The Latency Thread?

I believe that hearing the instrument acoustically together with the processed sound is responsible for a lot of the hullabaloo about latency.
Its definitely a big one, and back when Jim Roseberry was spreading the Gospel, and we first started making the Round Trip Latency ROundup page, it was a perverse relationship where nearly the shortest latencies were the most dangerous because of phase cancellation with parallel sources near the same volume being heard by the performer
 
Its definitely a big one, and back when Jim Roseberry was spreading the Gospel, and we first started making the Round Trip Latency ROundup page, it was a perverse relationship where nearly the shortest latencies were the most dangerous because of phase cancellation with parallel sources near the same volume being heard by the performer

I have the utmost respect for musicians like Royal Blood's Mike Kerr, who made pitch-shifters a staple of their sound, because yeah, the difference between what you feel from the instrument vs. what comes out of the speaker when using those throws me off regularly.
 
Last edited:
You're a science guy yourself. I follow your posts on computing and gear because you're a knowledgeable guy. But let us present the facts.

For example, you made no mention of scan times nor microcontroller processing times, both of which typically add up to 3 or 4ms themselves. This means that the moment you hit the pad, to the moment you hear a sound, there is a baseline 3ms latency response as standard. Even if you're getting a round-trip of 2ms - more or less only possible via Thunderbolt at 88.2khz sample-rate, with specific devices - your total latency will be 5ms.

Not disagreeing about the scan times.
Would you rather trigger those SD3 samples with 6ms (2ms DAW) or 14ms (10ms DAW) of latency?
6ms is going to feel more immediate.

Both PCIe and Thunderbolt can potentially yield extremely low round-trip latency.
Thunderbolt provides external access to the PCIe bus.

Put another way, if there were no performance/cost "penalty" for running at lowest possible round-trip latency, who'd choose to run at high RTL?
 
To illustrate (and yes, for yourself, if you like) how sensitive we are when it comes to latency: Double a signal in your DAW of choice, pan both hard L/R, add some samples of delay (using track parameters, a 100% wet delay /w 0 feedback or whatever) just to one of the signals. That's how incredibly great at detecting the smallest latencies the combination of our ears and brains is.
It's likely also a matter of evolution. When the leaves of the jungle behind you were rustling back in the days, you simply had to know whether it was slightly left or right because that might've easily become the deciding factor for that approaching sable-toothed fellow to have a nice meal or to stay hungry.
For more scientific explanations, one could always look up the Haas effect.

Note: No, this isn't all too directly related to amp modeler and DAW latencies - it just goes to show that at least in general, our senses are absolute suited to deal with extremely small latencies.
That is a good example.
Short delays are how we locate sounds in 3D space.
There was an article I read (years ago)... that discussed using short delays to very accurately pan signals.
 
What I'd be curious about is the amount of latency added by things like the interface and DAW on top of the plugins or hardware. When I play guitar plugged into the modeler and listening to the headphone output, it seems like it's going to be closer to 2-3ms latency. If I plug into the interface and through a plugin, it seems like it's closer to 10-15ms latency, even if the nominal latency is only like 5ms or so.

Just testing right now through the HX Stomp, I added a simple delay with 100% mix and no repeats. You can bring it down to 0.0ms basically just the latency of the unit itself which should be under 3ms according to Leo Gibson. I don't really notice much until it's around 5ms, where it's almost like there's a bit of bloom on the attack. Even 10ms is tolerable. By the time I get to 20ms, it's definitely feeling a little more like a delay than straight signal. At 30ms, it reminds me of playing through plugins from a decade ago, where it really starts to bother me.
 
What is "phase misalignment" a product of?
Short delay between when a sound is heard by your left and right ears.

Yes, but the flight time for the sound to reach your ears is exactly the same. You're just feeding them two different (out of phase) signals.

Tread carefully around ad-hoc examples of latency in music gear. In general, if they involve multiple sounds at the same time, they're showcasing something else.
 
Last edited:
Some of the clubs we play... our drummer is put behind the plexiglass "wall of shame".
The slight delay caused by that drives me crazy.
It causes transitions (to my ears) to sometimes be just a little off.
Can I work with it? Yes.
Does it affect me in a negative way? Yes.
 
