Question on recording (dry signal vs wet signal), latency, etc

[sd0453]

Hi all,

 

I've had the Helix Native since it came out and love some of the tones I've been able to get out of it. I use an audio interface with very low latency (Focusrite) so I can noodle around with amplifiers. There isn't noticeable latency when I'm playing at all. However, recently I've had some challenges recording more complex songs, and I think it is due to some very subtle latency as the signal gets process in Helix Native. For instance, if I record something while monitoring the wet signal, my recordings always tend to be off just a bit. Everything sounds a little early on the playback. I'm guessing this is because I'm hearing the sound after it is processed, which is causing me to play the notes on my guitar just a little earlier then I normally would. I tested this by recording only a dry signal, with no processing, and voila, my recordings were perfectly in time. So since I know I can recording a dry signal in time, does that mean that I have to record all my music this way and then use Helix Native to process the signal later? It isn't ideal to do it this way, but maybe that's how everyone else does it as well. Would love to hear other people's strategies around this. 

 

 

[SaschaFranck]

First off: It's perfectly normal for players to compensate for latency, especially when you listen to the processed signal at volume (not even very loud, just decent) instead of "listening to your pick". So what you describe is a rather common thing.

Still, latency and its implications is a rather complexed field. The amount of "human latency compensation" varies a lot, depending on tempo, style, sound, mood, listening volume, etc. So it's nothing to really rely on. Which is why you want the lowest possible latencies you can get when going the software monitoring route.

Quite unfortunately, the Focusrites aren't really at the top of the latency race. In fact, by comparison, they're rather mediocre. Don't get me wrong, they're decent interfaces but for players aware of such issues, they're far from being ideal and possibly not made for software monitoring when "in the pocket" time is important (ideally it should be for every musician but let's not get into that just yet...).

In case you're interested in numbers and possible alternatives for lower latencies, this Gearslutz thread should be the largest kinda database on the subject you will find on the web:

https://www.gearslutz.com/board/music-computers/618474-audio-interface-low-latency-performance-data-base.html

I know, it's horribly long but there's some updated lists and such.

 

Alternatively, when it comes to the Helix, you may consider an HX Stomp. That would allow you to monitor through very low latency hardware while recording the dry guitar track, which you may then send into Helix Native. The most amazing thing being that you can use the very same patch you were playing on the hardware.

For me, the combination of a hardware Helix (Floor in my case) and Helix Native is the most relieving thing I have encountered during the last decade (or so) of guitar recording. No latency issues, yet full access to each and every sound detail once I'm done recording. It really doesn't get much better. And I'm defenitely not a guy to just praise things without good reasons.

[sd0453]

Thanks SaschaFranck. I think that sufficiently answers my question. So the way I understand it is....as long as I'm using Helix Native for my amp modeling and using an audio interface to record, I should probably only be recording a dry signal for best results. If I try to monitor the output from Helix Native, I'll have issues with latency in my ears versus what I'm actually recording, even if it is very subtle. What you are suggesting makes sense, which is that I should get the modeling hardware so that I can plug directly into the Helix for recording, and therefore I won't experience this latency while monitoring. This helps clarify things for me.

 

I suspect there are other users of Helix Native that haven't considered this. I'm glad I finally realized it after using it for so long! Yes, I was absolutely compensating for latency myself in many recordings. I often had to wait until the back end of a beat to nail it. Faster riffs with 16th notes were often problematic for me, and for a long time I thought I was just a terrible player. But after recording a dry signal I realized that I'm actually playing in the pocket. 

[SaschaFranck]

You could as well just try to borrow one of the very low latency interfaces, such as an RME Babyface or Fireface UC to find out whether that cures your issues. 

While there's no exact numbers, it seems that once latency is pretty low (let's assume <5ms for a roundtrip), many folks don't compensate for it anymore as the distraction between pick feel and delayed sound is gone (even if there still is latency).

[soundog]

...another latency workaround that doesn't involve buying more hardware:

 

1. Bounce the song you're working on to a stereo track (a stereo export/rough mix of all tracks except the guitar you want to record.

2. Open a new project. Make two channels: #1 import the exported bounced tune;  #2 a blank track with Helix Native plug-in (with software monitoring on).

3. Set your DAW buffers as low as they will go. (32? 64? try it!)

4. Record your guitar track dry (monitoring through Native). Latency shouldn't be a problem with such a light cpu load.

5. Close your temporary project, open your original project, import the new guitar track into a new channel.

6. Set your buffers (ie latency) as high as needed, and apply Native to your new dry track as needed.

 

This sounds like a lot of steps, but it goes fast and easy when you get the hang of it, and is useful for systems that are bogging down on more complex projects.

[SaschaFranck]

1 hour ago, soundog said:

3. Set your DAW buffers as low as they will go. (32? 64? try it!)

4. Record your guitar track dry (monitoring through Native). Latency shouldn't be a problem with such a light cpu load.

 

 

Latency doesn't only depend on the buffersizes, it's mostly about the driver quality.

There's *huge* differences between the real latency values of different interfaces, even if you run all of them at the same buffer settings. 

[soundog]

Well, yeah .... but the driver quality is a constant; the buffer setting is a variable (you can control it and it has a direct effect on the latency). It has not been my experience (on Mac_ that its mostly about the driver quality .... maybe for Windows?

 

On a related note, this very long (8 years running) on gearslutz contains a lot of useful info on audio interface latency, including a database of measurements and info on how to test your own. https://www.gearslutz.com/board/music-computers/618474-audio-interface-low-latency-performance-data-base.html