Modelling Amp Latency And Cumulative Factors

stratman3142

I know this is a bit sad and probably only for those with OCD, but I've often wondered what the latency might be through my Blues Cube Artist, so I did a measurement today.

I split the signal using a Boss LS 2 pedal, sent one signal directly into my audio interface and the other signal into another channel of my audio interface via an SM57 pressed up against the cloth on my Blues Cube Artist. Then I damped the strings on my guitar and made clicking noise whilst recording the waveforms in Reaper. These was a difference of about 2ms (I could only measure to 1ms resolution). The distance of the mic from the speaker cone might account for roughly 0.5ms.
So based, on my measurement, the latency is between 1 and 2 ms.

As a reference, I repeated the measurement with my Laney VC15 valve amp and (as expected) the difference was tiny (i.e. a small fraction of 1ms) which I assume results from the distance of the mic from the speaker

A latency of about 2ms alone isn't enough to bother me, but then if I use my Helix LT (which I believe has a latency of about 1.8ms) with my Blues Cube then the cumulative latency gets up to between 3 and 4ms. Then there's the distance I stand from my amp to take into account (i.e. 1ms per foot).

I know from home recording that I start to feel something is 'a bit off' at somewhere around 8 to 9ms latency. Just something to think about. Now where did I put my 'tin foil hat'.

ICBM

What you need to do is to split the signal after the guitar, feed half of it through your signal chain as normal, and the other half goes through a small clean analogue amp. Then take a line out from this, and from the Blues Cube. Mix the two together and run them into a headphone amp. You don't need much of the dry analogue signal - in fact too much will cause comb-filtering when combined with the main signal - but just enough that you hear the attack transient where it should be, and then the rest of the sound comes in over the top. This will fool your ears that you're hearing the whole sound without latency.

Three-ColourSunburst

If the 1-2 ms of latency caused by the amp is causing an issue, why not just stand a foot or so closer to the speaker?

I guess you would find playing on a large stage or running your signal though a chain of digital pedals to be really off-putting!

Interesting subject. Just read this, which seems to cover the main issues quite well.

http://whirlwindusa.com/support/tech-articles/opening-pandoras-box/

guitarfishbay

I always record with headphones on, even when the amp isn’t in the room or when I’m DI’ing with digital gear. Tightness is necessary for the genres I play so the lower the latency the better.

My understanding is most of the current modellers are 2ms or under, which in itself is essentially imperceptible. I agree once you get closer to 10ms playing tightly gets harder.

I think my round trip is 3.4ms at 32 buffer/48khz, and my Axe FX is under 2ms IIRC. So likely between 5 and 6ms total latency, which I’m ok with if I have headphones on.

At at some point a compromise generally needs to be made in a home recording situation. I think it’s pretty important to try and play into the ‘correct’ sound - much as you can reamp to your heart’s content later if you’ve taken a DI the dynamics of your performance will be influenced by the sound you monitor through.

Danny1969

You can feel the latency when playing through the Helix, it's not a big deal, it's tiny but it's there. Any digital device with AD and DA converters will have a tiny bit of latency 
On big stages the natural latency caused by the distance you are from the speakers is a big issue and very off putting, another good reason for using IEM's

stratman3142

I think my main concern is playing live with a modelling amp plus added serial latency given by other Fx. For example, putting a Helix LT through a Blues Cube. Then standing some distance from my amp.

I know that when I used to reluctantly get pulled out to the front for solos through all analogue rigs (and I wasn't in the monitors) that I could feel the latency on fairly large stages.

Still perhaps it's a good excuse for me to say I need to stick close to my amp

richardhomer

stratman3142 said:

I think my main concern is playing live with a modelling amp plus added serial latency given by other Fx. For example, putting a Helix LT through a Blues Cube. Then standing some distance from my amp.

Presumably if your signal chain were long enough, you could be safely in the bar while latency takes care of your encore....

Three-ColourSunburst

A question for the amp techs. All I have read says that the Blues Cubes aren't modelling amps, instead creating their sound by using analogue, solid state components to create circuits that respond to input from the guitar in a way that is analogous with the way the circuits of a valve amp create their sound. Apparently the only thing that is digital / uses modeling is the reverb circuit.  Given this exactly why would there be more latency in a Blues Cube than in a valve amp?

digitalscream

Surely, if you're playing on a large stage, you'll have monitors to get a consistent foldback, so the distance problem becomes moot?

ICBM

Three-ColourSunburst said:

A question for the amp techs. All I have read says that the Blues Cubes aren't modelling amps, instead creating their sound by using analogue, solid state components to create circuits that respond to input from the guitar in a way that is analogous with the way the circuits of a valve amp create their sound. Apparently the only thing that is digital / uses modeling is the reverb circuit.  Given this exactly why would there be more latency in a Blues Cube than in a valve amp?

It takes a while for the signal to pass through the marketing bullshit.



When you read utter technobabble like this -

“Power supply types such as B+ or C- are significant factors that deeply relate to the curve of power tube gain or saturation.”

(The Tube Logic Story - https://www.roland.com/global/promos/tube_logic/ )

- then you need to be sceptical about any other claims as to how the circuit works.

If you can easily measure the latency on a DAW screen then it’s digital.

mart

digitalscream said:

Surely, if you're playing on a large stage, you'll have monitors to get a consistent foldback, so the distance problem becomes moot?

Yes, but unless they’re IEMs your ears will still be at least 2 metres away from them. The speed of sound is about 300m/s, so it takes about 6ms for sound to travel that 2m distance to your ears. 

So I would suggest that most of us can cope with 6ms of latency without difficulty.

