Hello everyone!
I’m interested in bringing De-Feedback and/or Clear Voice Live into my SQ5+ setup to increase gain before feedback, reduce room reverberation and improve overall vocal clarity.
At the moment, it seems to me that a Waves card with SuperRack Performer running on an M-series MacBook Pro would be the most sensible option, considering the price to performance ratio, the flexibility of using different plugins, not having to carry around an external server, ease of use as well as the good looking user interface, and, of course, the very low latency.
Another option would be a 64 Dante card, which costs almost as much as the Waves card but apparently adds almost twice the latency. It also requires using a third party software to host the plugins, such as Live Professor, SuperRack Performer, or a DAW.
Or, if you only want to use De-Feedback and don’t want the flexibility of running different plugins in the future, you could buy Alpha Labs dedicated NUC along with the recommended Focusrite interface. However, that solution requires an additional conversion stages and introduces significantly more latency, which doesn’t seem ideal to me.
What are your current setups and experiences? Or are you using something else entirely that I’m not aware of?
Unless you think you will want/need to connect to a Dante network to connect to other devices in the future, the Waves I/O card is a better choice IMHO than the Dante card (in general, not just for DeFeedback). As you mentioned, for plugin use or even multitrack recording/playback, the Waves card is faster and doesn’t require paid licenses or hardware on the computer side of the connection. For example, if you want to multitrack 65-128 channels in Dante, you have to purchase their expensive DVS Pro license ($125/yr) or a hardware Dante interface. On the other hand, the Waves driver is completely free. Even if your channel count is smaller (64 channels or less), you would still have to pay at least $50 for the basic Dante DVS license (per computer).
Really the only benefit a Dante card offers over the Waves card is the “larger universe” of Dante products compared to the number of Waves products available. If you don’t foresee a need to use other Dante products in the future, then Waves is the way to go IMHO.
If you already have a Waves or Dante card, then I think trying to utilize it for DeFeedback use is a smart choice. However I don’t know if I would purchase a card solely for DeFeedback use. I would probably purchase/use a USB interface as suggested by Alpha Labs.
All this being said, I have not personally tried DeFeedback at all, so all of this is conjecture on my part.
2 Likes
I just ordered the 4-instance version from SweetTreeProductions, to plug directly into my SQ- USB. You shouldn’t need an interface if you’re dedicating it to a mixer. The mixer is the interface. I’ll report on it after I’ve used it a few times. I have a band I work with with 6 singers in front.
1 Like
Thank you for your message, Dave!
I should have clarified in my original post that using the SQ’s USB-B port wouldn’t be an option for me, as I use it for multitrack recording.
Furthermore, using the USB-B port to integrate De-Feedback into the SQ setup would introduce significantly more latency than using the Dante or Waves option card.
By installing a Dante or Waves option card, I could still use the USB-B port for redundant multitrack recording.
If you could let me know how much latency the 4-instance version introduces once you receive it, I’d really appreciate it.
How would the USB add more latency? The computer is already outputting USB to whatever interface you’re using. By going directly into the SQ I’m eliminating any converters (USB-Dante, D/A + A/D) which is where most of the latency comes from.
EDIT - I think I misunderstood what you are getting from SweettreeProductions. I assumed you were getting the computer AND a USB interface. However after reading your post several more times, I’m guessing that you are only getting the computer and NOT a USB interface.
While using a direct USB connection between the SQ and computer will ensure the audio path stays digital the whole time, that doesn’t mean it is a fast audio path. I believe that A&H uses the default ASIO USB drivers which are pretty latent/slow. All the code that makes the ASIO USB Driver “universal” also makes it inefficient.
You are also incorrect when you say that most of that latency comes from the A->D and D->A conversion. Those conversions can be extremely quick. Just look at A&H system latency. It is less than .7ms which includes a A->D and D->A conversion along with all of the system processing time too. An ASIO USB connection generally has a one way latency around 10ms or even higher. Clearly that extra latency is not just from an A->D and D->A conversion when A&H does it well under 1 ms (more than 10 times faster).
The truth is that using a USB interface (like the Focusrite Scarlett) can actually create a faster audio path than a direct USB connection. That is because many of the USB interface manufactures have optimized their USB drivers for their specific hardware (and cut out anything that isn’t needed) and they are faster than the stock ASIO USB drivers, even with an extra A->D and D->A conversion added in.
Alternatively, you can use Waves or Dante which will also retain a purely digital signal between the console and the computer. Using Waves to connect the two devices together will certainly have less latency than a USB connection because of the way Waves has optimized their drivers. Using a Dante connection with Dante DVS software running on the computer will result in latency that is about the same (or perhaps even more latent) than a USB interface (like the Focusrite Scarlett) connection because the DVS software isn’t that optimized. However if you use a Dante Hardware interface with the computer, the latency will be much lower than using DVS or a USB interface.
