Topic: Input Calibration

When I play a -10.0 dBFS sine wave out through the main (analog) 1/2 outs of my ADI-2 Pro FS R BE with the reference level set to +24 dBu and connect that via a balanced patch bay back to the analog inputs of the unit, with reference level also set at +24 dBu and no inserts, the input level measured in TotalMix FX is -10.1 dBFS. The same happens with louder or quieter signals (the return is always 0.1 dB lower than the send).

Meanwhile, if I use the analog outputs of the ADI-2 Pro FS R BE but the analog inputs of my Digiface AES (via balanced TRS to the front combo jacks), I get a drop of 0.8 dBFS (-10.0 dBFS at output --> -10.8 dBFS at input, in this example), unless I turn on the input PAD, in which case the drop goes down to -0.5 dBFS). I suspect this may have something to do with the input impedance change activated by the PAD.

In both cases, is this expected behavior? The trim gain on the ADI-2 Pro FS R BE is only available in 0.5 dBFS increments, so I can't just compensate there, and the steps are 1.0 dBFS on the Digiface AES, so that's even worse.

Not the end of the world--we are talking about a 10th of a decibel, after all, and I can alway compensate in my DAW, but given that the ADI-2 Pro FS R BE is spec'd for measurement purposes, this seems like a strange oversight.

2

Re: Input Calibration

To prevent overload in loopback (simplest case to demonstrate) the output level in RME units is never spot-on, but a bit lower. The input sensitivity the other way round. This leads to the 0.1 dB difference. The 0.8 dB are indeed caused by output impedance to input impedance, and of no concern - it is just the level that is changed, not the sound.

Regards
Matthias Carstens
RME

3 (edited by KaiS 2023-08-24 09:30:20)

Re: Input Calibration

AdamBier wrote:

...given that the ADI-2 Pro FS R BE is spec'd for measurement purposes, this seems like a strange oversight.

Measurement software doesn’t rely on hardware calibration these days.
Hardware has a tendency to drift which, for highest absolute precision, makes periodical recalibration necessary.

Furthermore there’s always the voltage drop across the output and input impedances, which needs to be taken into account.

Every measurement software I know has options to calibrate in- and outputs in very fine steps, in the digital domain.

Re: Input Calibration

Very interesting, thanks Matthias and KaiS! Any recommendations on measurement software for Mac? I'd like to learn more about this topic and evaluating the objective performance of various signal chains.

5 (edited by KaiS 2023-08-25 06:22:16)

Re: Input Calibration

AdamBier wrote:

Very interesting, thanks Matthias and KaiS! Any recommendations on measurement software for Mac? I'd like to learn more about this topic and evaluating the objective performance of various signal chains.

For iOS / iPad, iPhone:

AudioTools by Andrew Smith:
https://apps.apple.com/de/app/audiotool … d325307477

As an entry, its FFT part is available separate for $12:
https://apps.apple.com/de/app/fft/id298840058

With FFT you can measure frequency responses and analyze distortions and noise very accurately, it’s my by far most used part of the suite.


Have a look if it’s better to load the whole AudioTools suite and in-app purchase just FFT, to benefit from some free goodies, or buy just FFT.

Re: Input Calibration

Thank you very much, KaiS!

In case anyone finds this topic when searching in the future, I've found a couple of good solutions for the Mac:

(1) Plugindoctor (https://ddmf.eu/plugindoctor/) -- cross-platform and available as a plugin or standalone app. Very fast FFT response, though probably not as accurate as other tools.

(2) REW (I recommend the beta version at https://www.avnirvana.com/threads/rew-b … ads.4514/) -- also cross-platform (Java) and VERY full featured and free! There are a lot of great instructional videos on YouTube. Originally designed for home speaker EQ calibration, but you can create a loop of outboard gear using the inputs and outputs of your interface and take all sorts of measurements.