Guitar Hero/Rockband microphones as cheap USB DAC - Anyone tried it?

[edvard]

I just noticed that Rockband and Guitar Hero mics are showing up at my local Goodwill for downwards of 3 bucks.  "Hmmm... " thinks I, "I wonder if the chipset is keen enough to make a decent guitar-to-USB interface?"
Apparently, they are based on relatively common hardware, so they show up as USB audio devices to Windows and Linux machines (can't speak for Macs, nothing showed up in le Google...).  I suppose I'll have to hack in a FET source-follower to buffer the impedance, and I imagine USB power to be a bit noisy, but I think it can be done.
Has anybody done this already?  And how was the sound quality?

[mistahead]

If they are they are that cheap to get into your hands - yes it would be worth looking at, bearing in mind the upper ceiling on quality (based on original application).

USB power shouldn't be too noisy - but if so isolate it, its not hard stuff to do... If I had a couple around spare I'd give it a crack open and look over.

Correct me if I'm wrong (often) but wouldn't an IC DAC-on-a-chip be in use in these? Surely that's common enough application these days to bundle it up into one package and sell out of cheap plants?

[g_u_e_s_t]

if they use an electret microphone as the pickup element inside, then you can just rip it open, and use the JFET inside of the element as your preamp.  just add a 1M resistor and youre ready to go.

[edvard]

If they are they are that cheap to get into your hands - yes it would be worth looking at, bearing in mind the upper ceiling on quality (based on original application).

Oh, don't I know it... 2 things kept me from just snapping up a few while I was there;
1- Would it even work?  It's made for a console, so I didn't even know if it would work in my computer.  Google search fixed that.
2- Would it sound ok?  I can't find any other information on the chipset, so no specs as to dynamic range, or frequency response, but hey, they're so cheap I can afford to experiment.

Correct me if I'm wrong (often) but wouldn't an IC DAC-on-a-chip be in use in these? Surely that's common enough application these days to bundle it up into one package and sell out of cheap plants?

Judging by the size of these units, yes.  And seeing as it's common hardware, not a proprietary chip, it's probably some prefab thing you can get the same of from Alibaba.com or something, with your brand name on it.

if they use an electret microphone as the pickup element inside, then you can just rip it open, and use the JFET inside of the element as your preamp.  just add a 1M resistor and you're ready to go.

I was hoping to find some specs on the chipset or gutshots or something, but the Google was not helping me in that regard.  I unscrewed the wind screen from one just to see what kind of mic element it had, and it looked like a diaphragm type, so I may still need a buffer.

More info when I snap up a few tomorrow, and I'll post gut shots and whatnot.  Exelsior!!

[mistahead]

Well that's what I was thinking, in this case you've got your hands on the hardware to play with...

If you can ID that chip for the rest of us it would probably help out those of us who can't get the cheap hardware wrapped around it. never know - I could be able to offer some spec advice from back in the early days of USB sound cards...

[edvard]

Hardware acquired... Camera ready... Dissection begins in 5... 4... 3...

Actually, before I start digging in, let me give a few outside details and observations.  I picked up two microphones (in case one doesn't work) for $2.99 each, and it's worth mentioning that these both have VERY long cords, and they are made out of that nice noodley rubber stuff that good headphone wire is made of, not that Rat Shack stiff vinyl(?) crap that physically separates at the strain relief after a few months of use.

The first microphone is a Disney "High School Musical" microphone.  The mic itself is rather light, very "cheap" feeling.  The dongle that houses all the electronics is rather large with screws to fasten the casing together (yay!), and it says Logitech A-0234A.  The A-0234A is also sold by Logitech as the Vantage microphone for the PlayStation, and also has an alter-ego as one of those "bug antennae on a stand" types of desktop usb microphone.  Plugging it in, it shows as "Logitech USB Microphone" on my Linux system, and recording with Audacity is successful, though the sound is a bit muddy.

