Almost all smartphone manufacturers use microelectromechanical systems (MEMS) microphones in their devices. MEMS microphones typically have a sensitivity between 5 and 17.8 mV/Pa and can capture signals as low as 30 dB SPL and as high as 120 to 130 dB SPL (signal-to-noise ratio >60 dB). MEMS microphones also have a flat frequency response similar to ceramic and condenser microphones used in type 2 noise dosimeters. With the introduction of the iOS 6 operating system in late 2012, Apple allowed developers to bypass the high-pass filter that degraded the quality of acoustical measurements on older iPhones. This development also allows users of Apple smartphones to connect external microphones through the headset input jack.
I wonder what iOS 7 changed, and what the upcoming iOS 8 will allow.
The Android-based apps did not have features and functionality similar to the iOS apps. This is likely due to the development ecosystem of the Android marketplace and users’ expectations for free or low priced apps. A comprehensive testing procedure could not be carried out to show conclusive evidence of differences, since not all apps shared features and metrics that met our selection criteria. The limited testing showed a wide variance between the same app measurements on different devices. This can likely be attributed to the fact that Android devices are built by several different manufacturers and that there is a lack of conformity for using similar microphones and other audio components in their devices.
This is not surprising. The Android ecosystem is highly fragmented, and any app development would have to take into account all of the different devices available. Plus the fact that the version of the Android OS is different depending on what the smartphone manufacturer supports for its device.
The effect of the four different iOS devices used in this study on sound level measurements as demonstrated in Table III and Figs. 3(a) and 3(b) show that the older iPhone 3Gs model produced the best agreement for all the apps and noise levels tested (420 samples), with mean differences of 0.44 dB and −0.71 dBA between the apps and the reference microphone measurements. The variability in the results could be due to the different microphone elements in each device as Apple moved to a new supplier of microphones with the introduction of the iPhone 5 and iPAD 4th Generation devices. The differences could also be related to the introduction of a new operating system (iOS 6) that allowed developers to bypass speech filters and input gain control on older devices.
So it seems like Apple’s iOS devices are most accurate; however, app developers haven’t yet figured out how to best utilize the new mics to provide more accurate measurements.