The iPhone XR has three microphones: one near the top of the front face of the phone, and two on the bottom edge. The bottom microphones are higher quality and are automatically used when recording on all of the apps I tried. I haven’t been able to figure out what the top microphone is for, but after reading Bill’s post, my best guess is that it is for noise cancellation. The iPhone XR microphone can record a maximum of 32-bit audio at 48kHz, and has options for both mono and stereo recording. I determined this by testing out different settings in my recording app and seeing which settings it permitted and which it did not.
Apps
I searched here for recording apps on iOS, and I only needed to try out two before I found one that had everything I needed.
- Hokusai 2 records uncompressed, 32-bit audio and allows exporting in a .wav format, but it does not have an option to disable automatic gain control.
- TW Recorder records uncompressed 32-bit audio and allows exporting in a .wav format, plus it has an option to disable iOS processing, which “automatically adjusts the microphone gain”. I believe this is an AGC control, so I have settled on TW Recorder. TW Recorder is great because it has a good user interface with editing tools, and allows both stereo and mono recordings at custom (up to maximum) sample rate and bit depth. Michael Lee recommended this app to me.
Generating Audio:
I used Audacity to generate a sine sweep from 100 Hz to 3kHz over a period of 18 seconds with an amplitude of 0.8, and white noise for 18 seconds with an amplitude of 0.8.
Recording method:
- For my close recordings, I held the phone right above my left computer speaker so that the microphone was pointed directly at the sound source. I put my computer volume at about half of its maximum.
- For my far recordings, I placed the phone slightly away (about an inch) from the left computer speaker, and laid it flat on my computer. I set my computer volume at about ¾ of its maximum or slightly higher.
- My first location was in my dorm room with the door closed and the windows open. It’s a double, so the room is large.
- My second location was in a small practice room in the basement of Pierson college. Since the microphone was extremely close to the speaker, I am guessing the size of the room did not affect the results.
These graphs were generated in Audacity using the “Plot Spectrum” function, while selecting a portion of the sample.
White Noise Observations:
- All of the frequency analysis graphs of the recorded audio look significantly different from that of the sample audio
- Though the results are not overwhelmingly clear, the close recordings appear to have slightly flatter graphs than the far recordings, though
- All of the graphs peak around 8000Hz, then quickly drop off
Sine Sweep Observations:
- All of the graphs have a steep dropoff at 3000Hz, which was the highest frequency in the sine sweep
- The close recording have slightly flatter regions from 100Hz to 300Hz than do the far recordings
- These graphs are all much smoother than the white noise graphs
The only trend with reasonably significant evidence among all of the trials is that close recordings produce flatter spectra than do far recordings. Here is the link to the sound files I used.
I’m glad to hear that my post helped you out in figuring out what the mics do! I think it’s really funny how in the end Android and iPhone hardware ends up becoming really similar, with the two brands both having multiple cameras, mics, and everything like that. I thought what was particularly cool about your results was that the white noise recordings that you had showed a relatively smooth and flat shape, meaning that your microphones are really well balanced in frequency reception.
I think its very interesting to see how the Iphone microphones have evolved with the rest of the phone. Older models such as the 6s don’t record frequencies as well and as wide of a range as the newer models. This makes complete sense, with how crazy fast the new Iphones are upgraded, it’s just very interesting to see just how much the microphones have improved.
Hey Ethan!
Interesting to see the stereo settings here. I kept my settings at mono throughout. I initially thought that the mono setting might have caused the aggressive fluctuations in my graphs. After finding John had similar results, perhaps it also had to do with the iPhone 11 model. I like the point about the sudden peak and dip in the graphs, a very subtle but significant detail concerning the iPhone XR’s frequency range.
I was wondering what the results of a practice room would be! Seems like the curves aren’t as steep, which shows how much the outside environment can affect the iPhone’s results. This can lead to an inaccurate reflection of the iPhone’s recording performance. Glad you got to take advantage of that resource!