ToneControl

ICBM said:

What you need to do is to split the signal after the guitar, feed half of it through your signal chain as normal, and the other half goes through a small clean analogue amp. Then take a line out from this, and from the Blues Cube. Mix the two together and run them into a headphone amp. You don't need much of the dry analogue signal - in fact too much will cause comb-filtering when combined with the main signal - but just enough that you hear the attack transient where it should be, and then the rest of the sound comes in over the top. This will fool your ears that you're hearing the whole sound without latency.

good advice

This is how they set up delay towers in PAs - you can stand nearer to a delay tower, but still think the sound comes from the stage
(see the Hass effect https://en.wikipedia.org/wiki/Precedence_effect)

The second-arriving sound can be up to 15dB louder before the listener hears that as the main directional source
http://www.audiosonica.com/en/course/post/166/Live_sound-Delay_towers

Three-ColourSunburst

ICBM said:

If you can easily measure the latency on a DAW screen then it’s digital.

QED I guess.

Having read the page in your link, even Roland admits that some DSP is going on, even if the digital processing is trying achieve something other than just emulate pre-amp distortion.

The power amp section consists of a discrete solid-state amp, driven by a high-impedance circuit to interact with the speaker characteristics, another essential part of great guitar tone. This amp-speaker interaction is fine-tuned with DSP for very precise adjustment.

Also from Roland

The final result was Tube Logic – a marriage of solid-state analog technology and DSP that goes way beyond modelling

https://www.rolandcorp.com.au/blog/the-roland-blues-cube-range-explained

stratman3142

More nerdiness.

Picture 1 shows my test set up. The Unit Under Test is my Blues Cube Artist and this time I used the Line Out instead of putting a mic on the amp.

Picture 2 shows a zoomed-in view of the waveforms in my DAW (Reaper). You can (hopefully) see the scale (in seconds) along the top on the picture. This measurement shows that the Blues Cube has about 2ms latency (unless I'm missing something).

Picture 1

https://i.imgur.com/Wpkd7Lc.png

Picture 2

https://i.imgur.com/Z73nj9y.png

ICBM

That's correct, and it's what I've done to tell for sure whether something is digital or analogue too. 

It matters if you're trying to find the cause of odd tone in a parallel FX loop, in case you're wondering!

stratman3142

ICBM said:

That's correct, and it's what I've done to tell for sure whether something is digital or analogue too. 

It matters if you're trying to find the cause of odd tone in a parallel FX loop, in case you're wondering!

It might also matter if you sense a disconnect between what you're playing and hearing (that can't be explained by excess alcohol). The disconnected feeling might result from the cumulative latency associated with a series of digital Fx, a modelling amp, plus the distance you are from your sound delivery system.

The use of a wireless system would also add to things as well, I assume.

ICBM

stratman3142 said:

It might also matter if you sense a disconnect between what you're playing and hearing (that can't be explained by excess alcohol). The disconnected feeling might result from the cumulative latency associated with a series of digital Fx, a modelling amp, plus the distance you are from your sound delivery system.

Yes, and I don't think that should be simply dismissed just because it's within the range that "shouldn't" be audible. OK a few ms is only equivalent to the same number of extra feet from the speaker, but it probably becomes perceptible as a 'feel' issue well before you can actually 'hear' it.

A very simple way to test that would be to set up your amp with a (preferably analogue!) wireless system in a large room and walk away from the amp until you can sense something isn't quite right - that should give you the exact number of ms that matters, if you measure the distance to the amp.

stratman3142

ICBM said:

stratman3142 said:

It might also matter if you sense a disconnect between what you're playing and hearing (that can't be explained by excess alcohol). The disconnected feeling might result from the cumulative latency associated with a series of digital Fx, a modelling amp, plus the distance you are from your sound delivery system.

Yes, and I don't think that should be simply dismissed just because it's within the range that "shouldn't" be audible. OK a few ms is only equivalent to the same number of extra feet from the speaker, but it probably becomes perceptible as a 'feel' issue well before you can actually 'hear' it.

A very simple way to test that would be to set up your amp with a (preferably analogue!) wireless system in a large room and walk away from the amp until you can sense something isn't quite right - that should give you the exact number of ms that matters, if you measure the distance to the amp.

I already know the answer based on recording. The number is in the region of just under 10ms, and I coincidentally just posted the following comment (link below) on the Reaper forum in relation to a discussion on recording latency.

https://forum.cockos.com/showpost.php?p=1946898&postcount=35

Although, perhaps in a live situation, things aren't quite so much under the microscope.

Three-ColourSunburst

Does anyone have any figures for how much latency is generated by various digital pedals?

Three-ColourSunburst

Interesting article on latency as it affects orchestral players.

https://www.linkedin.com/pulse/so-what-latency-tim-shuttleworth

ICBM

Three-ColourSunburst said:

Interesting article on latency as it affects orchestral players.

https://www.linkedin.com/pulse/so-what-latency-tim-shuttleworth

"Perhaps the most interesting finding is that for casual musicians, meaning people not practiced in ensemble playing, the comfort zone is not zero delay. It’s between 8 milliseconds and 20 milliseconds. To make sense of this think about two musicians playing in a room separated 11 feet apart. Now think about them close mic’d, perhaps 2 feet from the microphone and wearing headphones. The sound in their headphones is arriving in about 2 milliseconds but the sound from across the room is arriving after 10 milliseconds. And their minds know that the sound should take 10 milliseconds – life has taught them that – so the headphone sound is too early and it’s disturbing."

That might be something to do with why I hate playing through headphones... especially in a band mix. One of my former bands did try that, in order to be able to practice in a flat rather than a rehearsal room - with electronic drums and everything DI'd - but it was horrible.