I bought an iPad a few weeks ago, and I’ve been using GarageBand on it to demo song ideas when I’m not around my normal home studio setup. It’s a neat program but I was getting a little tired of using the “Smart Guitar” and “Smart Bass.” I wanted to record some actual audio to the thing and to that end I bought an audio interface for the iPad.
The aforementioned interface shall remain nameless, but it plugs into the headphone jack and sends the guitar signal into the microphone input. There are two different audio interfaces that I’ve seen that work in this way, and they both use a little JFET amplifier to do impedance matching. This JFET is powered by the 1.5V that the iPhone / iPad uses to bias the electret mic in the headset. It’s a great idea, but it falls flat in practice. In my experience, the interface had very poor noise performance and lacked low end. The low end loss is understandable, since the phone uses a high-pass / low roll-off filter to keep down wind noise and rumble from the headset mic. The noise is a bit puzzling – didn’t anybody listen to these things? JFETs vary wildly due to their construction, but if I can get a bad one right off the shelf then surely others can too. I would not build and sell a product that was likely to perform poorly right off the bat.
I put the thing back in its little box and returned it.
But now I’m without an audio interface, and that screaming guitar lead won’t play itself… So I decided to tinker. I had read that the iPad supports class-compliant USB audio devices using the Camera Connection Kit, so I hooked up a Logitech USB headset. It seemed to work pretty well, and I could get audio into GarageBand with it just fine. I didn’t want to hack up my good headset so I poked around looking for a USB audio device.
I found this:
It’s a Logitech USB mic, and they were sold for the Karaoke Revolution and Rock Band games. The Rock Band mic is identical but has the RB “Mic” symbol on the little box. The box opens up with two screws and there’s a little PCB inside. I pulled the mic off its solder pads and wired up a little voltage divider to keep the guitar from overdriving the mic input. There’s also a bias voltage feed that powers the mic’s capsule that has to come out.
I ended up soldering about 6′ of George L’s cable to the board and re-using the strain relief from the mic cable. Took a little trial and error to pick the right voltage division ratio – right now I am using 150K / 15K, which divides by 11. It pads the input enough so that a low-output guitar can drive the iPad okay, but I have to turn down my hotter guitars and active basses down a little bit. It’s a decent compromise but you could probably use 220K / 15K if you wanted a little less input volume.
I’ll get pictures up this evening.