Iiiiinteresting. I’ll try to measure it at some point.
Just a question, Alpha labs suggest NOT running anything else on the defeedback server. How much computing bandwidth does Dante or Waves use up, and would it cause problems with a multi instance setup?
Dante or Waves in and of itself doesn’t require any additional software running on the machine. They both require an OS “driver” to be running (the software layer that connects the external system to the computer’s OS), but this is normal and required even when using a USB interface. So using any of those methods to pass audio to/from the computer really shouldn’t affect the use of De-Feedback (outside of the different latency times that each method is going to require).
I just got the hardware in. Plugged it in, added a USB cable, set the USB channels in the system, patched the USB channels in the SQ, set an insert on the channel. Made some feedback, pushed the IN button for the insert, feedback gone! Took about 5 minutes. Can’t wait to try it on a gig.
I looked up the pricing on the Waves and Dante cards for giggles. YIKES! 20% of the console cost!
1 Like
Which is why I said that I wouldn’t buy an I/O card if the only purpose for the card is running Defeedback. Of course these cards open up a lot of other possibilities, so you may end up wanting to get a card for other uses. If you do ever get an I/O card, then I would certainly explore the latency times using the card vs the USB connection. The card should generally result in a faster connection (especially if we are talking about the Waves card because it’s software driver is more optimized than the Dante software driver).
Although I have not tested it myself, I still strongly believe that running a USB interface (like the recommended Focusrite Scarlett USB interface) will result in lower latency times than running the system through the console’s USB connection. Yes it requires more hardware (which costs money to purchase), but if latency times matter in your situation, it can be money well spent.
I’ll be testing latency this week and will get back.
1 Like
Well, well, well. Bummer. I guess I’m stuck with more hardware and cables.
Using the SQ ASIO driver directly, I got 12.54 ms latency. Using the Focusrite Scarlett and it’s ASIO driver I got .53 ms. I’ll try to attach some photos and videos.
Using the internal pink generator with some routing, and inverting one channel, adjusting the delay until I get no pink. Thanks to a Dave Rat video for the method.
1 Like
Still haven’t used it on a gig. But the advantage of the Scarlet interface is it’s universal, and will work with any desk.
2 Likes
I used to use a scarlet 2i2 as my interface for a while before I got my Qu-5 mixer, they’re pretty reliable.
Dave54 - thanks for those latency numbers. Honestly I am shocked that the difference is so great. Which is another way to say that I am shocked at how little latency the Scarlet interface added.
Also, just to make sure people looking at this understand what it represents - this measurement method is accurate in showing how much latency the insert path adds (the audio patch out of the console, through the DeFeedback system and back). Therefore it is the correct number to use when comparing various DeFeedback routing methods, but it does not reflect the total “system” latency from a source (instrument/singer) to the PA’s output.
That’s because this measuring method does not show the console’s latency (roughly .7ms) or other places latency might be added in as well (wireless systems, speaker DSP, etc, etc, etc). It is important to consider the total latency when determining acceptable latency times when deciding if you can run something like the DeFeedback system.
To be blunt, what this means is that if someone wants to use USB output of the SQ console for this purpose, the total latency will actually be longer than the 12.54ms that was measured in this test. It would be at least 13ms and could be even longer. A digital wireless mic might add another 2-5ms in latency for example.
I was a little concerned that I hadn’t done the test correctly, so I ran it again.
Reference channel: Internal pink to local aux6 out to local Ip6 and out to LR.
Test channel: Internal pink to local aux7 out to local IP7 and out to LR. The Defeedback is on an insert, fed in from local Out1 to the devices and back at local Ip1. Insert is disengaged.
First test: Set the Gains to be equal on both channels and bring faders for Ip6 and Ip7 up to unity. Adjust LR for a comfy volume. Invert polarity on IP6 and fine adjust the level on either channel until there is no sound.
Second test: Engage the insert on IP7. Turn on delay and adjust until it goes quiet again. It took a tiny bit of volume re-adjusting, as the devices weren’t exactly at unity.
Result: .88ms added by exiting via analog into the Scarlett, via USB from the Scarlett to the Defeedback device, back via USB to the Scarlett and analog back to the mixer.
This suggests a total latency in use of 1.58ms. I can live with that.
2 Likes
I would completely agree. That is next to nothing when it comes to added latency.
Trying to use the built in USB interface is another story. IMHO, that would really be pushing the envelope when it comes to total latency for live “in the room” use.
Again, thanks for running those test and posting your results. It really has been educational!
2 Likes
