formats

SpectraFoo Transfer Functions 102 – audio interface verification

Armed with the key commands, you’re ready to zip through some measurements…

One of the first measurements I recommend performing is to measure the outputs of your audio interfaces against the inputs. This test verifies they match in gain and that there are no malfunctions. This measurement is about as simple as it can get.

(1) audio interface (two channel minimum & preferably with balanced input and outputs)
(2) patch cables (may be XLR to XLR / XLR to TRS / TRS to TRS / TRS to TS / TRS to RCA)

WARNING – IF YOU PATCH AN OUTPUT INTO AN INPUT WITH PHANTOM POWER TURNED ON YOU MIGHT BLOW UP YOUR INTERFACE!!!)

IF YOU’RE NOT ABSOLUTELY SURE WHAT THIS ALL MEANS, AVOID USING XLR CABLES AND YOU SHOULD BE SAFE (WITH THE EXCEPTION OF THE METRIC HALO 2882!!!!

This is what your setup might look like:

Firebox test rig

In this measurement process we measurement two signals passing through the same analog circuitry and D/A A/D converters. Assuming that everything is functioning and the cables are wired correctly, we are comparing identical signal paths so the measurement should be perfect. The expected measurement will have perfect frequency and phase response traces. Like this:

Firebox Out to IN

If you have a perfect phase trace but the frequency response trace doesn’t align to the 0db marker, you have a gain discrepancy. This could be due to different gain settings on the audio interface which is simple to correct & verify or you may have a patch cable that is wired wrong. In either one of these cases, the measurement might look like this.

Firebox Out In 12db

In this case, I created this gain mismatch by boosting the gain on one of the inputs. If I reverse the signal assignment settings, the phase response trace will still be perfect but the frequency response trace will indicate -12db instead of at +12db. Like this:

Firebox Out In 12db rev

If your audio interface has more than 2 channels, it would be wise to measure each one against another. As long as they all measurement exactly the same against another, you can move on to making more complicated measurements.