Helix HW Latency

[rd2rk]

@phil_m

 

4 years ago you posted this:

 

There's always going to be some latency with anything digital, but the total throughput latency of the Helix is very minor - like around 1ms.

 

Someone on TGP posted about the much better latency of the GT1000, and made me watch a long annoying video wherein a very well meaning guy tested very scientifically(?) and came to the conclusion that the latency on the GT1000 = .5ms vs Helix = 2ms.

 

BOSS GT 1000 vs LINE 6 HELIX vs a Real PLEXI: FEEL and LATENCY test! - YouTube

 

I won't torture you by making you watch the whole thing (like the TGP guy did to me!), just skip to 15:40.

 

What say ye?

 

This came up during a discussion of 4cm, where such tiny differences can add up quickly with multiple FX loops.

[somebodyelse]

Everything has a latency.

Every foot between the speaker and your ears adds 1 ms.

If the total, round trip latency adds up to less than 12 ms, you can't HEAR it.

 

It's as bad as the sample rate and bit depth BS,

If what YOU'RE doing works for you, why give a f*** what anyone else's gear can or can't do?

 

[datacommando]

17 hours ago, rd2rk said:

What say ye?

Latency? It hasn’t troubled me in the past 5 years, so I didn’t think I will be worrying about this sort of (scientific) nonsense in future. This was previously mentioned in respect of the Helix processed and dry audio signals into a DAW for re-amping. If you compare the signal peaks they are slightly off. It is easy enough to fix - not a problem.
 

Frankly, my Helix is very good  at what it does - I won’t lose any sleep.

[kringle]

Thanks for sharing the link, I personally always find this stuff interesting even if it doesn't really change anything or mean one is actually better than the other.

 

Just some thoughts I had:

• His methodology seems effective, however the fact he loaded an IR instead on the helix instead of just a clean signal may have impacted the latency.
• I was a bit surprised to see how far off the wave shape was and that it was inverted between the amp signal vs the modelers signal. I would have expected a fairly similar wave shape, but it was totally different.
• assuming numbers are accurate, the Boss at 0.5ms is phenomenal latency for anything digital.
• The helix may have some settings, padding, or other processing ahead of it's actual signal path. or maybe there is opportunity to speed up, 2ms is high for a CPU & Analog to Digital  converter so not sure that 2ms is even accurate.
• Potentially there should be some tests as you add a few blocks and see if things change or compound latency, things might level out.

Either way, they are both great modeler units and this is all insignificant to the ear and reality of just playing and using it... like I said, I do like to see stuff like this, so thanks again for sharing.

[phil_m]

The Boss’ lower latency has to do with it running natively at 96kHz compared to 48kHz, I believe. That is the one advantage to running at higher sampling rates. The Helix oversamples at much higher rates at various points in the signal path as well (hence the increased oversampling with 3.10), but I believe an empty path will run at 48kHz the whole way through.

[PierM]

Dunno, but I did the same test with my HXStomp and I get 0.001 (1ms). I assume it's gonna be the same with the Rack. Going with samples to improve measurability, at 48Khz sample rate, test gives me ~70 samples, which is around 1.4ms (70/48000=0.0014583333333333 Sec)

 

Difference with my test is;

• I didn't put any block in the path, just IN to OUT.
• Signal at the audio interface (Focusrite Scarlett 6i6) is exactly the same. Same signal, same phase.
• No arbitrary reference waveform, since the waveforms are the same, just in a offset (due the latency)
• Reference signal goes into a Y splitter, signal A goes into HX Stomp L Input. HX Stomp L Output goes into Scarlett. Signal B goes straight into Scarlett.

 

Adding a AMP/CAB does still give me around 1ms

 

 

 

[kringle]

PierM - that is more in line with what I'd expected. it's practically immeasurable in raw pass-through and your signals are the same.

[rd2rk]

9 hours ago, somebodyelse said:

Everything has a latency.

Every foot between the speaker and your ears adds 1 ms.

If the total, round trip latency adds up to less than 12 ms, you can't HEAR it.

 

It's as bad as the sample rate and bit depth BS,

If what YOU'RE doing works for you, why give a f*** what anyone else's gear can or can't do?