The second is a Guitar Hero microphone proper, and the manufacturer says "RedOctane", which means it's one of the original Guitar Hero microphones.  The microphone body is on the heavy side, but not heavy enough to imply "quality".  The dongle on the cord is much smaller than the Disney mic, and isn't held together with visible screws.  If there are screws, they are under the labels on either side of the dongle.  The part number is E-UR20, which comes up in a web search as Guitar Hero microphone, no other rabbit holes there.  Plugging it in, it shows up as the exact same model of Logitech USB Microphone(!), and it recorded successfully in Audacity as well, with much more clarity than the Disney mic, though turning up the volume, I can hear that annoying digital "whistle" from antialiasing or whatnot.  Doesn't bode well.


OK, unscrewing the body of the Disney microphone dongle (two of the screws were molded-in dummies O_o) we see this:



The indicated semiconductors are (in order of number of pins):
1- A 5-pin something labeled "AAF38" (no info from internet search)
2- An 8-pin DIP labeled "24LC641", which turns out to be a Microchip 24LC64-1; a 64k serial EEPROM:
http://ww1.microchip.com/downloads/en/DeviceDoc/21189T.pdf
3- A 28-pin DIP labeled "WM9708SCDS", which is a Wolfson Microelectronics AC '97 compatible 18-bit stereo codec chip:
http://www.wolfsonmicro.com/documents/uploads/data_sheets/en/WM9708.pdf
So, there is the possibility of coaxing stereo input and output from this with a bit of hackery, perhaps.
4- A square-package chip labeled "TAS1020B", which is a programmable Texas Instruments Stereo USB streaming controller:
http://www.ti.com/product/tas1020b

In doing my research, I found the last two chips were labeled "Obsolete" or "NRND", so prolly limited availability commercially, but totally available if you hack apart a Disney HSM microphone.  So, there you have it, and I suppose the Guitar Hero mic will be similar but with smaller packages.  I haven't found screws yet...

I didn't see anything reminiscent of a transistor or op-amp on the input side, so this will most likely have to be buffered.  So...

Next up: hacking the input.

[g_u_e_s_t]

the JFET will be inside the actual microphone element itself.  check out this post for complete disection photos:

http://www.openmusiclabs.com/learning/sensors/electret-microphones/

i wonder what the limitations on audio quality are.  the codec seems pretty good, but the onboard anti-aliasing filters are pretty terrible.

[Perrow]

Not to crash the party, but USB sound cards are $1.13 shipped on eBay 

The three dollar mark was passed on page 16 searching for "USB sound card" and sorting on "price + shipping".

Don't know if they're better or worse than GH/RB mics though.

[artifus]

I can hear that annoying digital "whistle" from antialiasing or whatnot.

http://blog.audioworkshop.org/diy-usb-isolator-made-easy/#.UgnZAo3qnnw
http://www.circuitsathome.com/mcu/usb/usb-isolator

Not to crash the party, but USB sound cards are $1.13 shipped on eBay  

The three dollar mark was passed on page 16 searching for "USB sound card" and sorting on "price + shipping".

Don't know if they're better or worse than GH/RB mics though.

the cheap ones tend to suffer from noise too, often lacking decent power supply filtering and isolation.

[markeebee]

I am a total duffer when it comes to digital stuff.  Get your excuses in first......

Would there be a way of synchronizing the clocking of these (or the cheap soundcards that Perrow identified) so that you have a multi-input device? Like, record a whole band at once?

I have an uneducated hunch that it might be possible with a System on Chip thingy like this http://www.eetimes.com/document.asp?doc_id=1278372, but here I'm bellowing through the open window of my own ignorance.

[artifus]

might be possible with asio4all: http://www.asio4all.com/ or with jack on mac/linux: http://jackaudio.org/

some report syncing issues, some don't. stumbled upon someone using a bunch of those perrow linked to a while ago - may have been on instructables?

*edit* yep: http://www.instructables.com/id/Cheap-USB-Simultaneous-Multitrack-Recording/?ALLSTEPS

i'd be concerned about noise.

[Perrow]

i'd be concerned about noise.

For what, $10-$20 for a multi track recorder, I'd at least think about giving it a shot.

Will (most likely) not be professional quality, but in the category "bang for buck" I think you can do pretty well. With some creative post processing I think you can get quite nice results.

[Perrow]

You'd probably get even better performance from that multi usb sound card setup if you'd make splitter cables and record both left and right channel from each card (plus double the amount of channels). Remember to use your built in sound card (which is probably better than those usb ones).

