I thought our new Karaoke machine made our voices sound worse than the were, if that’s even possible.
It so happens that I’m an acoustical engineer, so I actually have the knowledge and tools to measure the microphone performance. So, here goes…
Our karaoke machine is a Karaoke USA GF829, which seems popular, and was recommended to me.
I measured the frequency response of the microphone, and lo and behold, the response is pretty remarkable.
Below, you see the frequency response from 60 Hz up to 20 kHz. At the top (purple) is 0 cm from the mouth, then 1cm, 2cm 3cm, 4cm, 5cm, and 10cm. As you can see, the response is absurd when your mouth is right on the microphone (which is a good place for it to avoid feedback). The fundamental of your voice (the 100-200 Hz range) is boosted nearly 20 dB above a very important range (1kHz to 2kHz).
At the 10 cm distance, the fundamental drops by 30 dB, but the ever critical range from 1k to 2k only drops 20 dB.
So… really, this microphone is kind of terrible. Ideally it should be pretty flat when your mouth is right on the mic (or perhaps a bit of bass boost), and loose bass as you pull away to 10cm.
Now I’m going to have to build a proper karaoke microphone. I wonder if this one can be easily fixed with a passive filter.
Here is the data normalized to 1kHz. Even in the best case, there’s a big hole at 1kHz, which is definitely bad.
Next step: measure the karaoke machine’s loudspeaker.
Many thanks to the good folks at Listen Inc. for the use of their great SoundCheck software.
Here’s the .dat file for SoundCheck that has the raw data. karaoke-microphone-response.dat
And the .dat files for the ST59AP microphonest59ap.tar