 

 

I know all of that. None of it is entirely correct, but that's for another discussion.

I'm asking a question specifically related to the difference in latency between two specific pieces of hardware, which arose from a discussion I was having in another forum.

I asked the question of someone who is known to have some knowledge of the subject, and has commented on it previously.

The answer is of interest to me, and that's all that matters.

 

 

[rd2rk]

4 hours ago, datacommando said:

Latency? It hasn’t troubled me in the past 5 years, so I didn’t think I will be worrying about this sort of (scientific) nonsense in future. This was previously mentioned in respect of the Helix processed and dry audio signals into a DAW for re-amping. If you compare the signal peaks they are slightly off. It is easy enough to fix - not a problem.
 

Frankly, my Helix is very good  at what it does - I won’t lose any sleep.

 

Not losing any sleep, just asking a question.

It only really matters when I'm trying to advise someone who wants to use all 4 FX Loops for additional digital devices, as it adds up.

I won't be running out to buy a GT1000 to recover 1.5ms. :-)

[rd2rk]

2 hours ago, kringle said:

Just some thoughts I had:

• His methodology seems effective, however the fact he loaded an IR instead on the helix instead of just a clean signal may have impacted the latency.
• I was a bit surprised to see how far off the wave shape was and that it was inverted between the amp signal vs the modelers signal. I would have expected a fairly similar wave shape, but it was totally different.
• assuming numbers are accurate, the Boss at 0.5ms is phenomenal latency for anything digital.
• The helix may have some settings, padding, or other processing ahead of it's actual signal path. or maybe there is opportunity to speed up, 2ms is high for a CPU & Analog to Digital  converter so not sure that 2ms is even accurate.
• Potentially there should be some tests as you add a few blocks and see if things change or compound latency, things might level out.

 

1 hour ago, PierM said:

Dunno, but I did the same test with my HXStomp and I get 0.001 (1ms). I assume it's gonna be the same with the Rack. Going with samples to improve measurability, at 48Khz sample rate, test gives me ~70 samples, which is around 1.4ms (70/48000=0.0014583333333333 Sec)

 

Difference with my test is;

• I didn't put any block in the path, just IN to OUT.
• Signal at the audio interface (Focusrite Scarlett 6i6) is exactly the same. Same signal, same phase.
• No arbitrary reference waveform, since the waveforms are the same, just in a offset (due the latency)
• Reference signal goes into a Y splitter, signal A goes into HX Stomp L Input. HX Stomp L Output goes into Scarlett. Signal B goes straight into Scarlett.

 

Thanks for doing that @PierM. My thoughts were similar to that and @Kringle.

The only time this matters (to me) is when someone wants to use all 4 FX loops to for additional digital devices, the latency adds up.

I got interested because there's a guy on another forum that aggressively pushes this latency thing and when I called him on it, all he could do was link the video I linked.

This is a spec that it would be nice to see provided by Mfrs.

 

[rd2rk]

1 hour ago, phil_m said:

The Boss’ lower latency has to do with it running natively at 96kHz compared to 48kHz, I believe. That is the one advantage to running at higher sampling rates. The Helix oversamples at much higher rates at various points in the signal path as well (hence the increased oversampling with 3.10), but I believe an empty path will run at 48kHz the whole way through.

 

Thanks @phil_m. I thought maybe something like that, and maybe to do with their use of a proprietary purpose-made chip vs the SHARC.

[PierM]

18 minutes ago, rd2rk said:

 

Not losing any sleep, just asking a question.

It only really matters when I'm trying to advise someone who wants to use all 4 FX Loops for additional digital devices, as it adds up.

I won't be running out to buy a GT1000 to recover 1.5ms. :-)

 

I see your point.

 

I do consider myself VERY sensible to latency, but I have zero problems running Helix Rack + 2 loops (two mono) going into two external EDPs and the other 2 loops (1 stereo) going into an HX Stomp, which has another chain of pedals in its own FX Loop.

 

In all honesty, I feel and hear no latency that I could talk about, and that's Helix Rack + Stomp + 7 extra DSPs of other pedals running through their FX Loops.

