Recommended output from SQ5 to 3.5mm input (ATEM Mini)

[GeoffParticipant]

But should it only use Pins 1 and 2 of the XLR, as Mike indicated above? Why would the reversed signal (Pin 3 of a balanced XLR) need to be grounded?

I would have replied to this earlier, but my SQ5 was at the warehouse and unavailable until today. I’ve been meaning to do some testing to see how the outputs behave when unbalanced and this is the chance to do it.

I started by generating a 440 Hz tone internally and sending it to the L/R bus. Then I assigned the L/R bus to analog outputs 11 and 12 and also the AES output. All bus processing (EQ, dynamics) was bypassed. I measured analog levels with a McCurdy ATS-100 Extended Range VU/PPM meter and AES with a WBS POD-22 Loudness Meter. For sake of simplicity I’ve rounded off the results to the nearest dB.

If I increase the gain until the console meters’ last green LEDs just light, it is sending out +5 dBu on the analog outputs. The AES output reads -18 dBFS. If I unbalance the analog output )short pins 1 and 3), the level into the meter stays the same. If I disconnect pin 3 however, the level drops by 6 dB. The AES output of course does not change through all this.

This all seems to imply that the console uses a cross coupled output topology. The two legs are normally symmetrical, and out of polarity with each other, but if one is shorted to ground, the other doubles its gain to make up for the loss. This is handy for balanced TRS outputs, so that their gain stays the same and nothing is damaged if a TS plug happens to be plugged into them.

Up to this point all these tests were done with a bridging (>10 KΩ) load. If the load is reduced to 600Ω (unlikely, but not unheard of), the level drops by 13 dB when pin 3 is disconnected, 7 dB lower than the bridging example. With both legs connected, but pin 3 shorted to ground, the level stays at +4 dBu. Losing level isn’t necessarily a problem, but unpredictability is, IMHO.

The other gotcha with unbalancing the outputs is that it will change the maximum analog output level. In order to produce the analog equivalent of 0 dBFS, the analog outputs need to produce +23 dBu of undistorted sine wave. With everything balanced, the audible clip level (I was just using my ears, not a distortion meter) is a fraction of a dB below +23 dBu. With one leg shorted to ground it’s +22 dBu. Not bad, I was expecting it to be a full 6 dB lower.

Geoff

[DRParticipant]

Wow, thank you, Geoff, for digging into that and getting some measurements! So practically, if I’m interpreting this correctly, a dual XLR to 3.5mm TRS cable that shorts pins 1 and 3 should result in ~+5 dbU into the ATEM (this seems to match the +4 dbU line level output mentioned in the SQ5 technical datasheet), or at least a significantly higher output than if we were to disconnect pin 3 entirely. And if correct, I assume that same output topology would apply to the 1/4″ TRS outputs.

Given all this info, back to my original question, it seems that the dual XLR out, or using one or both of the 1/4 TRS outputs, with an appropriate cable should work fine. I might have to go back and review my gain staging throughout though, as my livestream out mix tends to be at a low volume unless I’m maxing out the faders at the various downstream points (e.g., Mix1 to Matrix max of 10 dbU, master Mix 1 slider, and individual channel sliders on Mix 1).

[GeoffParticipant]

This may be (mostly) a gain staging problem, as you say.

An easy way to get more gain on an output is to engage the compressor, set the threshold all the way up, and then add 10 or 15 dB of makeup gain. If that’s too much level, turn down the threshold until you get some gain reduction happening. You probably want a “leveller” of some sort on the streaming feed anyway, so that can be a start for dialing it in.

There is an outside chance that you have a ground loop between the console and ATEM, although the ATEM uses an external supply and doesn’t really have a ground, unless it’s getting one from one of its input sources. If that’s the case you’ll need to put a transformer(s) in the path to break that loop. In the setup I’ve been using for more than a year, there was no audible hum without a transformer, but every situation can be a bit different.

Geoff

[Author]

Posts