ADAT / delayed sounds problem

[ndurant]

I've just started using multiple outputs from my Receptor 2+ Pro using the ADAT output and a Behringer AD8000. Plugins generally seem to work fine, and can be assigned to different outputs as expected.

However, with SampleTank (and SampleTron/SampleMoog) I find that when using the ADAT outputs, sounds often fade in with a delay of maybe half a second. If I play the same note repeatedly, it sounds instantly however, but then if I move to a new note, it will suffer from the same delayed fade in effect. When routing to the normal analog output every note sounds instantly as expected.

Can anyone shine any light onto why this might be happening, and any possible workarounds? Any ideas much appreciated!!

[ndurant]

Just replying to my own thread here....I discovered a workaround to the problem, which is very curious.

If I route any of the SampleTank type plugins, or Kontakt to an ADAT output on its own, I get this strange delayed fade-in effect on some of the notes. However if I then add another Receptor channel with Line Input as the source, and route it to the same ADAT output, the strange delayed fade-in effect disappears and it sounds as expected. It's as if the problem only occurs if SampleTank or Kontakt is the only thing routed to that channel, and the problem goes if it's sharing the output with something else.

I'd love to hear if anyone can shed any light on why this might be happening...and whether it's just my machine or a general issue.

Spherical

[batgab]

It's a head-scratcher, ndurant! What you're saying is that the audio sounds like it's going through a gate with a half-second attack time; you play a note, the "gate" opens slowly, and as long as you keep playing repeated notes the "gate" stays open. You release a note and the "gate" closes. The summed Line In signal may be enough, even if it's noise, to open the "gate" and keep it open.

I'm not saying there's a gate in your signal path, just characterizing your description. Different plug-ins routed to the same ADAT output channels don't exhibit the problem, I presume.

Is it possible that when you use SampleTank and Kontakt you are routing audio through the ADAT outs AND either the line or UniWire outputs so that your DAW is summing signals from multiple sources that may be opposite polarity (temporarily), causing cancellation? Probably a very remote chance, but I'm just trying to throw out some ideas!

Does changing your latency/buffer setting change the fade-in period?

[ndurant]

Thanks for your ideas batgab. Your description as behaviour like a "gate" is a very good one - it behaves very much like a gate in fact.

I don't think your suggestion of opposite polarity/cancellation is happening with my setup, as I'm just passing the 8 ADAT outputs direction from my DA converter (Behringer ADA8000) into an analog mixer, and in the case I'm describing here I'm only getting a signal on two channels (i.e. stereo), and not mixing with anything else. No DAW involved.

The effect seems to disappear whenever anything else is fed into the same ADAT output, including silence, and so if I put any other plugin onto the same ADAT output, the problem doesn't appear for either plugin. By the way, I've now seen it happening with the Minimonsta plugin too, so I don't think this is plugin-specific (although I haven't seen it happen with VB3 yet, interestingly).

Another interesting thing is that if I route silence from the Line In input (which I'm not using) to the same ADAT output, it fixes the issue as I mentioned in my previous post. However if I turn the volume of that Line In channel right down, once the fader hits zero the problem comes back. Similarly if I mute that channel, the problem comes back.

I've tried different latency/buffer settings, and there's no perceivable difference in the strange effect, going right from 32 to 1024 samples.

Bizarre....

[batgab]

Your information on how mixing the Line In source enables normal VSTi audio (when the Line In channel is not turned down all the way) and disables normal audio (when turned down) is an important clue, especially when coupled with the buffer setting info. It seems that when the Line in source is not fully attenuated, there is enough analog noise - even if it's way down, like -100dB or something - that it causes your signal chain to pass the audio normally. Think of it this way: when using only the VSTi, whenever you are not playing a note, the digital audio output of the VSTi and the Receptor channel is zero. There would be no digital bits present in the Receptor channel digital audio output stream that represent audio signals. But when you add in the Line In source, now the Receptor is taking the analog input (even if no gear is plugged in), converting it to digital, and mixing in with your VSTi output. It's entirely possible that the Line In noise level, while very low, is sufficient, when A-to-D'd, to be present in the Receptor output stream, causing something in the audio chain from that point on is "opening the gate". The question is, is it the Receptor ADAT output, the AD8000 or your mixer that is gating your signal?

You could try plugging your AD8000 outputs into a different mixer/amp/whatever, just to see if your mixer might be the culprit. Unlikely, but you never know.

Does the problem present itself no matter which AD8000 channels you use? Is there any other D/A converter you could plug your receptor ADAT outputs into? Honestly, it wouldn't surprise me that a device like your AD8000 would mute its output if it was seeing no input signal, as long as its analog output signal had decayed sufficiently. I.e., a gate. Does Behringer have a forum you could search through re this possibility?

If you have a piece of gear with a S/PDiF input, you could check to see if the Receptor S/PDIF output exhibits a similar problem.

I might play around with my Receptor 1 to see if I can replicate this phenomenon. I don't have the same signal chain or VSTi's as you, however.

[ndurant]

Thanks for your further ideas batgab!

Your suggestion about the Behringer ADA8000 possibly gating the problem prompted me to do some digging around the Internet, and it looks like I've found the answer here:

http://behringer-en.custhelp.com/app/an ... og-outputs.

Because I know several people who use the ADA8000 with the Receptor without any problem, I assumed the ADA8000 was not causing the problem. However it seems Behringer added a gate on later models, and I guess I'm unlucky enough to have one of these newer ones.

Also the reason I was seeing the issue on all plugins except the VB3 organ plugin is because VB3 generates a small amount of wind noise from the Leslie simulator, which I guess was keeping the gates open.

Anyway, mystery solved, and I can just work around it by leaving a low-level "silent" channel feeding into the ADAT outputs. Thanks for your assistance batgab

[batgab]

Alright, yeah, mystery solved! Glad I could help.