 

People likes to think their gear is "potentially" working better than something else, even when that doesn't make any difference in a real life scenario.

 

Next I will test latency of dry VS my entire rig. :D

[rd2rk]

BUT, BUT, Steve Vai can hear 2ms! :-)

 

I hear you, My HXS in the office is connected to an ancient 8gb i5. When I fool around with AT5 or THU, the best I can do without glitches is 128spls, which results in 20.9ms. I can "feel" it, but it's not nearly so bad that I can't still have fun.

 

[PierM]

This is at least 3 minutes latency....XD XD

 

 

[hideout]

Latency? Who gives a rat's patootie?

If you can actually hear that delay or worse yet, if it actually affects your playing, you need psychiatric help.

[datacommando]

6 hours ago, rd2rk said:

BUT, BUT, Steve Vai can hear 2ms! :-)

 

BUT, BUT, Steve Vai can do this after surgery on his right shoulder, and carpal tunnel messing up his left hand.

 

I think I might get very drunk and bite my fingers off!

 

;-)

 

 

[rd2rk]

FOR THE LOVE OF GOD!

 

I'm not even sure he's human......I mean, ALIEN LOVE SECRETS!

 

How else can he be explained?

[cruisinon2]

3 hours ago, rd2rk said:

How else can he be explained?

 

Well there was that Crossroads thing...;)

[somebodyelse]

19 hours ago, datacommando said:

Latency? It hasn’t troubled me in the past 5 years, so I didn’t think I will be worrying about this sort of (scientific) nonsense in future. This was previously mentioned in respect of the Helix processed and dry audio signals into a DAW for re-amping. If you compare the signal peaks they are slightly off. It is easy enough to fix - not a problem.
 

Frankly, my Helix is very good  at what it does - I won’t lose any sleep.

It's actually easy to avoid in Reaper (and presumably other DAWs) as it comes with a latency compensation plugin for send/return purposes.

[rd2rk]

9 hours ago, somebodyelse said:

It's actually easy to avoid in Reaper (and presumably other DAWs) as it comes with a latency compensation plugin for send/return purposes.

 

The question has nothing to do with "ITB (DAW)" latency, two different animals. The question relates to input to output latency within the Helix/GT1000 themselves, which is primarily caused by the AD/DA conversions, and secondarily by the processor speed. Since converters these days are all pretty fast, as phil_m suggested, the difference between the Helix and GT1000 could be the oversampling, 48k vs 96k.

[theElevators]

You adjust your playing to the latency subconsciously.  Ever thought of more latency that your wireless system introduces, then how it has to travel to FOH, then to your monitors, that's some more latency.  Then you hearing the sound from the venue, that's more latency still.  

[rd2rk]

12 minutes ago, theElevators said:

You adjust your playing to the latency subconsciously.  Ever thought of more latency that your wireless system introduces, then how it has to travel to FOH, then to your monitors, that's some more latency.  Then you hearing the sound from the venue, that's more latency still.  

 

True.

But again, not relevant to the thread topic, which is simply a question about actual measurable latency WITHIN the HW, not whether or not it affects my playing, or your playing, or what to do about it. Just, simply, "What is the actual real world objectively measurable latency of the hardware?".

[cruisinon2]

What's the round trip latency for boomerang latency threads? ;)

[somebodyelse]

15 hours ago, rd2rk said:

 

The question has nothing to do with "ITB (DAW)" latency, two different animals. The question relates to input to output latency within the Helix/GT1000 themselves, which is primarily caused by the AD/DA conversions, and secondarily by the processor speed. Since converters these days are all pretty fast, as phil_m suggested, the difference between the Helix and GT1000 could be the oversampling, 48k vs 96k.

I see.

So I need YOUR permission to post on this thread/

F***ing lollipop.

[rd2rk]

7 hours ago, somebodyelse said:

I see.

So I need YOUR permission to post on this thread/

F***ing lollipop.

 

No. You need to check your negative attitude at the door.

In a thread about cats you post about dogs, then curse out the OP (that's me) for reminding you of what the thread is about?

Keep it up with the foul language and personal attacks, I'm sure you'll become very popular around here.

 

 

[cruisinon2]

Put 'em up! Put 'em up!...;)