Input sensitivity match

[Harmonie]

@Julf

Do I feel a trap here
I'm not a technician, but if you have no induced noise, hum IEC, ground, noise or else issues, then RCA could be a good way after all. Twice lower noise than balanced.

But then I foresee your logical mind saying:
If no noise, then balanced 2x of no noise = not any more noise ....
Is that your point

If a system is ideal, unbalanced will give lower noise.

 

[DonH56]

Differential adds a second active output, which doubles the signal and noise but, since the noise is uncorrelated, the signal increases by 6 dB and the noise by only 3 dB so you gain 3 dB in SNR. The catch is that if you do not need twice the signal then the net result may be higher noise because you are operating the differential connection nearer the noise floor (e.g. if you do not need 6 dB higher signal, and turn it down 6 dB, then you potentially lose 3 dB in SNR depending upon how everything is implemented). But I don't think "twice the noise" is accurate.

You still gain the benefits of common-mode noise rejection and reduced even-order distortion terms.

I usually use "differential" rather than "balanced" because there are a number of quasi-differential schemes that implement a balanced connection but are not truly differential. You give up a little to a lot in noise rejection and performance rejection depending upon how the circuits are implemented at each end. Beyond the scope of this thread, I think.

 

[dajoe]

Doubling the signal by adding the second output was my initial idea to get 1V RCA output to match the 2V input sensitivity of my desired amp. But it doesn't seem so straightforward at all after the feedback here.

So better to select a higher gain amp or live with higher distortion on those last 5-6 dB

 

[Julf]

Do I feel a trap here
I'm not a technician, but if you have no induced noise, hum IEC, ground, noise or else issues, then RCA could be a good way after all. Twice lower noise than balanced.

No trap, but...

If a system is ideal, unbalanced will give lower noise.

This is not really true. As @DonH56 explained, the signal increases by 6 dB and the noise by only 3 dB so you gain 3 dB in SNR by using balanced - assuming your gain structure is optimal. A non-optimal gain structure increases noise, no matter what the topology.