IntroAs a hobby guitar player, I haven't been very happy lately with my FCA202 Firewire interface. I have no problems with the sound quality, but the driver problems are nagging. The latency seemingly increases over time of usage and it doesn't deal well with the Mac going to sleep.
Garageband doesn't initialize it properly (apparently) if the FCA202 appears on the FW bus after Garagebands start and so on... So it's really more a comfort problem. My hope was that the drivers of a different vendor would be better supported then the FCA202, which relies on the built in CoreAudio support alone and apparently it doesn't get much love and affection from Cupertino.The Firewire drivers don't seem to get much love and affection either, given their sorry state...
My other comfort problem was, that I didn't know what to wish for christmas. My proposed solution to both problems: the Apogee Duet. Most likely overkill, yet hopefully killing.
Driver problems are gone
I can now unplug and replug the device under Garageband. Go to sleep and wake up and everything is working fine. So it's a complete success in that respect. The integration with Mac OS X is very good, the device makes you feel like its part of the system. When you turn the knob, sliders appear on the screen. Very neat.
Interestingly the Maestro package comes with an apple firewire audio package, that gets installed along the other stuff: AppleFWAudio-2.2.0fc9-e1. Curious.
I thought that maybe that driver would also cure my FCA202 blues, but to no avail.
Usage and sound quality
I recorded myself on guitar, straight through both interfaces with no extra gear using Garageband (no effects).
A problem appeared immediately that there is no way to have (sufficient) negative gain. Mind you the FCA202 has no gain at all and also no level indicator, but with the Duet it became apparent that my guitars SH-4 pickup can produce more output than the Duet can take and that I would need to turn the volume on the guitar down to be safe.
I have to hit the strings really hard though to produce overs. Nevertheless to be safe, I'd be turning the volume on the guitar down some. It's not really the optimal solution, since that kills some of the highs. A -20db pad on input would have been a nice feature I think.
After recording I tried to discern some differences in sound quality. I couldn't make out any, neither over the JBL Aliens nor over my AKG K240DF headphones. I don't claim that there aren't any differences, but I can't tell my recordings apart (with what I have available).
Later: Well the control knob on the Duet is a little sucky. You'd expect that there would be a constant time for presses to be recognized and interpreted. Unfortunately they are widely varying. It's supposed to work like this. Quick press means cycle the display, longer press (2s or so) mute/unmute, really long press (5s) reset. Sometimes you have to press for 5s just to cycle the display...
Measurements using FuzzMeasure
Then I used a demo version of FuzzMeasure to measure the harmonic distortion and the frequency response, to see how measurements would differ. I used 44.1Khz/24bit because that's what I am using in Garageband. Both interfaces do better when in 96KHz mode, but I haven't used that mode for recording yet.
- Orange: FCA202->FCA202
- Red: Duet->FCA202
- Green: Duet->Duet
Here's the setup used. I connected the line-out of the Duet once with it's instrument input (green) and once with the FCA202 input (red). I also connected the FCA202 once with itself (orange).
The -3dB offset is a longer story. It's mainly there because the Apogee is a little funny in that respect. If you record an instrument at +0dB and play it with line out +0dB you will have a -3dB offset. So I wanted to compensate for that. I set the FuzzMeasure amplitude to -3dB on output and the gain on the Duet to +3dB to make a nicer graph and also to avoid some possible distortions because of samples hitting +0dB. But that didn't work because FuzzMeasure uses CoreAudio for attenuation and doesn't scale the sample values. And the FCA202 can't be volume controled from CoreAudio (nor any other place), the Apogee Duet can though. So in the end, the FCA202 got 0dB samples whereas the Apogee got -6dB samples that it scaled back to -3dB.
So I would think that the FCA202 had a slight distortion disadvantage but also a slight noise floor advantage in the FCA202->FCA202 measurement test and that's why there is this 3dB offset in the frequency response chart. But it's really of no interest. It's interesting to note though, that the Apogee absolutely nails that -3dB line. A point for extra effort - with no musical consequences.
Here's how I interpret the measurements: I'd call that basically even for recording and a win for the Duet on playback.
The orange line is FCA202 into itself. As one can see the frequency response is slightly worse than in the other two measurements. But the the red graph shows, that recording with the FCA202 has basically the same frequency response quality as the Duet. Nevertheless "slightly worse" means 0.1dB at 10KHz +, which shouldn't matter IMO.
As to distortion, the FCA202 has apparently more harmonic distortion in the lower frequencies <400Hz and the Duet in the upper range >1K, so that should even out. I interpret it, that for recording (ADC) the converters are pretty much in the same ballpark (at least for these two measured properties). But for playback the DAC converter of the Apogee is a bit better.
Tests with Audacity
I used Audacity's - on 10.4, doesn't run on 10.5 for me - signal generator to create a file of three short testtones. One 2KHz Sawtooth, followed by a 16KHz sine and a 40Hz sine (all with a 0.8 amplitude value). For these heavy testcases the results are almost the reverse, the FCA202's ADC doesn't cut it, but the DAC isn't so bad. The Apogee Duet finally proves its moneys worth.
Settings used for Duet. I used these settings because the output is compatible with the FCA202 and it offsets the -3dB while recording.
Especially the sawtooth is a problem for the converters as you can see. The Apogee ADCs converter beat the FCA202 soundly in this test.
The top sample is the test tone sample. The samples below are in order FCA->FCA, FCA->Duet, Duet->FCA, Duet->Duet.
Close up of the 2KHz sawtooth. The Duet does very well. The FCA202 ADC don't like it.The FCA202 DAC doesn't mind.
Spectral display. Duet and FCA202 don't mesh as can be seen in the 16Khz test.
Headphones
Headphone volume is much better on the Apogee. The FCA202 can't drive my 600 Ohm (common are 55 Ohm) headphones fully, but the Apogee can produce the required output.
"Reamping"
One more problem with the FCA202 was reamping. It's not usable to me, because the amplifier picks up too much garbage noise. I assumed that this was over the firewire bus and I could fix that by cutting the firewire power from the cable and by using the supplied power supply. But I never tried it and that's a good thing because the idea is just bullshit.
It's one of these inscrutable grounding problems, that I don't know how to fix properly. Gotta read that "Sound Reinforcement Handbook" again...
So reamping isn't possible for me with the Duet setup either, because it doesn't fix my problem magically.
I fixed the problem by interposing a DI Box between the Duet and the amp and using the Ground Lift on the DI Box. The DI Box adds a little bit of noise to the noise floor, but it's acceptable. Harddisk noises be gone!
Conclusion
My personal conclusion is that since even Apogee can't really improve much upon the home recording quality that's available for less than €100, my decision for picking a firewire recording audio interface would be all about convenience and comfort first and price second. The Duet is pretty much there, but not yet perfect and its probably a bit too pricey. For home recording of guitars one can probably get more for less somewhere else.
References
The Apogee Codec at $8.20 at Newark 1/08
The Behringer Codec at $7.43 at Newark 1/08
iPod measurements goes into more detail about measuring in general.]]>