Prepping the signal (level) before the sound card is probably wise too.

ps. about $7.50 will get you a Logitech usb sound card on eBay.

[artifus]

You'd probably get even better performance from that multi usb sound card setup if you'd make splitter cables and record both left and right channel from each card

Prepping the signal (level) before the sound card is probably wise too.

would a decent powered usb hub provide suitable isolation?

hmmm... if noise is generated by lack of filtering and gets in at the input of the adc... could a pseudo balanced xlr wiring solution be found for a clean mono input? using phase of l & r inputs to cancel the noise?

quad op amp. buffer input to (buffered  output? and) invert to left, non invert to right of cheap usb. match gain, invert phase of one and bounce both to mono in daw?

[Lurco]

OP: wouldn`t that rather be "ADC" than "DAC", or am I getting something wrong?

[edvard]

Not to crash the party, but USB sound cards are $1.13 shipped on eBay.

Ah, no worries, no party to crash Bro, this was a "What if..." post rather than an accidental Holy Grail, and so far, I got what I paid for: cheap USB input and an extra cable.  In other words, it works, and maybe a wee bit better than I expected, but I wouldn't recommend doing the same unless (like many folks here) you're in incurable hardware hacker with nothing better to do at the moment.  
I gotta admit, the units on the page you linked look butt-ugly (even the photos show crooked assembly) and with only one stereo output, I would highly debate the "7.1 Channel" designation.  
However, stereo recording and playback in one unit is a definite plus, and it'll likely be built with newer hardware, so better sound may be a possibility.  
Thanks for the link Perrow; at that price, I almost can't afford not to snap up a few.   My next foray into USB territory may be one of those little doohickeys... Stay tuned!

OK, on with the hacking...
First of all, I realize that I could just hack off the mic, solder in a 1/4" jack and plug into a buffer box or whatnot, but the grand purpose of this experiment included using nothing but the USB power.  If you follow in my footsteps, do what suits you.
Due to a serious lack of the manly art of patience, my first attempt was to simply jack my guitar directly into the unit just to see what it would do.  It worked, but sounded dull and boomy (big surprise...).  Tonight, I breadboarded a simple source-follower circuit to jack into the USB unit's input.

The first thing I notice is the sound is quite a bit brighter and clearer compared to my first attempt at just jacking the thing in directly (yet another big surprise... not) with a very slight drop in volume (as expected with a follower circuit).  This is a good thing though, as casual strumming put the levels into the red, so a voltage divider at the output may also be in order.  Below is my schematic:


A few notes:
1- The sound is surprisingly clean.  I have a rather noisy on-board soundcard in this computer (almost-but-not-quite-inaudible digital "junk" sneaks in when I move the mouse) so I was quite pleased to hear that gone when turning everything up.  There was some expected background hiss, but barely enough to notice, and easily taken care of with a noise reduction/gate plugin.
2- The latency is less than desirable.  I don't have hard numbers, but it's an ever-so-slight delay, maybe 15 msec or so, not enough to put you off your game, but enough to need adjusting in production if you use this for recording*.  I attribute this to the whole USB processing system rather than the unit itself.  Do a search for "USB audio latency" and you'll see what I mean.
3- A pull-down resistor is needed at the input of the unit, otherwise you get bad hum.  See schematic.
4- As mentioned before, the output of the guitar is a bit 'hot' for the unit's input, so a voltage divider after the output cap tames things down a bit.  A series resistor between output cap and pull-down resistor works quite nicely.  Maybe just replace the two resistors with a trimmer or pot to tweak it just right.
5- Experiment with different JFETs, you want the ground-to-source voltage to be around 2.5 volts for maximum headroom.  I first tried a 2SK117, but the G-S voltage was 0.41 or so and it clipped with just casual strumming.  I swapped it for a 2N5485, and I read 2.1v; close enough, and no clipping.

So, all in all, a successful experiment, which turned out to be a pleasant surprise.  I'll be boxing this one up and putting it to use.  It certainly doesn't compare to professional multi-channel units with all the bells and whistles, but is certainly an inexpensive option for the hardware-hacking bedroom musician.

*I'm using Linux with Jack, so on Windows, the situation may be better or worse depending on your setup and drivers.  I'm used to ~2-5msecs latency from my crap on-board soundcard.  YMMV.