home environment. And the x-axis is the main values of the two outcome measures. Again same as before we have the two different red dash lines to indicate the upper and lower limits. They are equal to 1.96 standard divisions. And also we also plot a line of BIAS with black color here and the shadow error indicates the 95% of confident interval. Also same as before we have provided the statistic tables correlation coefficient BIAS and root mean square arrows. Also for as you can see most of points lines within the upper and lower limits it tells us that there's no significant differences between the two environments for category of low-nose scaling. Which means that we can simply use the remote at home measurements in a remote and remote at home smartphone measurements and they have equivalent results to the standard in-boost environment in-boost measurement. And if we look at the R values it has a higher value at high category and a relatively lower value at a lower category. And the BIAS decrease with increasing the low-nose category for the root mean square arrow values so it has a relatively high low values at high categories while at high values at low category. So in conclusion the smartphone based at home automatic measurements with ambient noise monitoring are comparable to the in-boost measurements so in future we want to include hearing powered listeners and to examine the destruction effects. Thank you very much for your attention. You are welcome to propose your questions.