Can You Use an iPhone's Internal Microphone for Acoustic Testing and Accurate Recordings?
When testing audio systems, we find it very useful to have pure, unadulterated recordings of people in different situations. In order to do this, we need a recording with an omnidirectional microphone, done under the right conditions. So, we sometimes resort to just doing our own recordings because we just can't find what we need on the internet. This of course raises the question, is an iPhone recording good enough? If so, great. If not, what simple recorder is? I do have a multi-thousand dollar measurement mic/preamp/data acquisiton system, but we'd like to make things easy and portable and know that we're getting good recordings.
Is the iPhone mic good enough? The microphone needs to have a few characteristics for us to get pure, pristine recordings with reliable characteristics:
The microphone should have the following characteristics:
Omnidirectional. This is necessary so we get a reliable frequency response. There are many, many great recording mics that are cardioid, hypercardioid, or bidirectional. However, those all change frequency response as you get closer or farther from the microphone. As you get closer, low frequencies get boosted. This is an unavoidable consequence of the way these microphones work, not a defect with the microphone. For those more interested, check out the proximity effect.
Low-enough noise: Since we're going to be talking right up close to it, mic self noise is not a huge issue.
Not exhibit any compression until AOP: We need to make sure the mic can handle loud sounds, without compressing the recorded audio at all, until the mic runs out of dynamic range, in which case we want it to clip hard.
Have excellent frequency response: This means that in a perfect world, it would have a flat frequency response from 20Hz to 20kHz. My measurement mic does. an iPhone definitely will not, for numerous reasons. However, for recording speech, if it's pretty flat from 100Hz to 10kHz, it'll be good enough.
The recording software isn't clever: We don't want the recording software to do any compression, either in dynamic range, or in wave encoding. i.e. it shouldn't have an AGC (Automatic Gain Control) and it should not store files in .MP3 or .OGG. FLAC is fine though.
For the first two, simple observation is good enough: we know that the iPhone's microphones are all omnidirectional, and we also know that we can talk on them and record without encountering much noise. The rest is a question though...
First up: Frequency response
I went through and checked each of 3 iPhones frequency response against our reference microphone. Since these are regular MEMS microphones, we expect a pretty flat frequency response down to under 100Hz, and perhaps some peakyness up above 10kHz.
As you can see from these plots, the frequency response is quite good on iPhone 6S, iPhone 7, and iPhone 10. We see a consistent 50-60Hz roll off at the low end, and decent behavior at the high end, except for seemingly the iPhone X, which starts diving down at 10kHz. Not sure why this is. The resolution on this test is 1/3 octave, so not really good enough to look at fine details.
Thre is a bit of an upward slope in the iPhone 6S and iPhone X, yet the iPhone 7 seems flatter. I'm not sure why this would be. These were all 3 measured in their respective cases, so that could also be a factor at higher frequencies, but will have no effect below a couple kHz.
Next up: Dynamic Range Compression
I did 3 tests, at 90dBSPLA, 96dBSPLA, and 99 dBSPLA (dB SPL A-weighted). Above 99dBSPL, the iPhone microphone clipped, at least with the recording software I was using (AudioTools, set to 0 dB input gain. The option goes from 0 up to 20dB, so I didn't seem to have the option to lower the sensitivity.)
Here we see there is little, if any compression on the microphone side. The energy goes from the reference (90 dBSPL), to 96, then to 99, as expected. It is a fraction of a dB off, but that can be explained away pretty easily with my measurement setup, and I'm not concerned about that.
And finally, the Recording Software
It's quite possible that there are many good recording apps out there, but we're using AudioTools. It seems to be good at not messing up the recordings, but the user interface can be hard to actually click on things.
Good news is, it does seem to simply record the samples, rather than doing any sort of compression. It also records as a .wav file, as opposed to the VoiceMemo recorder, which is some compressed format. So... just click on the Recorder button of audiotools, and mash buttons until you're able to do an audio recording and email it to yourself :-)
Conclusion: You definitely can do good voice recordings with your iPhone
Just make sure to use an app that doesn't filter or mess up your sound in any way. So... just go right ahead and record using AudioTools knowing that you'll get a pretty darn faithful recording. Keep in mind that an omnidirectional mic will do nothing to reduce the reverberation in a room like a directional microphone will. So, if you'd like to reduce the reverberation in the recording, go into the closet, or into the great outdoors.