Parallel Compression with hardware compressor

[simul]

Hi everyone! I searched this in many posts but couldn't find anything similar. Sorry if it's already been discussed.
I would love to use my external distressors for parallel compression on drums. I'm mixing ITB since 20 years but I'd like to add some hardware units to the workflow.
I set up an aux with an hardware insert with the right ins/outs and calculated the latency with the dedicated button.
Still I have lots of phasing issues (between the dry bus and the parallel one). Am I doing something wrong? First time using hardware insert in DP
I'm on DP 11,21 version, with RME 802s converters.
thanks a lot!

[HCMarkus]

It will be great if you can get the timing sorted, but I'd suggest you consider using a little analog mixer or your interface's internal mixer (if it has one) to handle the parallel paths. No worries that way.

[simul]

thanks for the answer. not sure I should setup the interface mixer, since I cannot set sends (except for the internal fx).
It's so simple in other DAWs and I can't figure why it's not working in DP.

[EMRR]

The only way this works for me is to have the dry also go analog so latency is fixed. Either rerecord or mix analog before coming back in.

[bayswater]

The only way this works for me is to have the dry also go analog so latency is fixed. Either rerecord or mix analog before coming back in.

I do something similar when using external processors (analog or digital — it doesn’t matter). Put an instance of the Apple delay plugin on the channels that are not going out and back in, and set the delay to match the round trip latency. That plugin can be set in milliseconds. A submix of these channels to an aux helps, so you can do it with one instance of the delay.

[mikehalloran]

Are you connecting via USB or FW?

What is your Mac? RTL (round trip latency) is much faster on a post 2017 Intel or Apple Silicon machine due to the faster CPU and ability to run much lower buffers. A 5.1 or 6.1 MacPro cannot run as fast.

The RME 802 has been out since 2014 and does not support the ultra-low RTL “Fast USB” standard unlike all those inexpensive interfaces using the Cirrus Logic chips released since Oct 2019.

The least expensive interface I know that uses this chip and has effects loops is the Mackie ProFx3 (the 6 channel version doesn’t have effects send/returns).
https://www.amazon.com/Mackie-ProFX-Mi ... ref=sr_1_4

It’s only 2 out/4in over USB so probably not what you need but it will establish proof of concept: If you can get the RTL low enough to be useable, then a hardware solution is possible—you just have to find the right one. Amazon’s 30 day no-hassle return is your friend here.

Most MOTU interfaces use CueMix to compensate for latency — another way to deal with this. BTW, CueMix and the Mackie ProFx series have built-in effects for tracking.

BTW, don’t be confused by all the marketing nonsense you will read about RTL. The differences between the slowest Fast USB 2 using Core Audio and the fastest TB3 with proprietary drivers is the same as moving your ears closer to or further from monitor speakers about 14”. Again, buffer size and sampling rate have a big effect.

I use an analog metronome track to determine acceptable RTL when setting up my system. There’s no such thing as Zero Latency but Near Zero comes very close.

[simul]

The only way this works for me is to have the dry also go analog so latency is fixed. Either rerecord or mix analog before coming back in.

I do something similar when using external processors (analog or digital — it doesn’t matter). Put an instance of the Apple delay plugin on the channels that are not going out and back in, and set the delay to match the round trip latency. That plugin can be set in milliseconds. A submix of these channels to an aux helps, so you can do it with one instance of the delay.

thank you very much! it works perfect with the auplugin on the dry track. it's just annoying for example when I do drums mixing: I usually set up at least 3-4 parallel auxes and if I wanna use the hardware processor, I have to put that audelay on every single aux that's not going out in analog. but yes, bussing everything that remains inside to a single aux it's a smart idea, thanks!

[simul]

Are you connecting via USB or FW?

What is your Mac? RTL (round trip latency) is much faster on a post 2017 Intel or Apple Silicon machine due to the faster CPU and ability to run much lower buffers. A 5.1 or 6.1 MacPro cannot run as fast.

The RME 802 has been out since 2014 and does not support the ultra-low RTL “Fast USB” standard unlike all those inexpensive interfaces using the Cirrus Logic chips released since Oct 2019.

The least expensive interface I know that uses this chip and has effects loops is the Mackie ProFx3 (the 6 channel version doesn’t have effects send/returns).
https://www.amazon.com/Mackie-ProFX-Mi ... ref=sr_1_4

It’s only 2 out/4in over USB so probably not what you need but it will establish proof of concept: If you can get the RTL low enough to be useable, then a hardware solution is possible—you just have to find the right one. Amazon’s 30 day no-hassle return is your friend here.

Most MOTU interfaces use CueMix to compensate for latency — another way to deal with this. BTW, CueMix and the Mackie ProFx series have built-in effects for tracking.

BTW, don’t be confused by all the marketing nonsense you will read about RTL. The differences between the slowest Fast USB 2 using Core Audio and the fastest TB3 with proprietary drivers is the same as moving your ears closer to or further from monitor speakers about 14”. Again, buffer size and sampling rate have a big effect.

I use an analog metronome track to determine acceptable RTL when setting up my system. There’s no such thing as Zero Latency but Near Zero comes very close.

thanks a lot. I'm on a PC running osx, it's actually super fast, with the latest i9 cpu. connection is a FW with the TB3 adapter. I run three 802 (one main and 2 via adat). I've been recording and mixing since 20 years on RME and I just trust them, but never tried this hybrid mixing setup (always worked ITB).
Do you think buffer size will have a difference in this? I always monitored thru the RME mixer, and always kept buffer in DP at 1024, since I never used software monitoring. I thought this Hardware insert issue was a matter of latency in the software, but I'm a noob at this, sorry

[bayswater]

The only way this works for me is to have the dry also go analog so latency is fixed. Either rerecord or mix analog before coming back in.

I do something similar when using external processors (analog or digital — it doesn’t matter). Put an instance of the Apple delay plugin on the channels that are not going out and back in, and set the delay to match the round trip latency. That plugin can be set in milliseconds. A submix of these channels to an aux helps, so you can do it with one instance of the delay.

thank you very much! it works perfect with the auplugin on the dry track. it's just annoying for example when I do drums mixing: I usually set up at least 3-4 parallel auxes and if I wanna use the hardware processor, I have to put that audelay on every single aux that's not going out in analog. but yes, bussing everything that remains inside to a single aux it's a smart idea, thanks!

Yes, it can get complicated or tedious if you have a lot of channels that need the delay. Two ways I’ve dealt with it: do a submix on all the channels that need the delay, or if the set up is repeated, build the delays into a template.

[mikehalloran]

Do you think buffer size will have a difference in this?

Yes. All other things being equal, absolutely — but we just don't know what else is going on with your setup.

The reason I recommended trying that Mackie is because it will tell you if your nine year old interface is part or all of the problem. If you live in the Silicon Valley, I'll loan you one of mine (I have both 10 & 12 channel versions) but if you don't, Amazon's 30 Day Return policy is as close as you'll get to a loan. There's a 6 channel version but it doesn't have hardware effects sends/returns.