Not what people refer to when discussing latency, and particularly digital latency, which is, by its very nature, frequency- and phase-invariant.
Agreed. It’s not the same as assessing the threshold for perceiving tactile latency in relation to an audio event.

Or is it vice versa…. :wat
:rofl
 
Last edited:
Yes, but the time to flight for the sound to reach your ears is exactly the same.
I'm beating a dead horse here... but if the time was mathematically exactly the same (to reach both your left and right ears), it would not be out of phase. You'd perceive that source sound to be dead center (between left and right).

You can use Sasha's example above (very small delay on one channel of a stereo signal) to effectively pan/position it.

As I'm sure you know, you can "phase align" tracks by sliding one either earlier or later (to where the peaks/valleys align).
 
I'm beating a dead horse here... but if the time was mathematically exactly the same (to reach both your left and right ears), it would not be out of phase. You'd perceive that source sound to be dead center (between left and right).

Of course it is. The kicker is that, when you apply signal A a delay to turn it into signal B, they are no longer the same signal. And then they reach your ears at the exact same time: there's zero latency at play.

I mentioned an example a while ago which, while not perfect, better illustrates latency as a phenomenon related to absolute, not relative timing: open Google, type "metronome" and hear it ticking at 100bpm for a while. Clap along with it. Now switch to 101bpm. Can you tell a difference? Could you, if someone else was setting the tempo without you knowing?

Because the difference between 100 and 101bpm is an extra ~6ms.
 
Last edited:
Of course it is. The kicker is that, when you apply signal A a delay to turn it into signal B, they are no longer the same signal. And then they reach your ears at the exact same time: there's zero latency at play.

I mentioned an example a while ago which, while not perfect, better illustrates latency as a phenomenon related to absolute, not relative timing: open Google, type "metronome" and hear it ticking at 100bpm for a while. Then switch to 101bpm. Can you tell a difference? Could you, if someone else was setting the tempo without you knowing?

Because the difference between 100 and 101bpm is an extra ~6ms.
By definition there needs to be a minimum of two observable events in order for latency to exist in the first place.
 
By definition there needs to be a minimum of two observable events in order for latency to exist in the first place.

Yes: action and time to reaction. The examples mentioned above involve two signals for which your ears react at the exact same time :D
 
Yes: action and time to reaction. The examples mentioned above involve two signals for which your ears react at the exact same time :D
All I’m saying is in your metronome example there’s no comparative latency occurring.
 
All I’m saying is in your metronome example there’s no latency occurring.

As mentioned, it's a not a perfect example, but it illustrates how difficult is for us to spot changes in absolute timing. The "latency" here is the beats coming in slower once you increase the BPMs.

But if you run two metronome tabs at the same time, with different BPMs, you'll immediately pick up the difference. Any example of "latency" involving multiple signals is more likely than not unrelated to latency at all.
 
Last edited:
200w.gif
 
That is, by definition, not latency. You're experiencing phase misalignment - and yes, your ears are incredibly sensitive to it.

It is phase misalignment caused by latency.
Or, latency you become aware of due to phase issues.
I've explicitely said it wouldn't be "tactile latency", just another means of perceiving latency.
What we're dealing with when using digital equipment in terms of latency, usually comes down to a mixture of several things, phase issues being one of them.
 
Not disagreeing about the scan times.
Would you rather trigger those SD3 samples with 6ms (2ms DAW) or 14ms (10ms DAW) of latency?
6ms is going to feel more immediate.

Both PCIe and Thunderbolt can potentially yield extremely low round-trip latency.
Thunderbolt provides external access to the PCIe bus.

Put another way, if there were no performance/cost "penalty" for running at lowest possible round-trip latency, who'd choose to run at high RTL?
Yes agreed, but if there isn't a choice, and you get what you get and there is no changing it, then I'm not gonna put a sick bit of music equipment back on the shelf because it gives me 10ms RTL instead of 6ms.

I will confess though that when I run my Helix in 4-cable-method, I do feel a slight "thing" ... not sure if it is latency, or tonal artifacts from two AD/DA conversions. Never tested it really. I just know that I prefer to run my Helix right up front of the amps and not in 4-cm.
 
Back
Top